Diabetes is a disorder associated with an increased risk of developing vascular and other health complications. Oxidative stress and inflammation are the major pathogenetic mechanisms considered to be implicated in these complications(Reference Jialal, Devaraj and Venugopal1, Reference Pickup2). Subjects with type 2 diabetes have been shown to have increased levels of lipid peroxidation, oxidative damage to DNA and protein oxidation(Reference Jialal, Devaraj and Venugopal1), presumably caused by an overproduction of free radicals and a decreased antioxidative defence. Enhanced production of free radicals is related to hyperglycaemia, insulin resistance and hyperinsulinaemia(Reference Ceriello3). High levels of glucose lead to an increased production of free radicals via different mechanisms such as glucose auto-oxidation and formation of advanced glycation end products(Reference Bonnefont-Rousselot4).
Besides oxidative stress, inflammation is also implicated in the development of complications in type 2 diabetes(Reference Pickup2). Cyclo-oxygenase catalyzed PG formation, and subsequently low-grade inflammation is suggested to be an early event in the development of type 2 diabetes that is further linked to oxidative stress(Reference Helmersson, Vessby and Larsson5). Elevated levels of high sensitivity C-reactive protein (hsCRP) and IL-6 are seen in subjects with type 2 diabetes(Reference Pickup, Mattock and Chusney6), and are also associated with an increased risk for developing the disease in future(Reference Duncan, Schmidt and Pankow7). However, whether inflammation and oxidative stress are related to glycaemic control is still not fully clarified.
The aim of the present observational study was to investigate the relationships between glycaemic control and levels of oxidative stress and inflammation in subjects with well-controlled type 2 diabetes.
Experimental methods
Subjects and study design
Participants were recruited to take part in an intervention study with antioxidant supplementation. Baseline results are described in the present article and results from the intervention study are presented elsewhere. Inclusion criteria were age 40–75 years, type 2 diabetes treated with either diet or diet and oral hypoglycaemic medication, glycated Hb (HbA1c) < 10 % and BMI < 35 kg/m2. Exclusion criteria were insulin-dependent diabetes, known CVD, acute inflammatory, liver, kidney or thyroid diseases as well as medication or supplementation that could affect oxidative or inflammatory status. Subjects gave their written consent to participate in the study. The study was approved by the Ethical Committee of the Medical Faculty at Uppsala University, Sweden.
Blood and urine sample were drawn in the morning after an overnight fast. Body height, weight, waist circumference (waist) and blood pressure were recorded at the same time.
Laboratory analysis
Blood glucose concentration was analyzed by an enzymatic technique (HemoCue). HbA1c was analyzed with high performance liquid chromatography. Plasma insulin was assayed with an enzymatic immunological assay (Mercodia, Uppsala, Sweden) in a Coda Automated EIA Analyzer (Bio-Rad Laboratories, Hercules, CA, USA). Serum cholesterol, HDL- cholesterol and TAG concentrations were analysed by enzymatic colorimetric methods (Thermo Electron Corporation, Vantaa, Finland) in a Konelab 20 Clinical Chemistry Analyzer (Thermo Electron Corporation). LDL- cholesterol was calculated according to Friedewald(Reference Friedewald, Levy and Fredrickson8).
Biomarkers of oxidative stress
Comet assay and 8-oxo-7,8-dihydro-2′-deoxyguanosine
A high-alkaline formamido pyrimidine glycosylase (FPG) version of the comet assay(Reference Gedik and Collins9) was used with some modifications. For the 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodG) analyses, DNA was isolated using a cold work-up procedure(Reference Hofer and Moller10) with small modifications, followed by enzymatic hydrolysis(Reference Nagy, Johansson and Zeisig11). The amount of 8-oxodG per undamaged 2′-deoxyguanosine was analysed using on-line electrochemical and uv detection after separation of the nucleosides with HPLC.
Frequency of micronucleated erythrocytes
A flow cytometry-based micronucleus assay in very young erythrocytes from humans, transferrin-positive reticulocytes(Reference Abramsson-Zetterberg, Durling and Yang-Wallentin12), was used. In this micronucleus assay, measuring the frequency of micronucleus-transferrin-positive reticulocytes (fMN-Trf-Ret), the very young erythrocytes were separated from the mature before analysis. The detection limit was approximately a difference of 0·2 of the background fMN-Trf-Ret. The background fMN-Trf-Ret was approximately 1‰.
Malondialdehyde
Plasma malondialdehyde (MDA) concentration was measured by HPCL and fluorescence detection as earlier described(Reference Ohrvall, Tengblad and Ekstrand13).
F2-isoprostanes
Free 8-iso-PGF2α, a major F2-isoprostane in urine, was analysed by a validated RIA developed by Basu(Reference Basu14). Levels were adjusted for urinary creatinine concentration.
Nitrotyrosine
Nitrotyrosine was assayed in plasma using a commercially available enzymatic immunological assay (Bioxytech, OxisResearch, Portland, OR, USA).
