Beneficial effect of oral administration of Lactobacillus casei strain Shirota on insulin resistance in diet-induced obesity mice

Authors


Yasuto Yoshida, Food Research Department, Yakult Central Institute for Microbiological Research, 1796 Yaho, Kunitachi-shi, Tokyo 186-8650, Japan. E-mail: yasuto-yoshida@yakult.co.jp

Abstract

Aims:  This study aimed at determining whether oral administration of a probiotic strain, Lactobacillus casei strain Shirota (LcS), can improve insulin resistance, which is the underlying cause of obesity-associated metabolic abnormalities, in diet-induced obesity (DIO) mice.

Methods and Results:  DIO mice were fed a high-fat diet without or with 0·05% LcS for 4 weeks and then subjected to an insulin tolerance test (ITT) or oral glucose tolerance test (OGTT). Oral administration of LcS not only accelerated the reduction in plasma glucose levels during the ITT, but also reduced the elevation of plasma glucose levels during the OGTT. In addition, plasma levels of lipopolysaccharide-binding protein (LBP), which is a marker of endotoxaemia, were augmented in the murine models of obese DIO, ob/ob, db/db and KK-Ay and compared to those of lean mice. LcS treatment suppressed the elevation of plasma LBP levels in DIO mice, but did not affect intra-abdominal fat weight.

Conclusions:  LcS improves insulin resistance and glucose intolerance in DIO mice. The reduction in endotoxaemia, but not intra-abdominal fat, may contribute to the beneficial effects of LcS.

Significance and Impact of the Study:  This study suggests that LcS has the potential to prevent obesity-associated metabolic abnormalities by improving insulin resistance.

Introduction

Insulin plays a significant role in the regulation of glucose homoeostasis, lipid metabolism and blood pressure. The failure of target organs to respond normally to the action of insulin is termed insulin resistance, which together with resultant compensatory hyperinsulinaemia leads to a cluster of metabolic abnormalities, such as glucose intolerance, hyperlipidaemia, hepatic steatosis and hypertension (Paoletti et al. 2006; Shoelson et al. 2006; de Luca and Olefsky 2008). Therefore, insulin resistance is considered the pathophysiologic basis of metabolic syndrome, a major risk factor of coronary heart disease and several other diseases. One major determinant of insulin resistance is an excess accumulation of visceral fat, which causes a chronic low-grade inflammation characterized by increased macrophage infiltration and pro-inflammatory adipokine production (Shoelson et al. 2006; de Luca and Olefsky 2008). Pro-inflammatory adipokines interfere with the insulin-signalling pathway in peripheral tissues and facilitate the development of insulin resistance (Shoelson et al. 2006; de Luca and Olefsky 2008).

Recently, the pathogenic role of increased Toll-like receptor 4 (TLR4) signalling in obesity-associated inflammation has been well documented (Brun et al. 2007; Cani et al. 2007, 2008). TLR4 is expressed on a variety of cell types, including immune cells and adipocytes, and serves a crucial role in host cell responses to microbial pathogens. TLR4 recognizes exogenous ligands, such as lipopolysaccharide (LPS), a major constituent of the Gram-negative bacterial outer membrane, and endogenous ligands, such as nonesterified fatty acids (NEFAs), and subsequently triggers innate immunity through nuclear factor kappa B (NF-κB)-mediated activation of inflammatory genes. Recent studies have shown that obesity leads not only to an increased release of NEFAs from adipose tissue but also to moderately elevated influxes of LPS into the circulation because of increased intestinal permeability (Brun et al. 2007; Creely et al. 2007; Cani and Delzenne 2009). Moreover, Cani et al. (2007) have shown that the chronic infusion of low doses of LPS in lean mice is accompanied by the infiltration of macrophages into adipose tissue, insulin resistance and steatosis. These findings suggest that chronic low-grade endotoxaemia (metabolic endotoxaemia) is involved in the pathogenesis of metabolic abnormalities and that therapeutic strategies to reduce the influx of LPS may help prevent the development of insulin resistance. However, as the common method for measuring LPS, the chromogenic Limulus amebocyte lysate (LAL) assay, is sensitive to several nonspecific activators and inhibitors in plasma, it is difficult to accurately quantify low levels of LPS in plasma (Ruiz et al. 2007; Moreno-Navarrete and Fernández-Real 2009; Sun et al. 2010).

