Pilot study: alterations of intestinal microbiota in obese humans are not associated with colonic inflammation or disturbances of barrier function

Authors


Prof. M. Gotteland, Laboratory of Microbiology and Probiotics, Institute of Nutrition and Food Technology (INTA), University of Chile, El Libano 5524, Macul, Santiago, Chile.
E-mail: mgottela@inta.cl

Abstract

Aliment Pharmacol Ther 2010; 32: 1307–1314

Summary

Background  Obesity is associated with low-grade inflammation contributing to insulin-resistance. Gut barrier alterations, described in animal models of obesity, probably favour inflammation. This has not been hitherto described in obese humans.

Aim  To evaluate gut permeability in asymptomatic obese and its association with plasma (C-reactive protein (CRP), arachidonate/eicosapentaenoate ratio) and faecal (calprotectin and leptin) markers of inflammation and microbiota alterations.

Methods  A total of 13 obese (age: 33.9 ± 11.5 years; BMI: 35.9 ± 5.0 kg/m2) and 11 control subjects (age: 30.3 ± 8.1 years; BMI: 23.5 ± 2.4 kg/m2) were recruited. Gut permeability was assessed by the lactulose-mannitol-sucralose test, plasma fatty acids by gas chromatography, faecal calprotectin and leptin by Elisa and faecal microbiota by G+C profiling.

Results  C-reactive protein was increased in the obese subjects (P = 0.01), but neither the plasma arachidonate/eicosapentaenoate ratio, the faecal levels of calprotectin and leptin, nor the gut permeability were altered. The faecal microbiota was altered in the obese (P = 0.0002), with predominance of bacterial populations having a lower G+C content and decreased concentrations of high G+C populations.

Conclusions  Asymptomatic obese individuals with systemic low-grade inflammation do not have evidence of colonic inflammation or gut barrier alteration; however, the biodiversity of their intestinal microbiota is affected.

Introduction

Obesity is a growing public health problem worldwide including the developing countries, and represents a heavy economic burden for their respective societies.1, 2 Obesity is associated with a low-grade inflammation, which contributes to the development of insulin resistance, the metabolic syndrome, type-II diabetes mellitus and, in the long-term, cardiovascular diseases.3 Several factors appear to be implicated in this inflammatory process. The plasma fatty acid (FA) composition, for example, which is influenced by the amount and quality of the dietary fat, has been associated with the low-grade inflammation of overweight and obese individuals.4, 5 In particular, the arachidonic acid/eicosapentaenoic acid (ARA/EPA) ratio has been proposed as a biomarker of inflammation and in the development of inactivity-associated insulin resistance.6 Another participating factor is the increasing mass of adipose tissue, which is infiltrated by activated macrophages and releases inflammatory mediators that increase the hepatic expression and release of acute phase reactants such as C-reactive protein (CRP) into the circulation. In addition, the secretory profile of the adipocytes changes radically in obesity, with lower levels of adiponectin and higher levels of inflammatory mediators including leptin.3, 7

Leptin expression is increased in the colonic mucosa of patients with Crohn’s disease or ulcerative colitis;8 these findings have not been described in obese subjects, even though the expression of gastric leptin is increased in obese rats.9 In this respect, it remains unclear whether the low-grade systemic inflammation affects the gastrointestinal tract in the obese subjects.10 Faecal calprotectin, a marker of neutrophil infiltration in the colonic mucosa, has been shown to be higher in the obese.11 The increased concentrations of circulating pro-inflammatory cytokines as well as the infiltration of the colonic mucosa by neutrophils could alter the gut barrier function.12 Increased gut permeability was described in animal models of obesity in association with elevated endotoxemia and non-alcoholic steato-hepatitis, a complication of obesity;13 it remains unknown whether this also takes place in obese humans.

