Changes in endotoxin levels in T2DM subjects on anti-diabetic therapies

Abstract

Introduction: Chronic low-grade inflammation is a significant factor in the development of obesity associated diabetes. This is supported by recent studies suggesting endotoxin, derived from gut flora, may be key to the development of inflammation by stimulating the secretion of an adverse cytokine profile from adipose tissue.

Aims: The study investigated the relationship between endotoxin and various metabolic parameters of diabetic patients to determine if anti-diabetic therapies exerted a significant effect on endotoxin levels and adipocytokine profiles.

Methods: Fasting blood samples were collected from consenting Saudi Arabian patients (BMI: 30.2 +/- (SD)5.6 kg/m2, n = 413), consisting of non-diabetics (ND: n = 67) and T2DM subjects (n = 346). The diabetics were divided into 5 subgroups based on their 1 year treatment regimes: diet-controlled (n = 36), metformin (n = 141), rosiglitazone (RSG: n = 22), a combined fixed dose of metformin/rosiglitazone (met/RSG n = 100) and insulin (n = 47). Lipid profiles, fasting plasma glucose, insulin, adiponectin, resistin, TNF-alpha, leptin, C-reactive protein (CRP) and endotoxin concentrations were determined.

Results: Regression analyses revealed significant correlations between endotoxin levels and triglycerides (R2 = 0.42; p < 0.0001); total cholesterol (R2 = 0.10; p < 0.001), glucose (R2 = 0.076; p < 0.001) and insulin (R2 = 0.032; p < 0.001) in T2DM subjects. Endotoxin showed a strong inverse correlation with HDL-cholesterol (R2 = 0.055; p < 0.001). Further, endotoxin levels were elevated in all of the treated diabetic subgroups compared with ND, with the RSG treated diabetics showing significantly lower endotoxin levels than all of the other treatment groups (ND: 4.2 +/- 1.7 EU/ml, RSG: 5.6 +/- 2.2 EU/ml). Both the met/RSG and RSG treated groups had significantly higher adiponectin levels than all the other groups, with the RSG group expressing the highest levels overall.

Conclusion: We conclude that sub-clinical inflammation in T2DM may, in part, be mediated by circulating endotoxin. Furthermore, that whilst the endotoxin and adipocytokine profiles of diabetic patients treated with different therapies were comparable, the RSG group demonstrated significant differences in both adiponectin and endotoxin levels. We confirm an association between endotoxin and serum insulin and triglycerides and an inverse relationship with HDL. Lower endotoxin and higher adiponectin in the groups treated with RSG may be related and indicate another mechanism for the effect of RSG on insulin sensitivity.

Figure 1

Correlations between log fasting endotoxin (EU/ml) and a) log triglycerides (mmol/l), b) HDL (mmol/l) and c) cholesterol (mmol/l) in the whole diabetic cohort. The lines of best fit are also shown: a) R² = 0.42, p < 0.0001, b) R² = 0.055, p < 0.001, c) R² = 0.1, p < 0.001).

Figure 2

Correlations between log fasting endotoxin (EU/ml) and a) log triglycerides (mmol/l), b) HDL (mmol/l) and c) cholesterol (mmol/l) in the non-diabetic (ND) group. The lines of best fit are also shown: a) R² = 0.192, p < 0.0001, b) R² = 0.007, not significant (NS); c) R² = 0.163, p < 0.001).

Figure 3

Correlations between log endotoxin (EU/ml) and a) log fasting insulin (ng/ml) and b) glucose (mmol/l) in the whole diabetic cohort. The lines of best fit are also shown: a) R² = 0.032, p < 0.001, b) R² = 0.076, p < 0.001. Correlations between log endotoxin (EU/ml) and c) log fasting insulin (ng/ml) and d) glucose (mmol/l) in the non-diabetic (ND) group. The lines of best fit are also shown: c) R² = 0.025, NS; d) R² = 0.009, NS.

Figure 4

The mean levels of endotoxin (EU/ml) in sera from T2DM patients, pre and post RSG treatment (n = 11, p < 0.05*).