Effect of dietary prebiotic supplementation on advanced glycation, insulin resistance and inflammatory biomarkers in adults with pre-diabetes: a study protocol for a double-blind placebo-controlled randomised crossover clinical trial

This is a 6.5-month randomised crossover controlled clinical trial (RCT) in which adults aged 40–60 years with diagnosed pre-diabetes will be enrolled. Potential study participants will be identified from General Practice (GP) clinics throughout South Gippsland, Victoria. The study design is presented in Figure 1.

Individuals aged between 40–60 years and diagnosed with prediabetes (Impaired Fasting Glucose or Impaired Glucose Tolerance) within the previous 12 months. Diagnosis will have been made at each individual’s local GP clinic after undertaking an Oral Glucose Tolerance Test (OGTT). Prediabetes was defined as a fasting plasma glucose concentration ≥ 6.1 and < 7.0 mmol/L followed by a 2-hour post glucose load glucose concentration < 7.8 mmol/L, or a fasting plasma glucose < 7.0 mmol/L followed by a 2-hour post glucose load glucose concentration ≥ 7.8 and < 11.1 mmol/L [38].

Individuals previously diagnosed with type 1, type 2 diabetes or impaired renal function (eGFR <90 mL/min/1.73 m²), individuals with known gastrointestinal pathology (coeliac disease, inflammatory bowel disease), pregnant women, smokers, individuals who have taken antibiotics, dietary prebiotic or probiotic nutritional supplements within the previous three months, individuals taking aspirin or Vitamin B, individuals who have made major dietary or lifestyle changes in the previous three months, individuals who are unwilling to provide blood, urine and stool samples or are unable to attend their local pathology collection centre.

The trial has received ethical approval from the Monash University Human Research Ethics Committee.

The minimum difference we wish to detect is 0.4 micromol/L serum CML (20% reduction in CML), with a standard deviation of 0.4 [39] with 5% Significance and 80% Power. This calculates as a total sample size of 18 individuals, plus 12 individuals to allow for withdrawals = 30 subjects required. Epidemiological studies have demonstrated a positive correlation between serum CML and all-cause and cardiovascular mortality, cardiovascular disease, glucose intolerance, impaired insulin secretion, renal impairment and diabetic vascular complications [40, 41]. A 0.4 micromol/L increase in serum CML concentration represented a 68% increased risk for all-cause mortality over seven years in a large prospective cohort study [42].

The study timeline is presented in Figure 2. Following recruitment and screening, consenting participants will undergo a baseline assessment at their local GP clinic. The baseline assessment will be undertaken by the research dietitian, who will collect demographic details, medical and social history (living situation, marital status, current occupation), physical activity questionnaire, dietary intake assessment, anthropometric measurements including Body Mass Index (BMI) and waist circumference. Participants will also be instructed to attend the local pathology centre to provide a stool sample, 24-hour urine collection, and have blood taken for analysis.

Randomisation of participants to intervention/placebo sequence will be completed by a local pharmacist using a permuted-block randomisation stratified by gender via the web site http://​www.​randomization.​com. The pharmacist will distribute the experimental and placebo nutritional supplements to participants in the appropriate sequence. Other than provision of supplements, the pharmacist will have no contact with study participants or involvement in data collection or analysis. Dietary supplements will be provided in sealed opaque packages which will be weighed at the conclusion of the study in order to assess compliance. The supplements will be packaged off-site by an external pharmaceutical packaging agency. Each package will contain a participant’s study identification number and will be labelled as either Supplement 1 or Supplement 2 corresponding to the first and second intervention periods. This will ensure all investigators and participants are blinded to the treatment. Blinding will cease only after statistical analysis of the data has been completed.

Participants will be randomly assigned to receive either 10 g of chicory-derived inulin/oligofructose powder (kindly provided by Beneo-Orafti Active Food Ingredients, Belgium) (intervention) or 10 g of maltodextrin powder (placebo) daily. Previous studies have demonstrated the bifidogenic effect of daily consumption of 5 g and 8 g dietary inulin supplements [43]. The inulin and maltodextrin powders to be consumed are both tasteless and can be mixed into hot or cold liquids or semi-solid foods. Participants will be instructed how to incorporate their supplement into their usual diet, and advised to gradually increase their dose over ten successive days until the target dose is reached. This stepped escalation in supplement dose aims to minimise gastrointestinal discomfort for participants, as a sudden increase in dietary prebiotic intake may result in increased stool frequency, abdominal bloating and flatulence until the bowel adapts to the increased fibre intake [44]. Written instructions will also be provided. Participants will be advised to consume each dietary supplement daily for 90 days, and otherwise maintain their usual dietary intake and level of physical activity. Gastrointestinal tolerance to the dietary supplements will be indicated by each study participant using a visual analogue scale as described below.