Finally, another factor associated with gut barrier dysfunction is the intestinal microbiota. Increased gut permeability has been detected in patients with digestive or systemic diseases in whom colonic microbiota disturbances have also been described.12, 14 It has been recently shown that the colonic microbiota of obese mice and humans is altered,15 being more efficient in extracting energy from foodstuffs.16 In addition, the microbiota may contribute to fat storage in the host through the inhibition of the Fasting-Induced Adipose Factor (Fiaf), a hormone capable of inhibiting the lipoprotein lipase in adipose tissue.17

Taking these into considerations, the aim of this study was to evaluate in obese subjects their gut barrier function and whether the eventual alterations in this parameter are related to the higher concentrations of circulating and faecal biomarkers of inflammation and to changes in the colonic microbiota.

Subjects and methods

Subjects

The study protocol was approved by the Ethics Committee for Research in Humans of INTA, University of Chile, in compliance with the Helsinki Declaration. Informed written consent was obtained from all subjects before their admission to the protocol. Asymptomatic, nonsmoking, obese (BMI>30 kg/m2) and normal weight volunteers (18.5–24.9 kg/m2) of either gender, and between 18 and 50 years of age were included in the study. Subjects with previous or current gastrointestinal pathologies, chronic pathologies such as diabetes, nephropathies or liver cirrhosis, or with a history of antibiotic or nonsteroidal anti-inflammatory drug use in the previous month, were excluded from the study.

Anthropometric data, clinical laboratory analysis and biological samples

Anthropometric data were obtained from all the subjects and their whole body composition was determined by air displacement plethysmography (Bod-Pod, Body Composition System; Life Measurement Instruments, Concord, CA, USA), as previously described.18 Blood samples were obtained after an overnight fast to determine biochemical and lipid profiles as well as ultrasensitive C-reactive protein (CRP). An aliquot of heparinized blood was centrifuged and the plasma was stored at −30 °C until processed; these samples were used to characterize total plasma concentrations of saturated, mono-unsaturated and polyunsaturated fatty acids and more particularly, the ARA and EPA concentrations as hallmarks of inflammation. A fresh stool sample was obtained from each subject and transported to the laboratory under anaerobic conditions. Part of this sample was stored at −30 °C for the determinations of the G+C profile of the bacterial DNA as a reflection of the microbiota biodiversity. A second aliquot was used to carry out an acid steatocrit to determine the approximate faecal fat excretion.19 The last part of the sample was lyophilized and stored at −30 °C until processed for determination of faecal calprotectin and leptin.

Determination of plasma fatty acids and ARA/EPA ratio

Total lipids from plasma were extracted and subsequently purified according to Bligh and Dyer;20 methyl ester derivatives of the total lipids were prepared according to Morrison and Smith, using tricosanoic acid (23:0) as an internal standard.21 Gas chromatography was performed using a Hewlett Packard 5890 series II Plus gas chromatograph equipped with a split injector and a flame ionization detector. Samples were injected on a 30-m DB-FFAP capillary column (J&W, Folsom, CA, USA), using hydrogen as carrier gas. The column temperature was increased from 140 to 240 °C at 2 °C/min and the detector temperature was 260 °C.

Evaluation of gut permeability

The evaluation of gut permeability was carried out as previously described,22 using lactulose and mannitol as markers of small intestine permeability and sucralose as marker of colonic permeability. After an overnight fast, the volunteers drank 200 mL of a solution containing 7.5 g lactulose, 2 g mannitol and 2 g sucralose; urine was collected for 5 h in a plastic container with 10 mL of 10% thymol in isopropanol and kept at 4 °C. The volume voided was measured and aliquots were frozen at −20 °C until processed. Sugar concentrations in the urine samples were determined with a Varian 3600 gas chromatograph equipped with a split/splitless injector and a flame ionization detector (Varian Instruments, San Fernando, CA, USA), using cellobiose and α-methyl-glucose (Sigma Chemical Co., St. Louis, MO, USA) as internal standards. Results were expressed in mg as the 5 h-urinary excretion of each sugar and as lactulose/mannitol and lactulose/sucralose ratios as biomarkers of intestinal and colonic permeability respectively.23

Determination of faecal calprotectin and leptin

Faecal calprotectin and leptin as markers of colonic inflammation were determined in lyophilized faecal samples by ELISA using commercial kits (Cell Sciences, Canton, MA, USA, and R & D, Minneapolis, MN, USA respectively). Both assays were performed in a blinded fashion, according to the manufacturers’ recommendations; the limits of sensibility were 1.6 ng/mL for calprotectin and 7.8 pg/mL for leptin. The intra- and inter-assay coefficients of variation for leptin were 3.2%. and 4.4% respectively, and 7% and 11.3% for calprotectin. Results were expressed per g of dry stool.