Lipopolysaccharide-binding protein (LBP) is a 60-kDa acute-phase protein synthesized in liver that acts as a central mediator in TLR4-mediated inflammatory responses by transferring LPS to CD14 to initiate signalling (Moreno-Navarrete and Fernández-Real 2009). As plasma LBP levels are closely correlated with plasma LPS in various clinical settings, LBP is considered an indicator of endotoxaemia and the intensity of TLR4 signalling triggered by LPS (Gubern et al. 2006; Moreno-Navarrete and Fernández-Real 2009; Sun et al. 2010). Indeed, it has been reported that plasma LBP levels correlate positively to obesity and insulin resistance in humans (van Dielen et al. 2001; Gubern et al. 2006; Ruiz et al. 2007; Moreno-Navarrete and Fernández-Real 2009; Sun et al. 2010), which suggests that the sustained activation of TLR4 signalling by LPS partially accounts for obesity-associated inflammation and the resultant metabolic abnormalities.

Lactobacillus casei strain Shirota (LcS) is a representative probiotic strain that is commercially available as a health food in a number of countries. In several murine models, the oral administration of LcS improves inflammatory disorders, including inflammatory bowel disease (Matsumoto et al. 2005, 2009), arthritis (Kato et al. 1998), type 1 diabetes (Matsuzaki et al. 1997a), systemic lupus (Mike et al. 1999) and indomethacin-induced small intestinal injury (Watanabe et al. 2009). Interestingly, several of these beneficial effects are exhibited by nonviable preparations, as well as viable LcS. For example, Matsuzaki et al. (1997b) reported that oral administration of heat-killed LcS suppresses the elevation of plasma glucose levels in an obese type 2 diabetic murine model (KK-Ay). However, it remains to be determined whether LcS affects insulin resistance resulting from obesity-associated inflammation.

This study aimed at assessing whether oral administration of LcS improves insulin resistance in a high-fat diet-induced obesity (DIO) mouse model, a relevant model of human obesity (Thakker et al. 2006). Moreover, the effects of LcS on glucose intolerance, hepatic lipid accumulation and plasma LBP levels were also examined.

Materials and methods

Strains and culture conditions

Lactobacillus casei strain Shirota YIT 9029 (LcS) was obtained from the culture collection of the Yakult Central Institute for Microbiological Research (Tokyo, Japan). To prepare bacterial cells for administering to mice, LcS was first precultured in ILS medium (Rogosa et al. 1951) at 37°C for 24 h. Seventy millilitres of the preculture was then inoculated into 7 l of ILS medium in a 10-l fermenter and incubated at 37°C for 24 h. The cultured cells were collected and washed three times with distilled water by centrifugation at 9000 g for 30 min at 4°C. The cells were heat-killed at 100°C for 30 min, lyophilized and stored at −20°C until use.

Mice and diet

All mice were maintained and treated in accordance with the guidelines of the Ethical Committee for Animal Experiments of the Yakult Central Institute for Microbiological Research. Ten-week-old male C57BL/6J DIO (DIO) mice, which had been fed a commercial high-fat (HF) diet (D12492; Research Diets, Inc., New Brunswick, NJ, USA) from 4 weeks of age, were obtained from Charles River Japan (Yokohama, Japan). Male C57BL/6J mice, B6.V-Lepob/J (ob/ob) mice and BKS.Cg-m+/+Leprdb/J (db/db) mice were purchased from Charles River Japan at 10 weeks of age. Ten-week-old male KK-Ay/Ta mice were purchased from Clea Japan (Tokyo, Japan). All mice were housed individually in plastic cages in a room with controlled lighting (lights on 08:30–20:30), temperature (25 ± 1°C) and humidity (60 ± 5%) under conventional conditions.