Visits to the GP clinic will be scheduled for each participant at the conclusion of both Intervention 1 and Intervention 2 treatment periods, in order to provide follow-up data. Information collected will include anthropometric measurements, dietary intake assessment and physical activity assessment. Each participant will complete a self-administered questionnaire designed to assess gastrointestinal tolerance to the dietary supplement. Completed questionnaires will be placed into sealed opaque envelopes in order to maintain blinding of the researcher collecting data during the follow-up visits. Participants will also be instructed to attend the local pathology centre to provide urine and stool samples, and have blood taken for analysis.

All adverse events will be documented.

The primary outcome of interest is the difference in serum AGE and AGE-precursor concentration (measured as CML and MG respectively) between experimental and control treatments. Secondary outcomes include HbA1c, insulin resistance (measured indirectly by homeostasis model assessment), antioxidant capacity (reduced glutathione), markers of oxidative stress on lipid molecules (urinary 8-isoprostanes), inflammatory biomarkers (serum IL-6, high sensitivity C-reactive protein, MCP-1, sRAGE), adhesion molecules (E-selectin), gut barrier integrity (TLR4), 24-hour urine and faecal CML and MG concentrations, faecal bacterial composition (Bifidobacterium spp., Lactobacillus spp., Roseburia spp., Faecalibacterium prausnitzii and Akkermansia muciniphila) and faecal short chain fatty acid concentrations. Lipid levels (serum total cholesterol, LDL, HDL, TG) and blood pressure will also be measured. Anthropometric measurements including BMI and waist circumference will be collected, in addition to comprehensive dietary and lifestyle data. Gastrointestinal side-effects will be assessed using a visual analogue scale adapted from Lecerf et al.[45]. The scale rates nine items including flatulence, bloating, rumbling, abdominal cramps, a global digestive tolerance score calculated from the five previous items, stool consistency, stool frequency and general well-being through disturbances in usual and professional activities as well as disturbances in personal and social activities. Items are scored on a 10 cm linear scale.

An experienced research dietitian will obtain a comprehensive dietary history from each study participant at baseline and at the completion of each intervention period. A dietary history is a structured interview method consisting of questions regarding habitual food intake. It involves a 24-hour diet recall followed by a food frequency questionnaire to clarify information about usual consumption over the previous month. Usual portion sizes will be obtained in household measures and with the use of photographic aids. The dietitian will review the dietary history with each participant, probing for details on portion sizes and cooking methods in order to improve data accuracy. Dietary AGE content will be estimated from an open source database which lists the AGE concentration of foods using validated analytical techniques [46]. Dietary macro and micronutrient intakes will be estimated using the Foodworks nutrient software program (Xyris Software, NSW, Australia). Each participant will also be randomly contacted by telephone during the study and asked to provide a 24-hour diet recall, and changes in urinary urea excretion will be monitored in order to validate the dietary history data collected [47]. Physical activity will be estimated by asking participants to complete the International Physical Activity Questionnaire (Short Form) prior to and at the completion of experimental and placebo intervention periods [48].

Body weight will be measured in participants wearing light clothes without shoes using a digital scale (Seca, Germany) to the nearest 0.1 kg. Height will be measured using a portable stadiometer (Seca, Germany) to the nearest 0.1 cm. BMI is calculated by dividing weight (kg) per height (m) square. Waist circumference will be measured at the midpoint of the lowest rib and iliac crest using a measuring tape to the nearest 0.1 cm. Body composition (total body water, fat mass, fat-free mass) will be determined using Bioelectrical Impedance Analysis (BodyStat-1500, Bodystat, Douglas, Isle of Man, United Kingdom). Ambulatory blood pressure will be measured using an electronic blood pressure machine (Omron Corporation, Kyoto, Japan), with subjects at rest in a seated position. All anthropometric measurements will be conducted in duplicate, with the mean measurement recorded.

Biological samples will be collected at baseline and at completion of each three-month supplement intervention period and stored at −80°C immediately after collection. Twenty ml of fasting venous blood will be collected from each participant by phlebotomy into a sodium fluoride EDTA tube, a heparin-lined vacuum tube and a clean glass test tube. Twenty-four hour urine collections and morning stool samples will be collected in sterile containers.