G+C profiling of the faecal microbiota

For the analysis of the profiles of faecal community structure based on percent of GC content, faecal samples were shipped under dry ice to Alimetrics (Espoo, Finland). Due to the high cost of the analysis, only random samples from six control and six obese subjects were analysed. Total bacterial DNA was isolated and purified from the samples, as previously described by Apajalahti et al.,24 and subsequently subjected to CsCl density gradient centrifugation in the presence of bisbenzimidazole for 72 h. The tubes were fractionated by gently pumping the gradients through UV flow cells, which recorded the total DNA content in each part of the tube representing a known density and the G+C content of the fractionated DNA. Determination of the percent GC content represented by each gradient fraction was carried out by regression analysis of the data obtained from gradients containing standard DNA samples with known percentages of GC (Clostridium perfringens, Escherichia coli, and Micrococcus lysodeikticus). The absorbance value of the percent GC profile curve was first calculated for each percentage from 20% to 80%. The integral of the percent GC profiles was then normalized to 100% and the relative abundance was calculated for each 5% increment of the percent of G+C.

Statistical analysis

Statistical analysis was carried out using the statistica Software Package (StatSoft, Tulsa, OK, USA). Anthropometric and biochemical data were expressed as means and standard deviation. As most of the other variables did not follow a normal distribution, results were expressed as median and interquartile range (IR); comparisons between groups were carried out by using the nonparametric Mann–Whitney U-test and correlations were determined by the Spearman rank test.

Results

Anthropometric and biochemical parameters

A total of 11 controls and 13 obese volunteers were recruited in this study. Anthropometric and biochemical characteristics of both groups are shown in Table 1. As expected, most of the variables evaluated were altered in the obese subjects; in particular, CRP concentrations were significantly higher compared with the controls (P = 0.03) and a mild correlation was observed between the plasmatic CRP and the percentage of body fat (r = 0.52). All volunteers completed a 24 h food intake recall; no differences in fat intake were observed between both groups except in one obese individual who had an abnormally high intake. The fatty acid profile in the diet was not different when comparing both groups with a surprisingly high proportion of mono- and polyunsaturated fatty acids. However, the plasma concentrations of total saturated, monounsaturated and polyunsaturated fatty acids (Mean ± s.d.) were significantly higher in the obese group than in the control group [1353 ± 84 μg/mL vs. 955 ± 91 μg/mL (P = 0.006), 998 ± 73 μg/mL vs. 660 ± 79 μg/mL (P = 0.004) and 1761 ± 99 μg/mL vs. 1230 ± 108 μg/mL (P = 0.0012)]. ARA [Median (IR)] in the normal subjects [186 (153–214 μg/mL)] did not differ from those of the obese [203 (191–234 μg/mL); P = 0.21]; similar results were observed for the concentrations of EPA [19.9 (12.9–37.6 μg/mL) vs. 35.5 (20.6–64.5 μg/mL) for the control and obese respectively; P = 0.23]. As a result, the plasmatic ARA/EPA ratio remained unchanged in the obese subjects compared with the control [11.0 (5.7–11.8) vs. 7.0 (4.4–11.2), respectively; P = 0.36]. Faecal fat excretion assessed by the acid steatocrit did not yield any differences between the groups (data not shown).

Table 1.   Anthropometric and biochemical characteristics of the normal weight and obese subjects
 Normal weight subjects (n = 11)Obese subjects (n = 13)P-value*
  1. (Mean ± s.d.).

  2. * Mann–Whitney U-test.

  3. † Chi-square test.