DIO mice freely received tap water and an HF diet. After a 14-day adaptation period, the mice were assigned randomly to two groups and further fed an HF diet or an HF diet supplemented with 0·05% (w/w) LcS for 5 weeks. Body weight was recorded once a week, and diet intake was recorded every 2–3 days. The experimental diets were prepared in solid form, stored at 4°C and fed to the mice within 5 weeks of preparation. After 4 weeks of treatment, mice were subjected to an insulin tolerance test (ITT) or an oral glucose tolerance test (OGTT) to evaluate insulin resistance and glucose intolerance, respectively. At the end of the experimental period, mice were fasted for 6 h, anesthetized with diethyl ether and then exsanguinated from the heart. The liver was then perfused with saline in situ, excised and weighed. The mesenteric, perirenal and epididymal fat was removed from each mouse and weighed.

C57BL/6J, ob/ob, db/db and KK-Ay mice received a normal mouse-chow diet (MF diet; Oriental Yeast, Tokyo, Japan) for 1 week under the conditions described earlier for DIO mice and then were subjected to blood sampling from heart.

Insulin tolerance test

Diluted regular human insulin (Humulin R; Eli Lilly Japan K.K., Kobe, Japan) was injected intraperitoneally into mice at a dose of 0·75 units kg−1 body weight under nonfasting conditions. Blood samples were collected with heparinized capillary tubes from the tail vein before and 30, 60, 90 and 120 min after insulin injection and immediately placed on ice. Plasma was prepared by centrifugation at 830 g for 15 min at 4°C and applied to the assay of plasma glucose.

Oral glucose tolerance test

The evaluation of glucose intolerance was performed using the OGTT, as previously described (Andrikopoulos et al. 2008). Briefly, after 6 h of fasting, glucose was administered to mice at a dose of 2 g kg−1 body weight by oral gavage. Blood samples were collected with heparinized capillary tubes from the tail vein before and 15, 30, 60 and 120 min after glucose administration and immediately placed on ice. Plasma was prepared by centrifugation at 830 g for 15 min at 4°C and applied to the assay of plasma glucose and insulin.

Chemical assays

Plasma glucose was quantified using the glucose CII-test Wako (Wako Pure Chemical Industries, Osaka, Japan). Insulin and LBP in plasma were assayed using a mouse insulin colorimetric enzyme-linked immunosorbent assay (ELISA) kit (Revis Insulin kit; Shibayagi, Gunma, Japan) and a mouse LBP ELISA kit (HyCult Biotechnology, Uden, The Netherlands), respectively.

Total fat was extracted from liver as previously described (Folch et al. 1957) and weighed. The fat extracts were dissolved in methanol, and triglyceride (TG) and total cholesterol (TC) were enzymatically analysed using the Triglyceride E-test Wako and HDL-cholesterol E-test Wako (Wako Pure Chemical Industries), respectively.

Statistical analysis

All data are expressed as the mean ± SEM. Statistical significance was determined using either Student’s t-test or Dunnett’s multiple comparison test, and < 0·05 was considered significant.