Serum, urinary and faecal CML will be measured using a competitive ELISA (AGE-CML ELISA, Microcoat, Penzberg, Germany) [49]. This assay has been validated [50], is specific, and shows no cross-reactivity with other compounds [49]. The within assay and between-assay coefficient of variation are both less than 5%, respectively. Methylglyoxal will be measured by HPLC. Serum total cholesterol and triglyceride concentrations will be determined by enzymatic colorimetric assay (Technicon Instruments, Ltd., New York, N.Y., USA), while HDL cholesterol will be determined enzymatically in the supernatant after precipitation of other lipoproteins with dextran sulphate-magnesium. LDL-cholesterol will be calculated using the Friedewald formula. Plasma glucose levels will be determined by using an automated glucose oxidase method (Glucose analyser 2, Beckman Instruments, Fullerton, California). Insulin will be measured by enzymatic colorimetry (WAKO Pure-Chemical Industries, Osaka, Japan). Insulin Resistance (IR) will be estimated by the homeostasis model assessment (HOMA) index as [FI × (fasting glucose/22.5)], where FI is insulin in microunits per millilitre and fasting glucose is in millimoles per litre [51, 52]. HbA1c will be measured by autoanalyser (Roche Diagnostics, Mannheim, Germany). Plasma IL-6, MCP-1, E-selectin, hsCRP, TLR4, glutathione (GSH), and myeloperoxidase will be measured by commercial ELISA kits (Biosource International, Camarillo, CA, USA). Urine 8-isoprostanes will be measured by ELISA (Oxford Biomedical Research, MI, USA).

Stool samples will be homogenised in a blender and stored at −20°C for SCFA analysis. Samples will be thawed and 5 g aliquots placed in Centriprep fluid concentrators, MWCO 30,000 kDa (Amicon Inc., Beverly, MA, USA). Samples will be centrifuged for 30 minutes at 1000 × g, room temperature and supernatants placed in 15 ml polypropylene tubes. 0.3 ml of 25% m-phosphoric acid will be added to each tube, samples will be vortexed and incubated at room temperature for 25 minutes. Samples will be centrifuged at 5000 × g for 15 minutes at room temperature. Supernatants will be decanted and frozen overnight. The following day, samples will be thawed and the pH of each sample adjusted to 6.5 using 4 N KOH. Oxalic acid will be added at a final concentration of 0.03% and SCFA concentrations determined by gas chromatography with use of a Hewlett-Packard 5880A gas chromatograph (Hewlett Packard, Palo Alto, CA, USA) containing an 80/120 Carbopack B-DA/4% Carbowax 20 M column (Supelco Inc., Bellefonte, PA, USA).

Quantitative Real-time PCR will be used to determine faecal concentrations of Bifidobacterium spp., Lactobacillus spp., Roseburia spp., Faecalbacterium prausnitzii and Akkermansia muciniphila. The primers used will be based on the following 16S rRNA gene sequences: Bifidobacterium spp: F-CTCCTGGAAACGGGTGG and R-GGTGTTCTTCCCGATATCTACA [53], Lactobacillus spp: F-AGCAGTAGGGAATCTTCCA and R-CACCGCTACACATGGAG [54], Roseburia spp: F-CGKACTAGAGTGTCGGAGG and R-GTCATCTAGAGTGTCGGAGG [55], Faecalbacterium prausnitzii: F-GGAGGAAGAAGGTCTTCGG and R-AATTCCGCCTACCTCTGCACT [56], and Akkermansia muciniphila: F-CAGCACGTGAAGGTGGGGAC and R-CCTTGCGGTTGGCTTCAGAT [57]. PCR amplification and detection will be achieved with an ABI 7300 Real-time PCR system (Applied Biosystems, Foster City, CA, USA) using Mighty Amp for Real-time (SYBR Plus) and Rox Reference Dye (Invitrogen, Carlsbad, CA, USA). Each assay will be performed in duplicate in the same run. The cycle threshold of each sample will then be compared with a standard curve (performed in duplicate) made by diluting genomic DNA (tenfold serial dilution). Prior to isolating the DNA, the cell counts will be determined in culture and expressed as “colony forming units” (CFU). Data will be expressed as log CFU/g of faeces.

