IgG against specific bacterial antigens in obese patients with diabetes and in mice with diet-induced obesity and glucose intolerance

Abstract

OBJECTIVE

High fat diets increase the risk for insulin resistance by promoting inflammation. The cause of inflammation is unclear, but germfree mouse studies have implicated commensal gut bacteria. We tested whether diet-induced obesity, diabetes, and inflammation are associated with anti-bacterial IgG.

MATERIALS/METHODS

Blood from lean and obese healthy volunteers or obese patients with diabetes were analyzed by ELISA for IgG against extracts of potentially pathogenic and pro-biotic strains of Escherichia coli (LF-82 and Nissle), Bacteroides thetaiotaomicron, and Lactobacillus acidophilus, and for circulating Tumor Necrosis Factor α (TNFα). C57Bl/6 mice were fed low- or high- fat diets (10 or 60% kcal from fat) for 10 weeks and tested for anti-bacterial IgG, bodyweight, fasting glucose, and inflammation.

RESULTS

Obese diabetic patients had significantly more IgG against extracts of E. coli LF-82 compared with lean controls, whereas IgG against extracts of the other bacteria was unchanged. Circulating TNFα was elevated and correlated with IgG against the LF-82 extract. Mice fed high-fat diets had increased fasting glucose levels, elevated TNFα and neutrophils, and significantly more IgG against the LF-82 extracts.

CONCLUSIONS

Diabetes in obesity is characterized by increased IgG against specific bacterial antigens. Specific commensal bacteria may mediate inflammatory effects of high-fat diets.

Introduction

Insulin signaling is sensitive to inflammation [1], and inflammatory stimuli can induce glucose intolerance [2-3]. High fat diets are thought to initiate inflammation in expanding adipose tissues [4-6], presumably through direct activation of innate immune receptors by dietary saturated fatty acids [7-8]. However, germfree mice are resistant to diet-induced obesity and inflammation [9-12], suggesting that inflammation is dependent on interactions between the diet and commensal bacteria. High fat diets promote absorption of bacterial lipopolysaccharides (LPS) [2; 13], which can induce adipose tissue inflammation and insulin resistance [2-3], and also promote translocation of bacteria into visceral adipose tissue [14]. Alternatively, inflammation due to high-fat diet/bacteria interactions may originate in the intestine and subsequently cascade to surrounding visceral fat [9].

While there is mounting evidence for a direct role of the gut microbiome in diet-induced insulin resistance, this has not been conclusively demonstrated in obesity-associated diabetes. To shed more light onto this issue we tested whether diabetes in obesity is associated with inflammatory immune responses against specific gut bacteria.

Materials and Methods

Test subjects

Plasma was obtained from 32 obese individuals participating in a Health Management Resources (HMR®) weight loss clinic and from 10 healthy lean volunteers at Biospecialty Corp. (Colmar, PA, USA). Donor parameters are listed in Table 1. Half of the obese donors had diabetes and were being treated with insulin, insulinotropes, metformin, glyburide, PPARγ agonists, or combinations thereof. Nine obese subjects with diabetes and 5 obese controls were on statins, none were smokers. All samples were obtained with approval from relevant Institutional Review Boards and with informed written consent. Samples were stored at -86°C until use.

Table 1.

Relevant parameters of plasma donors

	Lean (n=10)	Obese (n=16)	Obese, diabetes (n=16)
Body-mass index	24.8 ± 3.0	41.1 ± 10.4	40.6 ± 7.1
Gender	1F, 9M	7F, 9M	7F, 9M
Age	42.3 ± 11.2	50.8 ± 17.0	59.2 ± 7.2

Anti-bacterial IgG and TNFα measurements

Total (free and soluble-receptor bound) TNFα was measured in 2x diluted human plasma with an ELISA from eBioscience (BMS223HS; sensitivity 0.13 pg/mL) and in mouse plasma with a multiplex ELISA (Millipore Corp.). To detect anti-bacterial IgG, we developed an ELISA as follows: Extracts of overnight cultures of Escherichia coli strains LF-82 (a pathogenic strain isolated from a patient with Crohn's disease [15]) and Nissle (a non-pathogenic strain), Bacteroides thetaiotaomicron, or Lactobacilus acidophilus grown in Lysogeny Broth, were prepared using a detergent-based bacterial protein extraction kit (“B-Per”; Pierce Biotechnology). The extracts likely contained a mix of lipid, protein, and sugar antigens from cytoplasm, membranes and cell walls. Extracts (10 μg protein / well) were coated onto 96 well flat-bottom ELISA plates (BD-Falcon) in carbonate buffer (pH 9.6). After blocking (“NAP” buffer; G-Biosciences), 400x dilutions of human or 100x dilutions of mouse plasma were added in triplicate, and bound IgG was detected with alkaline phosphatase-conjugated anti-human or mouse IgG (Fc specific) antibodies (Sigma-Aldrich). A chromogenic substrate (p-Nitrophenyl phosphate; Sigma-Aldrich) was added, and the color reaction was stopped with 3M sodium hydroxide. Absorbance at 450 nm (A450) was measured in a Bio-Rad microplate reader.