Results

In a preliminary experiment, it was confirmed that 12-week-old male DIO mice, which had been fed a commercial HF diet from 4 weeks of age, developed obesity, insulin resistance and glucose intolerance (data not shown). Therefore, LcS [0·05% (w/w) in HF diet] or HF diet alone (control group) were fed to the 12-week-old DIO mice for 4 weeks, and the effects of LcS on insulin resistance and glucose intolerance were examined by the ITT and OGTT, respectively. In the ITT, although the nonfasting plasma glucose levels were similar in both groups of mice, plasma levels of glucose at 30, 60, 90 and 120 min after insulin injection were significantly lower in the LcS group than in the control group (< 0·01 or 0·05) (Fig. 1). A similar trend was also observed in the OGTT, with plasma glucose levels after glucose loading being significantly lower at 30, 60 and 120 min in the LcS group (< 0·01 or 0·05) (Fig. 2a), although no significant differences in the plasma levels of insulin were observed between the two groups (Fig. 2c). The effects of LcS administration were clearly reflected in the area under the curves (AUC) determined from the glucose and insulin concentration curves, which showed a significant reduction in the AUC of plasma glucose (0–120 min) (< 0·01) (Fig. 2b), but no effect on the AUC of plasma insulin (0–30 min) (Fig. 2d).

Figure 1.

 Effect of LcS on insulin sensitivity in diet-induced obesity (DIO) mice. DIO mice, which had been fed a high-fat (HF) diet from 4 to 12 weeks of age, were fed an HF diet (control) or an HF diet containing 0·05% (w/w) LcS for an additional 4 weeks. Under nonfasting conditions, an insulin tolerance test was then performed and plasma glucose levels were measured at several time points (0–120 min) after insulin loading. Values are expressed as the mean ± SEM (n = 10 mice in each group). *< 0·05, **< 0·01 (Student’s t-test). •, Control; ○, LcS.

Figure 2.

 Effects of LcS on glucose and insulin responses in an oral glucose tolerance test (OGTT) in diet-induced obesity (DIO) mice. DIO mice, which had been fed a commercial high-fat (HF) diet from 4 to 12 weeks of age, were fed an HF diet (control) or an HF diet containing 0·05% (w/w) LcS for an additional 4 weeks. After mice were fasted for 6 h, an OGTT was performed and plasma glucose and insulin levels were measured. The time course of plasma glucose (120 min) (a), area under the curves (AUC) of plasma glucose (b), time course of plasma insulin (30 min) (c) and AUC of plasma insulin (d) were then plotted. Values are expressed as the mean ± SEM (n = 12 mice in each group). *< 0·05, **< 0·01 (Student’s t-test). •, Control; ○, LcS.

While no significant differences were observed in food intake, body weight gain, intra-abdominal fat weight and liver weight between the two groups (Table 1), the total fat and TG levels in liver were significantly lower in the LcS group than in the control group (< 0·05) (Fig. 3).

Table 1.   Effect of LcS on body weight, diet intake and tissue wet weights in DIO mice
Measurement parameterControl (n = 10)LcS (n = 10)
  1. LcS, Lactobacillus casei strain Shirota; DIO, diet-induced obesity; HF, high fat.

  2. DIO mice, which had been fed a commercial HF diet from 4 to 12 weeks of age, were fed an HF diet (control) or an HF diet containing 0·05% (w/w) LcS for an additional 5 weeks. Values are expressed as the mean ± SEM.

Body weight (g)
 Initial32·3 ± 0·632·3 ± 0·5
 Final39·2 ± 1·138·2 ± 0·8
Body weight gain (g/4 weeks)6·88 ± 0·665·92 ± 0·49
Diet intake (g/day)2·63 ± 0·082·49 ± 0·06
Tissue weight (g)
 Mesenteric fat0·79 ± 0·080·73 ± 0·05
 Perirenal fat0·99 ± 0·060·99 ± 0·03
 Epididymal fat2·38 ± 0·102·43 ± 0·08
 Liver1·67 ± 0·061·63 ± 0·04
Figure 3.

 Effects of LcS on wet weights of abdominal fat and liver in diet-induced obesity (DIO) mice. DIO mice, which had been fed a commercial high-fat (HF) diet from 4 to 12 weeks of age, were fed an HF diet (control) or an HF diet containing 0·05% (w/w) LcS for an additional 5 weeks. After mice were fasted for 6 h, liver was perfused with saline and then excised. Total fat (a), triglyceride (TG) (b) and total cholesterol (TC) (c) levels in livers were then determined. Values are expressed as the mean ± SEM (n = 10 mice in each group). *< 0·05 (Student’s t-test).

