In the present work we have found that the inflammatory enzyme COX-2 was highly expressed in internal mammary arteries from diabetic patients. Moreover, since our aim was to evaluate which serum biomarker(s) was (were) responsible for COX-2 expression, we performed a translational study in which we recruited a group of 116 atherosclerotic patients with or without DM as a risk factor. Data from this study showed that DP had increasing levels of glucose and Hb1ac, as expected, but also triglycerides (TG). From this, we studied the correlation of these markers with the vascular expression of COX-2 by a bivariate analysis and we found that TG levels correlated with the expression of COX-2.
Some authors have shown previously that high levels of glucose induced COX-2 expression in several cell types. High glucose-induced COX-2 expression in monocytes is regulated by NFk-B, protein kinase-C and p38 MAPK signalling pathways . Protein kinase-C has also been involved in COX-2 expression in rat vascular smooth muscle cells in culture . However, a comprehensive analysis of the intracellular mechanism underlying COX-2 expression in human vascular smooth muscle cells is out of the scope of this manuscript.
We evaluated in human vascular smooth muscle cells in culture the effect of different concentrations of glucose in the cellular medium. We found that glucose 25 mM slightly but significantly increased the expression of COX-2 in cells from non-DP. Although sporadically glucose levels in DP may reach the concentration of 25 mM, the glucose levels in blood from our diabetic patients were in a range of 8-10 mM. In fact, the bivariate analysis of glucose (or Hb1ac) versus COX-2 expression showed no correlation, which indicates that the expression of COX-2 in arteries from DP should be related with other factors. Another possibility was that the expression of COX-2 could be related to changes in the lipid profile of our patients. As shown in Figure 3, the levels of triglycerides positively correlated with the expression of COX-2. This fact indicates that it is the lipid profile rather than glycaemia which might regulate the expression of COX-2 in diabetic patients. This, along with the fact that levels of triglycerides were higher in diabetic patients, leads to the conclusion that elevated triglyceride levels in diabetic patients, but not glucose, is responsible for the overexpression of COX-2 in arteries from diabetic patients. To corroborate this, when human VSMC from non-DP were treated with different concentrations of triglycerides we found that, unlike glucose, treatment with triglycerides induced COX-2 expression in cells from ND patients.
Dislipemia, associated with high levels of triglycerides and low concentrations of HDL contributes to a proinflammatory state . In this context, the adipose tissue, in addition to storing calories as triglycerides, also secretes a large variety of pro-inflammatory proteins  which contributes to cardiovascular disorders.
The consequences of COX-2 expression in atherosclerotic vessels are not fully understood. The induction of COX-2 in vivo is generally associated with deletereous responses. However, in the presence of endothelial dysfunction (i.e. in diabetes) the local induction of COX-2 in the underlying smooth muscle cells may compensate for the reduced thrombo-resistance of that section of the vessel and may also compensate for the decrease in nitric oxide-dependent vasorelaxation observed in diabetic arteries . Thus, endogenous PGI2 release as a result of COX-2 expression is considered beneficial in the cardiovascular system since it decreases VSMC proliferation , cholesterol accumulation and platelet activation and increases vasodilation [21, 22]. Interestingly, PGI2 synthesis from human aorta samples decreases as a function of progressing atherosclerotic lesion, whereas PGE2 increases in parallel . PGE2 is a proatherogenic eicosanoid when released in advanced atherosclerotic plaques since it may induce the release of metalloproteinases (MMP) such MMP-2 and MMP-9, enzymes capable of degrading all macromolecular constituents of the extracellular matrix  and thus participate in atherothrombosis. In this work, we evaluated the release of basal PGI2 and PGE2 in human VSMC isolated from DP and non-DP. We found that, although the basal levels of PGE2 were similar in both groups of patients, the release of PGI2 decreased in cells from diabetic patients. Some postulations may be made about the surprising fact the fact that PGI2 levels in cells from diabetic patients was lower than in non-diabetic ones. One of the most outstanding candidates for PGI2 inhibition in the diabetic scenario is the peroxynitrite oxygen reactive form (ONOO-), which has been shown to perform a selective nitration PGI2 synthase in in vitro models of diabetes and therefore inhibit PGI2 synthesis . This increase in ONOO- is thought to take place by means of eNOS uncoupling in diabetes , which may be related to a decreased eNOS expression in the internal mammary arteries of diabetic patients who underwent by-pass surgery. Moreover, the analysis of COX-2 and MMP-9 indicated a correlation between these proteins. The latter might indicate that, persistent overexpression of COX-2 in diabetic patients might lead to a deleterious effect.
According to our results, it is intriguing to systematically assess plasma and urine levels of eicosanoids such as LTs and TXs, plasma and RBC membrane levels of antioxidants such as SOD, catalase and glutathione as well as plasma levels of NO in DP and non-DP undergoing CABG surgery. This could help to broaden our knowledge about how diabetes affects the balance among lipids, inflammation, eicosanoids, oxidative stress and subsequent endothelial funcition and will be the goal of coming research efforts.