Outcome analyses will be undertaken on an intention-to-treat basis. Data will be presented as means ± SD. The Kolmogorov-Smirnov goodness-of-fit test will be used to test for normal distribution, and data not normally distributed will be log-transformed. Correlation analyses will be performed using the Pearson correlation coefficient. Significance of changes during the study will be assessed by comparing change of means between placebo and prebiotic treatment periods by paired sample t-tests. Trial data will be analysed using a linear mixed model design based on repeated measures to account for fixed factors such as treatment sequence (inulin – placebo vs placebo - inulin) and treatment period (intervention 1 vs intervention 2) in addition to participants as a random factor. Significant differences will be defined as a value of P < 0.05 based on two-sided tests. Any differences in physical activity levels, anthropometry, energy or nutrient intake during the course of the trial will be identified using ANOVA. Gastrointestinal symptom data obtained by visual analogue scale will be analysed using the Wilcoxon signed rank test. Effect sizes including 95% confidence intervals will be calculated for all significant outcomes. Data analysis will be performed using SPSS 20.0 software (SPSS, Chicago, IL).

Kinetic studies have estimated that up to 30% of dietary AGEs consumed are intestinally absorbed [64]. Under circumstances of increased intestinal permeability, it is likely that greater quantities of dietary AGEs and their reactive dicarbonyl precursors may be able to gain entry into the systemic circulation. Elevated levels of circulating proinflammatory cytokines and ROS frequently observed in individuals with prediabetes are known to compromise tight junctions between cells, disrupting the integrity of the intestinal barrier and enabling the absorption of larger, potentially deleterious compounds [65]. Prebiotic fermentation products may reduce gastrointestinal permeability and as a result limit the absorption of exogenous AGEs.

The SCFAs acetate and butyrate are intricately involved in the maintenance of gut epithelial integrity. Acetate increases colonic blood flow and butyrate is the principal fuel for colonocytes, assisting to optimise epithelial cell health [66]. Butyrate reduces gastrointestinal permeability by enhancing the activation of the peroxisomal proliferator-activated receptor gamma (PPARgamma) gene, a nuclear receptor involved in the attenuation of inflammation in colonic epithelial cells [67]. Butyrate also upregulates the expression of mucin-associated genes important in maintaining the integrity of the intestinal mucosal barrier [68]. Oligofructose supplementation in mice has been shown to increase the expression of zonula and occludin, proteins important in the maintenance of tight junctions between gastrointestinal cells [69], and a dietary inulin intervention reduced markers of intestinal permeability in healthy adult males [70]. Butyrate is a histone deacetylase inhibitor and is likely to promote the transcription of these tight junction proteins. Prebiotic-induced changes in gut microbiota also increase endogenous production of Glucagon-like peptide 2 (GLP-2), which enhances gut barrier function by promoting the proliferation of crypt cells [71, 72].

Increased production of ROS stimulate endogenous AGE formation by oxidising glucose and unsaturated fatty acids to generate reactive dicarbonyls. Experimental drug treatments which attenuate oxidative stress have previously demonstrated reductions in serum AGE levels [73]. Cytokine production at sites of inflammation stimulate immune cell activation of NADPH oxidase (NOX) and production of myeloperoxidase, enzymes involved in the oxidation of amino acids to form AGE precursors [74]. Activated immune cells can also secrete the high-mobility group box 1 (HMGB-1) protein which is capable of binding to RAGE, thereby inducing further inflammation [75].

A high fat diet (independent of the level of obesity in the host) is associated with negative changes in bacterial communities within the colon [76]. In response to a high fat meal, bacterial lipopolysaccharide (LPS) translocates from the intestine into the host circulation, resulting in ‘metabolic endotoxemia’ [77, 78]. LPS is a major component of the outer membrane of gram negative bacteria, and is a potent activator of the mammalian immune system. LPS interaction with immune cells stimulates macrophage over-production of ROS, enhances the secretion of pro-inflammatory cytokines, and contributes to weight gain and development of insulin resistance [79, 80]. Individuals with type 2 diabetes have been found to possess endotoxemia levels 2-fold higher than people without diabetes [81]. Metabolic endotoxemia also positively correlates with total energy intake and fasting insulin levels in the general population [82]. In mice with high fat diet-induced metabolic endotoxemia, nutritional supplementation of the diet with prebiotics restored intestinal levels of gram positive bacteria (particularly Bifidobacterium species) and subsequently improved glucose tolerance and reduced circulating concentrations of LPS and pro-inflammatory cytokines [83]. Human trials involving dietary prebiotic supplementation have successfully reduced serum LPS levels [45] and markers of lipid peroxidation (a process which generates AGE precursors), possibly through the reduction of ROS production or the direct antioxidant ability of some Bifidobacterium and Lactobacillus bacterial species [84, 85].