Finally, the plasma levels of LBP, a marker of endotoxaemia, in four different murine models with obesity and/or diabetes, DIO, ob/ob, db/db and KK-Ay mice, were compared to those of lean C57BL/6J mice fed a normal diet. All obese and/or diabetic models used in this study had significantly higher levels of LBP than lean mice (< 0·05 or 0·01) (Fig. 4). The oral administration of LcS for 5 weeks to DIO mice suppressed the elevation of plasma LBP levels (P < 0·05) (Fig. 5).

Figure 4.

 Plasma lipopolysaccharide-binding protein (LBP) levels in several murine models of obesity. Four-week-old C57BL/6J male mice were fed either a high-fat or commercial normal diet (ND) for 8 weeks to establish diet-induced obesity and lean mice, respectively. Eleven-week-old male ob/ob, db/db and KK-Ay mice were fed an ND for 1 week. Subsequently, the plasma LBP levels in all groups of mice were determined. Values are expressed as the mean ± SEM for each group (= 10 mice in each group). *< 0·05, **< 0·01 vs lean mice (Dunnett’s test).

Figure 5.

 Effect of LcS on the plasma lipopolysaccharide-binding protein (LBP) levels in diet-induced obesity (DIO) mice. Twelve-week-old DIO mice were fed a high-fat (HF) diet or an HF diet containing 0·05% LcS for 5 weeks, and the plasma LBP levels were determined. Values are expressed as the mean ± SEM (n = 10 mice in each group). *< 0·05 (Student’s t-test).

Discussion

The present study has demonstrated that oral administration of LcS accelerated the lowering of plasma glucose levels during an ITT in DIO mice with previously established insulin resistance, indicating that LcS ameliorated insulin resistance. LcS treatment also reduced the elevation of plasma glucose levels after glucose loading in DIO mice, without significant changes in plasma insulin levels. These findings show that LcS has the potential to improve glucose intolerance through the amelioration of insulin resistance.

A number of studies have revealed that a reduction of visceral fat in animal models and humans is associated with increased insulin sensitivity (Fantuzzi and Mazzone 2007). LcS, however, showed no significant effect on body weight and intra-abdominal fat weight, suggesting that mechanisms other than the suppression of obesity are involved in the attenuation of insulin resistance.

Probiotics are traditionally defined as live micro-organisms that, when administered in adequate amounts, confer a health benefit on the host. LcS is a representative probiotic strain that has been shown to be effective in the prevention and treatment of a diverse spectrum of gastrointestinal disorders (Miyazaki and Matsuzaki 2008). Modifications of gut microbiota and host immune responses are considered as major mechanisms of probiotic action. The former modification requires the viability of probiotics in the gut lumen, although their viability is not necessarily required for immune modification (Matsuzaki et al. 1997a; Matsumoto et al. 2005; Nomoto 2005). Recently, several studies have proposed the gut microbiota as an environmental factor involved in the development of obesity-associated metabolic abnormalities (Turnbaugh et al. 2006; Cani and Delzenne 2009), suggesting the possibility that probiotics may affect metabolic abnormalities by modulating gut microbiota. In our study, however, the nonviable preparation of LcS improved insulin resistance and glucose intolerance, which strongly suggests that these effects of LcS are mediated through a mechanism other than the modification of gut microbiota.