Female (%)54.553.9N.S.†
Age30.3 ± 8.133.9 ± 11.5N.S.
Weight (kg)63.1 ± 10.292.3 ± 12.30.0000
BMI (kg/m2)23.5 ± 2.435.9 ± 5.00.0000
Waist circumference (cm)78.7 ± 7.5112.5 ± 9.60.0000
Body fat (%)25.1 ± 7.348.9 ± 9.30.0000
Fat body mass (kg)15.6 ± 3.843.1 ± 11.20.0000
Lean body mass (kg)47.2 ± 11.354.9 ± 10.6N.S.
Uric acid (mg/dL)4.73 ± 1.954.62 ± 1.08N.S.
Glucose (mg/dL)86.3 ± 5.798.7 ± 18.30.04
Alkaline phosphatase (U/L)63.2 ± 21.679.8 ± 16.80.05
GOT (U/L)20.4 ± 3.624.2 ± 6.6N.S.(0.1)
Total cholesterol (mg/dL)175.6 ± 29.1190.6 ± 28.4N.S.
HDL-cholesterol (mg/dL)3.54 ± 1.024.72 ± 0.900.01
LDL (mg/dL)103.2 ± 26.5115.2 ± 27.1N.S.
HDL (mg/dL)52.4 ± 12.641.3 ± 8.00.02
VLDL (mg/dL)20.1 ± 12.634.1 ± 11.30.01
Triglycerides (mg/dL)101.1 ± 63.7170.7 ± 56.60.01
CS/TG2.16 ± 0.781.25 ± 0.470.002
C-reactive protein (mg/L)1.01 ± 0882.54 ± 2.070.03

Gut permeability and inflammation

Results of the gut permeability tests are shown in Table 2. The urinary excretion of lactulose and mannitol was not altered in the obese subjects and accordingly, no changes in the lactulose/mannitol ratio, a widely accepted marker of small intestine barrier function, were observed in these individuals compared with the normal-weight controls. The determination of the whole gut permeability (including the colon) by measuring urinary sucralose excretion did not show any changes, nor was the lactulose/sucralose ratio, a marker of distal gut damage, altered. Furthermore, the faecal concentration of calprotectin, which is a marker of colonic inflammation, was not altered in the obese group. Faecal leptin was detected in only three subjects (1 obese and 2 normal-weight; 16.5; 12 and 24.1 ng/g of dry stool respectively) and these results could not be statistically analysed.

Table 2.   Gut permeability and faecal calprotectin in normal-weight (control) and obese subjects
BiomarkersControl group (n = 11)Obese group (n = 13)P-value
  1. Results were expressed as Median (Interquartile range).

  2. P-values were from Mann–Whitney U-test.

Gut permeability
 Mannitol (mg)115 (83–186)151 (120–190)0.47
 Lactulose (mg)5.8 (5.5–7.3)5.3 (4.7–9.2)0.54
 Sucralose (mg)11.4 (6.9–17.3)13.9 (11.0–16.2)0.23
 Lactulose/Mannitol (%)0.011 (0.009–0.021)0.012 (0.009–0.018)0.70
 Lactulose/Sucralose (%)0.14 (0.11–0.19)0.13 (0.12–0.18)0.62
Faecal marker of inflammation
 Calprotectin (μg/g dry stool)69.5 (55.2–84.0)79.9 (62.8–89.0)0.41

GC profiling of the faecal microbiota

The fractionated bacterial DNAs obtained after a 72-h CsCl density gradient centrifugation were analysed for their percentage of G+C content through UV detection. Figure 1 shows the G+C content profiles in both groups of subjects, while the lower panel shows the continuous curve of P-values obtained by running the t-test for normalized absorbance values at each G+C percent. The results clearly show that both groups displayed significantly different microbial communities in their stool samples. Compared with the G+C profile of normal weight subjects, those from the obese showed a significant increase in the relative abundance of bacterial populations with 23–37% G+C content in their DNA and a significant decrease in the relative abundance of those with 40–47% and 57–61% of G+C content. The most dominating faecal bacteria displayed a G+C content in their DNA of 41.7 ± 1.4% and 36.2 ± 1.0% in the normal and obese subjects respectively (P = 0.0021). Interestingly, the G+C peak values correlated negatively with the CRP levels (r = −0.68; P < 0.02).

