Probiotic supplementation prevents high-fat, overfeeding-induced, insulin resistance in humans

Abstract

The purpose of this study was to determine whether probiotic supplementation (Lactobacillus casei Shirota [LcS]) prevents diet-induced insulin resistance in humans. Seventeen healthy individuals were randomised to probiotic (n = 8) or control (n = 9) groups. The probiotic group consumed an LcS-fermented milk drink twice daily for 4 weeks whereas the control group received no supplementation. Subjects maintained their normal diet for the first 3 weeks of the study, after which they consumed a high-fat (65% energy) high-energy (+50% kcal) diet for 7 days. Whole body insulin sensitivity was assessed via an oral glucose tolerance test conducted before and after overfeeding. Body mass increased by 0.6 ± 0.2 kg in the control group (p < 0.05) and 0.3 ± 0.2 kg in the probiotic group (p > 0.05). Fasting plasma glucose concentrations increased following 7 days of overeating (control group only; 5.3 ± 0.1 vs. 5.6 ± 0.2 mmol/L, p < 0.05) whereas fasting serum insulin concentrations were maintained in both groups. Glucose AUC increased by 10% (from 817 ± 45 to 899 ± 39 mmol/L/120 min, p < 0.05) and whole body insulin sensitivity decreased by 27% (from 5.3 ± 1.4 to 3.9 ± 0.9, p < 0.05) in the control group, whereas normal insulin sensitivity was maintained in the probiotic group (4.4 ± 0.8 and 4.5 ± 0.9 before and after overeating, respectively, p > 0.05). These results suggest that probiotic supplementation may be useful in the prevention of diet-induced metabolic diseases such as type II diabetes.

Introduction

Insulin resistance is a major characteristic of obesity and type II diabetes. A number of metabolically active tissues and several potential mechanisms have been implicated in the pathophysiology of insulin resistance, such as impaired GLUT4 translocation to the cell membrane and reduced skeletal muscle glucose uptake ⁽¹⁾, elevated hepatic glucose production ⁽²⁾ and impaired β-cell function, which leads to reduced insulin secretion ⁽³⁾. In addition to these well-described defects, there is emerging evidence to suggest that changes in the gut microbiota might also play an important role in the development of human metabolic disease, through a mechanism that is linked to increased gut permeability, metabolic endotoxemia and systemic low-grade inflammation ^(4,5-7).

A clear association has been demonstrated between metabolic disease and compositional changes in the gut microbiota, with a lower abundance of Firmicutes and a higher proportion of Bacteroidetes and Proteobacteria in type II diabetic patients when compared with non-diabetic controls ⁽⁸⁾. Similar results have been found in rodent studies in which insulin resistance has been induced via short-term high-fat overfeeding ⁽⁷⁾. As well as altering the composition of the gut microbiota, high fat diets have been shown to elevate systemic lipopolysaccharide (LPS) concentrations ^(9,10), whereas selective modification of the gut microbiota, through prebiotic supplementation, has been shown to reduce high-fat diet-induced metabolic endotoxemia and lower intestinal permeability in obese diabetic mice ⁽⁶⁾. Furthermore, probiotic supplementation has been shown to improve glycaemic control and reduce metabolic endotoxemia in diet-induced obese mice ⁽¹¹⁾. Therefore, prebiotic and/or probiotic supplementation may be a useful strategy to improve metabolic health and prevent diet-induced insulin resistance and type II diabetes in humans.

Consumption of probiotic yogurt has already been shown to reduce fasting blood glucose concentrations and glycosylated haemoglobin (HbA1c) levels in type 2 diabetic patients ⁽¹²⁾, but whether or not probiotics can also prevent diet-induced insulin resistance in otherwise healthy individuals is not yet known. Therefore, we tested the hypothesis that 4 weeks of supplementation with probiotics (Lactobacillus casei Shirota [LcS]) would prevent insulin resistance induced by short-term, high-fat, overfeeding in healthy young males and females. In this paper, we provide novel evidence that probiotics preserve glycaemic control and prevent insulin resistance during a dietary challenge consisting of severe lipid overload, suggesting that probiotics might be useful in the fight against human metabolic disease.

Methods

Subjects

Seventeen healthy individuals volunteered for this study (14 males and 3 females; their physical characteristics can be seen in Table 1). They were recruited from the student population and the local community. No payments or other incentives were made for participation. The sample size for this experiment was estimated based on pilot data from our laboratory in which we observed a large decrease in insulin sensitivity after short-term, high-fat overfeeding. The inclusion criteria required subjects to be physically active (exercising at least 3 times per week for more than 30 minutes at a time), non-smokers, free from cardiovascular or metabolic disease and not taking any medication, weight stable for at least 6 months, and with a normal body mass index (BMI: 18.5-24.9 kg/m²). Subjects were excluded from the study if they had taken any probiotic or prebiotic supplements within the previous 3 months. Due to the nature of the dietary intervention, vegetarian and vegans were excluded from the study. This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the Loughborough University Ethical Committee for Human Participants. Written informed consent was obtained from all subjects.