SCFAs produced as a bacterial by-product of prebiotic fermentation are absorbed into the host circulation, effecting the expression of a wide range of genes in distal tissues associated with cell proliferation, differentiation and apoptosis. SCFAs are ligands for the G-protein coupled receptors (GPRs) GPR41 and GPR43 [86] on immune cells. These receptors are involved in down-regulating inappropriate immune cell production of pro-inflammatory cytokines, chemokines and ROS [87]. The SCFAs acetate, proprionate and butyrate exhibit a variety of anti-inflammatory actions through inhibition of NFκB activation, prevention of LPS-stimulated TNFα production in neutrophils and suppression of cytokine production [31, 88‐90].

Numerous other immune modulating effects have been observed secondary to gut bacterial activity including the production of anti-inflammatory compounds such as polysaccharide A [91], peptidoglycan [92] and conjugated linoleic acid [93], and the induction of T-regulatory cells [26].

An energy-restricted diet resulting in weight loss has been shown to reduce serum AGE levels in overweight and obese individuals [94]. The consumption of prebiotics in human clinical trials has promoted self-reported satiety [95], weight reduction, reduced production of the orexigenic hormone grehlin and stimulated expression of the appetite-reducing hormone peptide YY (PYY) [96]. Interaction with GPR41 by the SCFAs proprionate and butyrate increases satiety [97], upregulates PYY production and modulates the expression of leptin, a hormone important in controlling energy intake and expenditure [31].

In mice, the selective growth of certain lactobacillus species in the colon reduced body fat storage through the up-regulation of fiaf (fasting induced adipose factor) gene expression and inhibition of lipoprotein lipase (LPL) [98‐100]. These findings may have the potential to affect weight reduction in humans and subsequently reduce circulating AGE levels.

Inulin enhances the proliferation of lactic acid producing bacteria capable of synthesising B-group vitamins, some of which have an antioxidant capacity [101]. These vitamins can be utilised by the human host to neutralise ROS. Vitamins B1 and B6 trap the carbonyl groups of highly reactive AGE precursors before they can react with proteins [102]. Some Lactobacillus and Bifidobacterium species are efficient scavengers of the lipid peroxidation product malandialdehyde, protecting the host from excessive accumulation of this toxic AGE precursor [84]. Inulin also exhibits antioxidant properties independent of altering gut bacterial growth and is able to scavenge a number of ROS, which may help to reduce lipid peroxidation in the stomach [103].

The SCFA butyrate, produced as a bacterial by-product of inulin fermentation, has been shown to increase colonic glutathione production [104]. Glutathione is an antioxidant co-factor required for glyoxalase I activity, an enzyme which degrades the AGE precursor methylglyoxyl. Increased production of ROS is also thought to deplete glutathione levels. Through the reduction of oxidative stress, prebiotics may assist in the maintainance or upregulation of the glyoxylase pathway.

Both transient and chronic elevations in blood glucose increase endogenous AGE generation. Activation of GPR43 in adipocytes by proprionate inhibits lipolysis and lowers glycemia in healthy individuals [105]. Butyrate has been shown to reverse diet-induced insulin resistance in animal studies [106], possibly by enhancing PPARgamma expression which increases fatty acid oxidation in muscle.

Glucagon-like peptide 1 (GLP-1) is an incretin hormone released from intestinal L-cells in response to consumption of carbohydrates and fats. GLP-1 potentiates glucose-induced insulin secretion, reducing post-prandial blood glucose levels. GLP-1 also enhances satiety and slows gastric emptying. Prebiotic feeding in rats promoted L-cell differentiation in the colon and increased GLP-1 production [107, 108], probably through an increase in bacterial production of butyrate.

The consumption of inulin as a dietary supplement may also influence circulating AGE concentrations independently of its prebiotic function. High glycemic index (GI) diets and their resultant hyperglycemic effect have been shown to enhance AGE formation in healthy individuals [109]. Inulin is a soluble fibre which when consumed daily may play a role in the reduction of the GI of the diet. Many soluble fibres are known for their ability to delay gastric emptying and slow the rate of intestinal nutrient absorption, reducing the GI of the carbohydrates in the meal. Regular consumption of prebiotic soluble fibre reduces fasting and post-prandial serum glucose levels in people with impaired glucose tolerance [110] and type 2 diabetes [111].