Biomarkers of inflammation
High-sensitivity C-reactive protein
High-sensitivity C-reactive protein measurement was performed in plasma by a latex-enhanced reagent (Dade Behring, Deerfield, IL, USA) with the use of a Behring BN ProSpec analyzer (Dade Behring).
Interleukin-6
IL-6 was analysed in plasma by a high-sensitivity ELISA kit (IL-6 HS, R&D Systems, Minneapolis, MN, USA). Samples and standards were pipetted in a microtitre plate coated with monoclonal antibody against IL-6. After incubation and washing enzyme substrate solution was pipetted and followed by anti-IL-6 antibody. The colour reaction was proportional to the bound IL-6.
Prostaglandin F2α-metabolite
Urinary 15-keto-dihydro-PGF2α, a major metabolite of primary PGF2α, was analysed, by a validated RIA developed by Basu(Reference Basu15). Levels were corrected for urinary creatinine concentration.
Statistical analysis
The statistical software JMP version 3.2 (SAS Institute, Cary, NC, USA) was used. Probability values < 0·05 were regarded as statistically significant. The unpaired t test or the Wilcoxon two-sample test was used to analyse sex differences. The correlation coefficients (Pearson's or Spearman's coefficients) were calculated when analysing correlations.
Results
Clinical characteristic
Fifty-six participants (thirty women and twenty-six men, forty-eight non-smokers and eight smokers) were included in the study. Significant differences between sexes were found for weight (men>women), cholesterol, HDL- and LDL-cholesterol (women>men), 8-iso-PGF2α and nitrotyrosine (women>men) and IL-6 (men>women). Smokers had a higher IL-6 level compared with non-smokers and almost every biomarker for oxidative stress and inflammation tended to be higher in smokers. Twenty-one persons were treated with diet only, thirty-five with the addition of anti-diabetic medication (sulfonylureas or other insulin-stimulating compounds or/and biguanides). Baseline characteristics are presented as means and standard deviations: age 62·3 (sd 7·0) years; BMI 28·3 (sd 3·8) kg/m2; weight 81·9 (sd 13·8) kg; waist 99·6 (sd 11·1) cm; systolic blood pressure 143 (sd 14) mmHg; diastolic blood pressure 78 (sd 9) mmHg; HbA1c 6·1 (sd 0·9)%; fasting blood glucose 7·8 (sd 2·3) mmol/l; insulin 69·3 (sd 40·2) pmol/l; TAG 1·7 (sd 0·9) mmol/l; LDL-cholesterol 2·9 (sd 0·9) mmol/l; HDL-cholesterol 1·1 (sd 0·2) mmol/l; cholesterol 4·7 (sd 1·0) mmol/l; 8-oxodG/106 dG 0·96 (sd 0·45); FPG-sites 30·2 (sd 15·1) % tail; fMN-Trf-Ret 0·90 (sd 0·50) ‰; MDA 0·68 (sd 0·08) μmol/l; 8-iso-PGF2α 0·19 (sd 0·09) nmol/mmol creatinine; nitrotyrosine 245 (sd 401) nmol/l; hsCRP 3·1(sd 4·2) mg/l; IL-6 2·5 (sd 2·2) ng/l; 15-keto-dihydro-PGF2α 0·24 (sd 0·10) nmol/mmol creatinine.
Relationships between glycaemic control and indicators of oxidative stress and inflammation
Urinary 8-iso-PGF2α was positively correlated to fasting blood glucose (r 0·33, P = 0·01) as well as HbA1c (r 0·29, P = 0·03). In addition, there was a positive correlation between HbA1c and nitrotyrosine (r 0·43, P = 0·01). A negative correlation was seen between HbA1c and MDA (r − 0·32, P = 0·017). Waist (r 0·37, P = 0·005), BMI (r 0·32, P = 0·016) and weight (r 0·36, P = 0·006) correlated positively with hsCRP. Correlations were found between 8-iso-PGF2α and MDA (r − 0·33, P = 0·012), hsCRP and IL-6 (r 0·61, P < 0·0001) and 8-iso-PGF2α and 15-keto-dihydro-PGF2α (r 0·27, P = 0·048).
Excluding smokers (n 8) from the correlation analyses decreased power of the study but did not change main results. No change in main results of correlation analyses was found when studying the subjects treated with diet only or treated with diet plus diabetic medication. An exception was a positive association between FPG-sites and blood glucose (r 0·7, P = 0·0004) in subjects treated only with diet.
Discussion
The study showed positive associations between glycaemic control (blood glucose and HbA1c) and urinary 8-iso-PGF2α and HbA1c and nitrotyrosine, demonstrating a significant biological relationship between glycaemic control and oxidative stress. It was also found that abdominal obesity and low-grade inflammation (hsCRP) were closely related to each other. Despite the inclusion of a considerable number of biomarkers of oxidative stress and inflammation, only a few associations were found among these markers.