Table 1.

Subject characteristics before and after 7 days of overeating

	Control group (n = 9; 7 males, 2 females)				Probiotic group (n = 8; 7 males, 1 female)
	Baseline		Overfed		Baseline		Overfed
	Mean	SE	Mean	SE	Mean	SE	Mean	SE
Age (years)	24	2	-	-	25	2	-	-
Height (m)	1.72	0.02	-	-	1.77	0.03	-	-
Body mass (kg)	72.1	4.8	72.7*	4.8	73.4	2.3	73.7	2.4
BMI (kg/m2)	24.2	1.2	24.4*	1.2	23.5	0.6	23.6	0.7

Values are means ± SE. BMI, body mass index.

^*Significantly different from baseline, p < 0.05.

Pre-testing/pre-screening

Subjects attended the laboratory 3 weeks prior to the start of the study for an initial pre-screening, which included an assessment of body mass index and fasting plasma glucose and insulin levels.

General study design

After the initial pre-screening visit, subjects were randomly assigned to one of two groups (control group, n = 9 [7 males, 2 females]; probiotic supplemented group, n = 8 [7 males, 1 female]). The probiotic that was used in this study was Lactobacillus casei Shirota (LcS; commercially available as the fermented milk drink Yakult Light). Both groups maintained their habitual food intake during the first 3 weeks of the study (days 1-21), but the probiotic group also consumed 2 × 65 mL of Yakult Light each day. An oral glucose tolerance test (OGTT) was performed on day 22 for assessment of baseline insulin sensitivity. After completing the first OGTT, subjects were provided with a high-fat (65% energy) high-energy (approximately 50% increase in energy intake) diet for 7 days. The probiotic group continued to consume 2 × 65 mL of Yakult Light throughout the 7-day overfeeding period. We selected a 4-week LcS supplementation period as previous experiments have demonstrated that this is sufficient to alter the composition of the gut microbiota in humans ^(13,14,15). On day 29, both groups of subjects returned to the laboratory for a second OGTT for assessment of post-overfeeding insulin sensitivity.

Experimental procedures

On the experimental days (days 22 and 29: OGTTs), subjects arrived at the laboratory in the morning (between 7.00 and 9.00am) after an overnight fast of at least 10 hours. After voiding and being weighed, a 20 gauge Teflon catheter (Venflon, Becton Dickinson, Plymouth, UK) was inserted into an antecubital vein of one arm to allow repeated blood sampling during the 2 h oral glucose tolerance test. A fasted blood sample (10 mL; t = 0) was obtained before subjects ingested a 25% glucose solution (75 grams of glucose dissolved in 300 mL of water). Additional 10 mL blood samples were obtained at 15, 30, 45, 60, 90 and 120 minutes after glucose ingestion. Blood samples were divided equally between Vacutainer tubes containing either K₂EDTA or a clotting catalyst (Becton Dickinson, Plymouth, UK) for separation of plasma and serum, respectively. EDTA tubes were stored on ice whereas serum tubes were left at room temperature until complete clotting had occurred. Blood tubes were then centrifuged at 2,300 g for 10 min (EDTA, 4°C; serum 20°C) and the resulting plasma or serum removed and stored at −20°C until later analysis.

Diet records, analysis and compliance during overfeeding

Subjects were provided with standardised forms and digital kitchen scales for the purpose of recording weighed food intake on 3 days each week during the pre-experimental period (days 1-21). Subjects also received detailed written and verbal instructions on how best to complete these records. These records were then used for the assessment of habitual energy intake and diet composition so that individualised diets could be planned and prepared for the 7-day overfeeding period. The analysis of diet records and planning of food intake for the overfeeding period was performed using WISP V4.0 (Tinuviel Software, Anglesey, UK). The overfeeding period was designed to increase energy intake by 50% compared to habitual food intake and consisted of foods that were predominantly high in fat (65% of energy consumed as fat). During the overfeeding period, all food was purchased and prepared by the research team and then delivered to the subjects. Subjects were instructed to eat everything that was provided to them, not to eat any other additional foods, and to return any unwanted food so that it could be weighed, and the diet values adjusted if necessary. All subjects were informed of the importance of strict diet adherence and we are confident that the subjects were fully compliant with the diet intervention.