Figure 1.

 Mean percentage G+C profiles of bacterial communities in stool samples from normal weight (blue line) and obese (red line) subjects. Bars in profiles are standard errors of the mean. Lower panel: Significance of difference between both groups. Y-axis shows the P-value of difference between the groups at each %G+C value. The solid black line represents the P-value of 0.05.

Discussion

This study evaluates whether a low-grade systemic inflammation is associated with gut inflammation and alterations of gut barrier function and faecal microbiota in obese individuals. The low-grade inflammation in our obese subjects was confirmed by their higher plasma CRP, which correlates positively with their body fat mass, as previously observed.25 The ARA/EPA ratio is another marker of inflammation because ARA is associated with the production of pro-inflammatory eicosanoids, while EPA exerts mainly anti-inflammatory activities.26 It has also been that in human volunteers, this ratio correlates with the HOMA index of insulin resistance.6 On the other hand, EPA prevents the disruption of the gut barrier function induced by proinflammatory cytokines27 and the development of intestinal inflammation in rodents.28 In addition, the proportion of plasmatic EPA as a fraction of total fatty acids is lower in obese.26 However, our results show that the ARA/EPA ratio remained unchanged in the obese subjects. The cause of these discrepancies with the other studies remains unclear, but it is possible that they are due to differences in the quality of the dietary fat.

On the other hand, Poullis et al. observed that faecal concentrations of calprotectin are significantly increased in obese subjects.11 However, our results did not reveal any increases in this marker in our obese subjects. The discrepancy may be due to the fact that the subjects of Poullis were much older than ours. In addition, we measured faecal leptin as another marker of colonic inflammation; in fact, leptin expression is higher in the inflamed colon from patients with inflammatory bowel diseases than in healthy controls,8 and luminal leptin has been involved in NF-κB activation in intestinal epithelial cells and in the development of cell damage and mucosal neutrophil infiltration in mice.8 However, leptin was undetectable in most of our subjects, both obese and controls; this may be due to either the absence of leptin release by the colonocytes or its degradation or utilization by the microbiota. Considered together, these results suggest that there are no detectable inflammatory processes in the colonic mucosa of young, otherwise healthy, obese adults.

Another aim of this study was to evaluate the gut barrier function, as its alterations contribute to a number of pathologies, including steathohepatitis.10, 12 Evaluations of gut barrier function in obesity are scarce. An increase in the small intestine permeability was observed ex vivo in genetically obese mice, accompanied by a redistribution of some tight junction proteins in the epithelium.13 However, more recent studies indicate that the intake of high fat diet (HFD) rather than obesity is important in causing gut barrier alterations. In fact, the intestinal permeability and tight-junction integrity are affected in rats fed a HFD (70%) for 4 weeks, and these deleterious events are prevented by the administration of antibiotics or the prebiotic oligofructose.29, 30 Using obese OLETF rats and their lean counterpart LETO, Suzuki and Hara observed that their intestinal permeability was comparable at baseline, but that it increased in both of them after feeding the HFD for 16 weeks, in association with a lower expression of some tight-junction proteins in the small intestine, but not in the caecum or the colon.31 To our knowledge, our results are the first evaluation of gut permeability in obese humans. In contrast with the animal study by Brun et al., we did not detect any evidence of alterations of intestinal or colonic barrier function. It could be argued that the low sample size may limit the interpretation of our results; nevertheless, not even a trend was observed (P = 0.70 and P = 0.62 respectively). In addition, our results coincide with the absence of changes in faecal calprotectin and leptin in these subjects. In relation with the eventual influence of dietary fat on this parameter, it is important to note that the higher plasma fatty acids concentrations observed in the obese group, even if they reflect in some extent their dietary fat intake, may also depend on the total lipid body stores and in consequence, it is difficult to conclude to what extent their fat intake influences gut barrier function. These results suggest that the alterations of gut permeability in obesity may represent a late event in older subjects.