To the best of our knowledge, no other study investigating subjects with type 2 diabetes has reported so many biomarkers of oxidative stress and inflammation simultaneously. Totally six different biomarkers for oxidative stress and three for inflammation were measured in this patient group in order to clarify the relationships between glycaemic control, oxidative stress and inflammation since the latter are two vital pathologies that are considered to be the integrated parts of the metabolic syndrome(Reference Pickup, Mattock and Chusney6, Reference Esposito, Ciotola and Schisano16).
Glycaemic control related to oxidative stress
The direct relationship between glycaemic control and oxidative stress found in the present study has also been shown elsewhere. One report described a highly significant correlation between blood glucose and urinary 8-iso-PGF2α and a reduction of 8-iso-PGF2α associated with improved glycaemic control(Reference Altomare, Vendemiale and Chicco17). However, there are also investigations not showing such association(Reference Helmersson, Vessby and Larsson5). These differences could be explained by various levels of glycaemic control in the patient groups. The negative correlation between HbA1c and MDA found in the present study was opposite to findings by Altomare et al., who observed a positive correlation in patients with type 2 diabetes(Reference Altomare, Vendemiale and Chicco17). MDA is generally considered as a less specific marker of oxidative stress than the most reliable indicator of oxidative stress, 8-iso-PGF2α (Reference Basu18).
As far as we know, the correlation shown between HbA1c and nitrotyrosine has not been shown elsewhere. However, the lack of such correlation has been reported previously by Ceriello et al. (Reference Ceriello, Mercuri and Quagliaro19), who at the same time observed a direct correlation between plasma glucose and nitrotyrosine.
In the present study, we found a lack of associations between glycaemic control and oxidative stress as measured by 8-oxodG and FPG-sites. Hinokio et al. (Reference Hinokio, Suzuki and Hirai20) reported a positive correlation between glycaemic control and 8-oxodG (in urine and blood mononuclear cells) in diabetic subjects, maybe due to the higher level of HbA1c compared with the present study. The positive relationship between glucose and FPG-sites found in the present study in subjects treated with diet only has also been seen by other investigators(Reference Dincer, Akcay and Alademir21). To our knowledge, the present study is the first one investigating the fMN-Trf-Ret in subjects with type 2 diabetes. Healthy subjects have earlier been examined(Reference Abramsson-Zetterberg and Zetterberg22) with the same method and showed a similar level of frequency as in the present study.
Glycaemic control and inflammation
No correlations were found between markers of glycaemic control and inflammation in the present study as supported by Pickup et al. (Reference Pickup, Chusney and Thomas23) but contradictory to observations by Ford(Reference Ford24). The association found between diabetes and CRP disappeared when adjusted for BMI and waist, indicating that inflammation is more related to obesity than features for diabetes(Reference Helmersson, Vessby and Larsson5).
The well-known relationship between obesity and low-grade inflammation was also observed in the present study. A highly significant correlation between hsCRP and obesity was found, which also has been shown by others in subjects with type 2 diabetes(Reference Ford24). A strong association between IL-6 and CRP was also observed in the present study. The positive correlation between the cyclo-oxygenase-mediated inflammatory marker PGF2α-metabolite and the oxidative stress biomarker 8-iso-PGF2α found in the present study has also been observed by others(Reference Helmersson, Vessby and Larsson5), showing a link between free radical generation and inflammatory response.
Limitations of study
A limitation of the present study was the absence of a healthy reference population. Comparisons regarding oxidative stress and inflammation therefore had to be made with other investigations but with care since study conditions and methodology could differ in many aspects, especially grade of glycaemic control and obesity. Furthermore, the present study did not measure the glucose tolerance or record the diabetes duration, parameters that could have impact on the investigated relationships between metabolic disorders, oxidative stress and inflammation in subject with diabetes.
In conclusion, the present study indicated that glycaemic status was associated with oxidative stress even in well-controlled diabetes subjects. Furthermore, inflammation was more related to abdominal obesity than to glycaemic control. The relatively small study group with a well-controlled type 2 diabetes and moderate obesity, giving a narrow range of HbA1c, blood glucose, waist and BMI, increased the probability that existing relationships may not have been detected. In spite of these limitations, interesting relationships were found.
A large number of biomarkers of oxidative stress and inflammation were investigated, but only a few associations were found between the markers. This could be due to the fact that none of these biomarkers biosynthesizes via similar pathways or simultaneously owing to their diverse nature and origin.
Acknowledgements
We acknowledge funding from Vinnova (the Swedish Governmental Agency for Innovation System) Semper AB and Procordia AB. Thanks to the staff at KPL (Centre for Human Studies of Foodstuffs) Siv Tengblad, Eva Sejby and Barbro Simu for excellent technical assistance and Rawya Mohsen and Lars Berglund for statistical analyses. E. R. was employed by financial supporter Semper AB (employment completed 2005-01-01). E. R. has no other conflicts of interest. All other authors have no conflicts of interest to declare. Data have been presented as an abstract at the 24th International symposium on Diabetes and Nutrition, Salerno, Italy, 2006. E. R., B. V., S. B., R. Å., C. J. and L. A.-Z. contributed to research and writing of manuscript. L. M. and A. S. contributed to research.