Plasma levels of the endotoxaemia marker LBP are positively associated with obesity and insulin resistance in humans (van Dielen et al. 2001; Gubern et al. 2006; Sun et al. 2010) and are increased in obese subjects with metabolic complications, such as type 2 diabetes and nonalcoholic steatohepatitis (Gubern et al. 2006; Ruiz et al. 2007; Moreno-Navarrete and Fernández-Real 2009). Although these findings suggest that metabolic endotoxaemia is involved in the development of obesity and associated metabolic abnormalities, significant increases in plasma LBP have not been reported in obese animal models. The present study has demonstrated that DIO mice and three different genetic models with obesity and/or diabetes, ob/ob, db/db and KK-Ay mice, display higher levels of plasma LBP compared with their respective age-matched lean mice. DIO, ob/ob and db/db mice have been reported to have increased intestinal permeability and accompanying metabolic endotoxaemia (Brun et al. 2007; Cani et al. 2007, 2008). Therefore, these observations suggest that plasma LBP levels can serve as a marker for metabolic endotoxaemia in not only humans but also murine models of obesity. Moreover, the oral administration of LcS suppressed the elevation of plasma LBP levels in DIO mice, suggesting the possibility that LcS treatment attenuated metabolic endotoxaemia. Recent evidence indicates that metabolic endotoxaemia contributes to the development of obesity-associated inflammation and resultant metabolic abnormalities observed in high-fat DIO mice (Cani et al. 2007, 2008). Therefore, LcS treatment might improve obesity-associated inflammation.

An increase in intestinal permeability has been proposed as a mechanism for the development of metabolic endotoxaemia (Brun et al. 2007; Cani et al. 2007, 2008). Interestingly, growing evidence indicates that increased intestinal permeability and influx of foreign antigens from the gut lumen also play a pathogenic role in inflammatory disorders, including inflammatory bowel disease and type 1 diabetes (Sharma et al. 2010). Oral administration of LcS has been shown to prevent the development of such inflammatory diseases in a number of mouse models (Matsuzaki et al. 1997a; Matsumoto et al. 2005). It is assumed that a reduction in intestinal permeability might be a common mechanism underlying the protective effects of LcS in inflammatory diseases. As heat-killed preparations of several Lactobacillus strains have been reported to reduce intestinal permeability in human intestinal epithelial cell lines and a murine model of colitis (Seth et al. 2008; Miyauchi et al. 2009), it remains to be resolved whether LcS can improve intestinal permeability in inflammatory diseases.

Several probiotic strains have been reported to exert an antiobese effect in DIO murine models (Lee et al. 2007; Ma et al. 2008; Tanida et al. 2008); however, few studies have examined the effects of probiotics on insulin resistance. Recently, Ma et al. (2008) reported that a viable probiotic mixture (VSL#3) containing bifidobacteria, lactobacilli and Streptococcus thermophilus, but not heat-killed VSL#3, improves obesity, insulin resistance and hepatic steatosis in DIO mice. This result contrasts with the present study, as heat-killed LcS improved insulin resistance and hepatic fat accumulation without significant changes in body weight gain and intra-abdominal fat weight in DIO mice. Possible reasons for the differences between these studies are as follows: 1) the effects on obesity or insulin resistance vary depending on the probiotic strain, and 2) the viability of the probiotic strain is necessary to improve body weight. To verify these possibilities, further studies comparing the effects of several probiotic strains on obesity-associated markers and their dependency on viability are necessary.

In conclusion, we have demonstrated that the oral administration of LcS improves insulin resistance, glucose intolerance and fatty liver in DIO mice. Previous studies have also shown that LcS suppresses the development of diabetes in obese type 2 diabetic KK-Ay mice and that LcS-fermented milk exhibits a hypolipidaemic effect in Syrian hamsters (Matsuzaki et al. 1997b; Kikuchi-Hayakawa et al. 2000). Therefore, this probiotic strain is expected to prevent obesity-associated metabolic abnormalities through the improvement in insulin resistance. Additional studies are necessary to elucidate the detailed mechanisms of action and the efficacy of LcS treatment on human subjects.

Acknowledgements

We thank Drs K. Shida, M. Nanno, R. Tanaka and H. Sawada at the Yakult Central Institute for Microbiological Research for helpful advice and useful discussions related to the study. We also thank the members of the Animal Experiment Research Laboratory at our institute for maintenance of the mice.

Ancillary