Blood analysis

Plasma samples were analysed using commercially available spectrophotometric assays for glucose (Glucose PAP, Horiba Medical, Northampton, UK) and triglyceride (Triglyceride PAP, Horiba Medical, Northampton, UK) concentrations using a semi-automatic analyser (Pentra 400, Horiba Medical, Northampton, UK). The coefficient of variation (CV) for plasma glucose analysis was between 0.4 and 0.8% for samples with high (10.9 mmol/L; glucose ingestion) and normal (4.7 mmol/L; fasted) glucose concentrations. The CV for plasma triglyceride analysis was between 1.6 and 3.4% for samples with normal (1.1 mmol/L) and low (0.5 mmol/L) triglyceride concentrations.

Serum insulin concentrations were analysed using an enzyme-linked immuno-sorbent assay (ELISA EIA-2935, DRG Instruments GmBH, Marburg, Germany). The CV for serum insulin analysis was 3.0% for a mid-range sample (54 μU/mL; glucose ingestion).

Calculations

Plasma glucose and serum insulin concentrations from the OGTT were used to determine whole-body insulin sensitivity using the Matsuda insulin sensitivity index (ISI):

$$\mathrm{ISI}=\frac{10000}{\sqrt{\begin{array}{c}(\mathrm{FPG}\cdot \mathrm{FPI})\cdot \left(\mathit{\text{mean}}\mathrm{OGTT}\mathit{\text{insulin}}\mathit{\text{concentration}}\right)\hfill \\ \cdot \left(\mathit{\text{mean}}\mathrm{OGTT}\mathit{\text{glucose}}\mathit{\text{concentration}}\right)\hfill \end{array}}}$$

where FPG is the fasting plasma glucose concentration, FPI is the fasting plasma insulin concentration and 10000 represents a constant that allows numbers ranging between 1 and 12 to be obtained. The square root conversion is used to correct the nonlinear distribution of values ⁽¹⁶⁾.

Area under the curve (AUC)

AUC was calculated using the trapezoidal rule with zero as the baseline.

Statistics

The primary outcome measurement was a change in insulin sensitivity (determined from plasma glucose and serum insulin concentrations during an oral glucose tolerance test). Secondary outcomes measurements were changes in body mass, body mass index and fasting plasma triglyceride concentrations. Data analysis was performed using SPSS version 21.0 for Windows (SPSS Inc, Chicago, IL, United States). Data are expressed as means ± SE. In order to compare potential differences in the metabolic responses to overfeeding, a two-way (pre vs. post-overfeeding) repeated measures analysis of variance (ANOVA) was conducted with a between subject variable (control vs. probiotic group) with post-hoc analysis where appropriate. Statistical significance was accepted if p < 0.05.

Results

Weight gain and BMI with overeating

By the end of the 7 day overfeeding period, subjects in the control group had gained 0.6 ± 0.2 kg in body mass (p < 0.05) whereas the increase in body mass for the probiotic supplemented group was smaller and not significant (0.3 ± 0.2 kg, p > 0.05). No significant changes in BMI occurred as a result of 7 days of overeating. Data for body mass and BMI are provided in Table 1.

Fasting plasma substrates and serum insulin

Fasting plasma glucose concentrations were increased following 7 days of overeating (control group only; 5.3 ± 0.1 vs. 5.6 ± 0.2 mmol/L, p < 0.05) whereas fasting serum insulin concentrations were well maintained. Fasting plasma triglyceride concentrations were decreased following overeating (both groups; p < 0.05). Fasting data provided in Table 2 were obtained from the zero time point (before glucose ingestion) of the OGTTs that were performed before and after overfeeding.

Table 2.

Fasting plasma substrate and serum insulin concentrations before and after 7 days of overeating

	Control group (n = 9; 7 males, 2 females)				Probiotic group (n = 8; 7 males, 1 female)
	Baseline		Overfed		Baseline		Overfed
	Mean	SE	Mean	SE	Mean	SE	Mean	SE
Glucose (mmol/L)	5.3	0.1	5.6*	0.2	5.8	0.1	5.8	0.1
Insulin (μU/mL)	12	1	12	2	12	1	12	2
TG (mmol/L)	1.4	0.3	0.9*	0.1	1.4	0.2	1.0*	0.2

Values are means ± SE. TG, triglyceride.

^*Significantly different from baseline, p < 0.05.

Energy intake and diet composition

Habitual energy intake and diet composition was similar between the control and probiotic supplemented groups (Table 3). The intentional overfeeding period increased energy intake by 51.9 and 51.3% for control and probiotic groups, respectively (p < 0.05). The majority of this additional energy was provided by an increase in dietary fat intake.