Finally, we also compared the microbiota of our subjects by determining the profile of the G+C percent content of their faecal bacterial DNA. This method reflects the biodiversity of the microbiota, taking into account that the G+C DNA content is specific for each bacterial genus or cluster. It is important to note that by using faecal samples, we could only evaluate the biodiversity of the intraluminal bacteria and not of those associated with the colonic mucosa. Although this method has been previously applied to studies in humans,32–34 this is the first report using it in obese subjects. Our results clearly indicate that the bacterial communities differed in both groups of individuals; however, the experimental design used in the study does not allow to determine whether these alterations occur early in the disease and may be a factor implied in the development of obesity or whether they are secondary to the disease itself. The increase in the 23–37% G+C content range observed in the obese is consistent with higher numbers of bacteria belonging to low G+C Clostridium clusters and/or to some species of Lactobacillus, whereas the decrease in the 40–47% range could represent lower concentrations of bacteria belonging to Clostridial cluster XIVab (e.g. C. coccoides), to the Ruminococcus family or to Bacteroides. In addition, the significant decrease in the higher G+C region (57–61% range) observed in the obese subjects is consistent with a lower presence of Actinobacteria including Bifidobacteroidaceae, the Clostridial cluster IV or the Faecalibacterium prausnitzii genus. From these results, studies evaluating the biodiversity of the colonic microbiota in obese subjects have yielded contrasting results. Ley et al. observed a higher Firmicutes/Bacteroidetes ratio in obese humans15 and its normalization after 52 weeks of a hypocaloric diet.35 Corroborating these results, caloric restriction and increased physical activity for 10 weeks decreased the BMI in overweight adolescents in whom it was associated with increases of the B. fragilis and the Lactobacillus groups and lower counts of the C. coccoides group and some Bifidobacterium in their microbiota.36 A similar reduction of Bacteroides in obese subjects was observed by Armougom et al.37 In contrast, Schwiertz et al. observed a decrease in the Firmicutes/Bacteroidetes ratio in overweight and obese subjects, together with lower counts of the C. leptum group and of the Bifidobacterium genus.38 Such discrepancies could be due to the methodologies used for the analysis of the microbiota or to differences in the characteristics of the subjects or of their dietary habits, particularly its fat content. In relation to this point, Hildebrandt et al. evaluated the influence of diet, genotype and obesity on the microbiome composition by using knockout mice for RELM-β (Resistin-Like Molecule-β); these animals remain lean when fed a HFD, contrary to their wild-type control.39 They observed microbiome alterations when these animals were switched to the HFD, independently of their obese or lean genotype, with a decrease in the Bacteroidetes and an increase in both the Firmicutes and Proteobacteria. Accordingly, the intake of a HFD, rather than as a factor generating obesity per se, could be important in driving changes in the intestinal microbiota. As already mentioned, it is difficult to conclude in the obese group whether they had a higher dietary fat intake and whether this could modify their colonic microbiota. Finally, we observed an inverse correlation between CRP concentrations and G+C abundance, suggesting that bacterial populations with high DNA GC contents may modulate inflammatory processes in the host. These observations confirm those of Sokol et al. who detected low concentrations of F. prausnitzii in patients with Crohn’s disease, a microorganism that exhibits anti-inflammatory effects through its capacity to inhibit NF-kappaB activation.40

In conclusion, young asymptomatic obese subjects with slightly elevated CRP as a parameter of systemic, low-grade inflammation, do not show changes in their EPA/ARA ratio or evidence of colonic inflammation or alterations in their gut barrier function. However, the biodiversity of their intestinal microbiota is affected. These results have to be interpreted taking into account that this is a pilot study carried out in a low number of subjects.

Acknowledgements

Declaration of personal interests: The authors thank the volunteers for their cooperation in the study. Declaration of funding interests: This study was supported by Grant 1080519 from Fondecyt-Conicyt, Chile.

Ancillary