Mouse studies

Male C57Bl/6 mice, ordered at 5 weeks of age (Jackson Laboratories), were housed three per cage in a specific pathogen-free animal facility with a 12h light / dark cycle, and were used at 6 weeks of age. One group was fed a diet with 60% of kcal from fat (diet D12492 from Research Diets Inc.), the other a diet with 10% of kcal from fat (D12450B). The animals were euthanatized after 10 weeks, after a fasting (4h) blood glucose measurement (TrueTrack glucose meter; Home Diagnostics Inc). Plasma anti-bacterial IgG and TNFα were measured as described above; blood neutrophils with a Coulter counter (Perkin Elmer). All animals were handled in accordance with good animal practice as defined by the relevant national and local animal welfare bodies, and experiments were approved by the Institutional Animal Care and Use Committee.

Statistics

Results are expressed as mean ± S.E.M and were analyzed with GraphPad Prism v5.04. Groups were compared with unpaired Student's t-tests or Anova and Bonferroni's post-hoc analysis. Statistical significance was assumed when P<0.05.

Results

Increased IgG against extracts of E. coli LF82 in plasma of obese patients with diabetes correlates with TNFα

IgG against E. coli LF-82 extracts was lowest in lean subjects and highest in obese subjects with diabetes, with significant difference between obese diabetics and lean controls (Figure 1A; P<0.05). IgG against the other bacterial extracts (E. coli Nissle, B. thetoiotaomicron and L. acidophilus) was not different between groups. TNFα levels in obese diabetic patients were significantly higher than in obese controls (Figure 1B; P<0.05), and TNFα correlated with IgG against the LF-82 extract (Figure 1C; P<0.05).

Figure 1.

IgG against extracts of E. coli (strains LF-82 and Nissle), B. thetaiotaomicron, and L. acidophilus in plasma from lean and obese controls and obese patients with diabetes (A). Shown are A450 (average ± S.E.M) obtained with plasma from lean controls (“L”; n=10), obese controls (“O”; n=16) and obese patients with diabetes (“OD”; n=16). Values are normalized for those of lean controls. Panel B shows TNFα in the blood of obese controls and obese diabetic patients; panel C shows a positive correlation between IgG against extracts of E. coli LF-82 and TNFα. Asterisks indicate statistically significant differences between groups (T-test, P<0.05)

Increased IgG against extracts of E. coli LF-82 in plasma from mice fed high-fat diets

As expected, mice on the high fat diet gained more weight (Figure 2A; P<0.005) and had higher fasting glucose levels (Figure 2B; P<0.001), suggesting impaired glucose homeostasis. Mice on the high fat diet also had elevated neutrophil counts (Figure 2C; P<0.001) and circulating TNFα (Figure 2D; P<0.05), indicating systemic inflammation. They also had significantly higher IgG against the LF-82 extract (Figure 2E; P<0.05), but IgG against the other extracts was not significantly different.

Figure 2.

Body weight (A), fasting glucose (B), % neutrophils in the white-blood fraction (C), total TNFα (D), and anti-bacterial IgG (E) in blood and plasma from C57Bl/6 mice (n=6 per group) on low- or high- fat diets for 10 weeks. Shown are average ± S.E.M. Asterisks indicate statistically significant differences between groups (P<0.05).

Discussion

Our study made two novel observations. First, diet-induced obesity and glucose intolerance in mice was associated with increased IgG against antigens of pathogenic E. coli. Second, IgG against such extracts was significantly elevated in obese individuals with diabetes, but not without diabetes, and IgG correlated with TNFα. This would suggest that specific components of the intestinal microbiome can contribute to diet-induced metabolic inflammation and that profiling of IgG against bacterial antigens could help predict diabetes in obese subjects. However, it is unclear whether IgG responses are cause or consequence, and the identity of relevant bacteria and bacterial antigens remains unknown.

Recent studies have established a role for the gut microbiome in diet-induced metabolic inflammation of adipose tissue [9; 14]. However, while the composition of the gut microbiome changes during diet-induced obesity and insulin resistance [12; 16], it is unclear which species are responsible. We hypothesized that such species could be identified by analyzing cognate Immunoglobulin G. Indeed, IgG against extracts of E. coli LF-82 were increased in obese individuals whereas IgG against non-pathogenic E. coli or other bacteria was not elevated. However, it is not possible to conclude that it is the LF-82 strain against which IgG was directed. LF-82 was isolated from the intestine from one particular patient with Crohn's Disease [15] and it is unlikely that this particular strain is present in all humans, let alone in C57Bl/6 mice. Our extract likely contained several antigens that could be shared among various potentially pro-inflammatory strains or species and cross-react with IgG. We are currently attempting to determine the nature of these antigens.

Importantly, one could argue that increased anti-bacterial IgG simply reflects translocation, and there are indications for increased gut leakiness in obesity [17-18]. However, each engagement of translocated bacteria with cognate IgG has the potential to induce an inflammatory response through activation of FcγRIIa and other IgG receptors. Over time, such repeat inflammatory insults could set the stage for chronic inflammation and insulin resistance.

Abbreviations

ELISA

Enzyme Linked Immunosorbent Assay

IgG

Immunoglobulin G

LPS

Lipopolysaccharides