Table 3.

Daily energy intake and diet composition before and during 7 days of overeating

	Control group (n = 9; 7 males, 2 females)				Probiotic group (n = 8; 7 males, 1 female)
	Baseline		Overfed		Baseline		Overfed
	Mean	SE	Mean	SE	Mean	SE	Mean	SE
Energy (kJ)	9934	759	15091*	1079	10751	581	16265*	865
Protein (g)	97	8	117*	9	103	6	127*	8
Carbohydrate (g)	292	17	213*	13	317	11	226*	15
Fat (g)	91	11	259*	11	103	12	281*	15

Values are means ± SE.

^*Significantly different from baseline, p < 0.05.

Oral glucose tolerance test

The plasma glucose and serum insulin responses to an oral glucose tolerance test before and after overfeeding can be seen in Figure 1. In the control group, overfeeding caused a 10% increase in glucose AUC (from 817 ± 45 to 899 ± 39 mmol/L/120 min, p < 0.05, Figure 1A) whereas no change was observed in the probiotic supplemented group (baseline; 866 ± 49 mmol/L/120 min, overfed; 888 ± 53 mmol/L/120 min, Figure 1B). A similar trend was observed for the insulin responses (Figures 1C & 1D) but this did not reach statistical significance.

Figure 1.

Plasma glucose (A, control group; B, probiotic group) and serum insulin (C, control group; D, probiotic group) concentrations during a 2 hour oral glucose tolerance test conducted before and after 7 days of overeating. Values are means ± SE (n = 9, control group; n = 8 probiotic group).

Insulin sensitivity

After 7 days of overfeeding, whole-body insulin sensitivity of control subjects was impaired by 27% (decrease from 5.3 ± 1.4 to 3.9 ± 0.9, p < 0.05) whereas no change in insulin sensitivity was observed in the probiotic supplemented group (Figure 2).

Figure 2.

The Matsuda insulin sensitivity index calculated during an oral glucose tolerance test conducted before and after 7 days of overeating for control and probiotic supplemented groups. Values are means ± SE (n = 9, control group; n = 8 probiotic group). *Significantly different from baseline, p < 0.05.

Discussion

The main finding of the present study is that high-fat overfeeding decreased insulin sensitivity by approximately 27% in healthy young males and females, but that supplementation with the LcS probiotic prior to and throughout the overfeeding period preserved glycaemic control and maintained insulin action. These results provide further indirect evidence that compositional changes in the gut microbiota are involved in the development of human metabolic disease and that probiotic supplementation could be used to prevent insulin resistance caused by excessive consumption of high fat foods (i.e., a Westernised diet).

Numerous rodent studies have reported that short-term (3 days to 4 weeks) adherence to a high-fat diet induces insulin resistance ^(17,18-23), and similar results have been reported in a small number of recent human studies ^(2,24,25), which is why we chose to adopt this model in the present study. Although the underlying mechanisms for the development of insulin resistance remain unclear, a popular theory is that excessive consumption of high-fat foods can lead to an accumulation of reactive intramyocellular lipids, such as ceramide and diacylglycerol, which inhibit insulin signalling (insulin/IRS-1/PI3-K/AKT pathway) and impair GLUT4 translocation to the cell membrane, thereby reducing skeletal muscle glucose uptake ^(26,27-28). Another potential mechanism is systemic low-grade inflammation, as the prevalence of insulin resistance is associated with an increase in lipopolysaccharide (LPS; a component of the cell wall of gram-negative bacteria) concentrations in the blood ^(29,30). Furthermore, in experimental models of metabolic endotoxemia, using intravenous or subcutaneous LPS infusion, it has been found that LPS acts as a trigger for pro-inflammatory cytokine production ^(4,31), which in turn leads to impaired insulin action through increased serine phosphorylation of the insulin receptor substrate-1 (IRS-1) ^(31,32). Interestingly, the increase in systemic LPS that is associated with the metabolic syndrome is thought to originate from the gram-negative bacteria living in the gut ^{(5,6-7,10,30,33)}. Normally, the intestinal epithelium would act as a physical barrier to prevent LPS translocation into the circulation. However, there is a reduced expression of intestinal epithelial tight junction proteins (ZO-1 and occludin) and an increase in gut permeability (i.e., a decrease in gut-barrier function) in animal models of insulin resistance and type II diabetes ^(5,6). This decrease in gut-barrier function may be causally linked to diet-induced changes in the composition of the gut microbiota, as high-fat diets have been shown to increase the ratio of Bacteroidetes [gram-negative] to Firmicutes [gram-positive] and increase gut permeability in mice that became diabetic as a result of the diet ⁽⁷⁾. On top of this, selectively increasing the abundance of Bifidobacterium spp., through prebiotic supplementation, has been shown to reduce intestinal permeability in obese diabetic mice ⁽⁶⁾. In the context of the present study, this tends to suggest that the ability of probiotic supplementation to prevent high-fat diet-induced insulin resistance is most likely due to the maintenance of a favourable gut microbiota and preservation of gut-barrier function, although further mechanistic work will be necessary to confirm this.

The main limitation of this study is that we cannot confirm whether the high-fat diet altered the composition of the gut microbiota, impaired gut-barrier function, or lead to an increase in systemic inflammation. Nor can we confirm the impact of LcS probiotic supplementation on these physiological outcomes. Although it was not possible to identify the underlying mechanisms in the present study, we have discussed our observations and speculated upon the possible causes in the context of what has already been shown in animals. It is also important to keep in mind that this was a preliminary, proof of concept investigation into the efficacy for probiotic supplementation to maintain human metabolic health. Therefore, it was important to assess changes in whole body insulin sensitivity (i.e., the primary outcome measurement), and to confirm our experimental hypothesis, before conducting the more costly and time consuming mechanistic experiments. In future studies it will be necessary to perform faecal DNA analysis in order to determine diet and/or probiotic-induced changes in gut microbiota composition, as well as measuring markers of gut-barrier function and changes in systemic and tissue inflammation. Nonetheless, the preliminary finding that 4 weeks of LcS probiotic supplementation maintained glycaemic control and prevented insulin resistance during a dietary challenge that consisted of severe lipid overload (approximately 270 g of fat per day) suggests that diet-induced changes in the gut microbiota are likely to play an important role in human metabolic health and disease. Anti-diabetic properties of certain probiotics have been reported in several rodent studies ^{(11,34,35-37)} and one human study ⁽¹²⁾. In that study, 64 type II diabetic patients consumed either a probiotic yogurt containing Lactobacillus acidophilus La5 and Bifodobacterium lactis Bb12 or a placebo, probiotic free, yogurt for 6 weeks. The authors reported a significant decrease in both fasting blood glucose and glycosylated haemoglobin (HbA1c) concentrations as well as an improvement in antioxidant status. We did not find a reduction in fasting glucose or insulin concentrations following probiotic supplementation, most likely because of the fact that we tested healthy young subjects with normal baseline values. However, our results demonstrate that regular consumption of LcS probiotics may offer protection against the development of diet-induced insulin resistance. Not all studies have reported metabolic health benefits following LcS supplementation. Tripolt et al., reported that 12 weeks of supplementation with LcS failed to improve insulin sensitivity or enhance β-cell function in subjects with the metabolic syndrome ⁽³⁸⁾. The discrepancy between their previous findings and our current investigation may be related to the recruitment of a clinical population as opposed to healthy volunteers. With this in mind, it is reasonable to suggest that LcS probiotics could be used more effectively in the prevention, rather than the treatment, of human metabolic disease.

We observed a significant decrease in fasting plasma TG concentrations after seven days of overfeeding, which is consistent with the findings of another very recent study ⁽³⁹⁾. However, the reduction in fasting plasma TG is most likely due to changes in the macronutrient composition of the diet, rather than the amount of energy consumed. Removing carbohydrate from the diet and replacing it with an isoenergetic amount of fat has been shown to reduce plasma TG concentrations ⁽⁴⁰⁾, whereas increasing the amount carbohydrate in the diet has been shown to have the opposite effect and actually increases fasting plasma TG concentrations ⁽⁴¹⁾. This effect has been termed carbohydrate-induced hypertriglyceridemia ⁽⁴²⁾. In the present study, the amount of carbohydrate that was consumed during the overfeeding period was reduced in both absolute and relative terms when compared to the subject’s habitual food intake (~220 vs. ~305 g/day), and the additional energy intake was met through a large increase in the amount of fat that was consumed, suggesting that the reduction in fasting TG may be related to the removal of carbohydrate from the diet. Without the use of stable isotope tracers it is difficult to speculate on the reason for a reduction in plasma TG concentration after overfeeding, but it must be related to changes in TG turnover; be it a decrease in synthesis rate, an increase in clearance rate, or a combination of both.

In conclusion, we have shown that probiotic supplementation has the potential to prevent high-fat diet-induced insulin resistance in humans, and this warrants confirmation in a larger, placebo-controlled study. Further mechanistic studies are also necessary to confirm the underlying mechanism for this effect.