Vascular Biology of Atherosclerosis in Patients with Diabetes: Hyperglycemia, Insulin Resistance, and Hyperinsulinemia

Polyol Pathway

The two major enzymes of the polyol pathway include aldose reductase (AR), the first and rate-limiting enzyme of this pathway, and sorbitol dehydrogenase (SDH). By the action of these enzymes, glucose is metabolized to sorbitol and fructose, respectively. In the process, as shown later, AR action results in the conversion of nicotinamide adenine dinucleotide phosphate, reduced form (NADPH) to nicotinamide adenine dinucleotide phosphate (NADP ⁺ ) and the action of SDH consumes nicotinamide adenine dinucleotide (NAD ⁺ ) to yield NADH.

Compared with human or rat tissues, in the mouse the levels of AR are significantly lower; hence, a strategy to specifically test the role of AR in atherosclerosis used transgenic mice, which expressed human-relevant levels of AR on the major histocompatibility type 1 promoter, thereby exerting global overexpression of the enzyme. When these mice were bred with mice deficient in the low-density lipoprotein (LDL) receptor and made diabetic with streptozotocin, a significant increase in atherosclerosis, both by percentage of aortic arch lesion area and by en face analysis of the entire aorta, resulted after 6 weeks of a high-cholesterol diet, without a change in total cholesterol or triglyceride or in levels of very low-density lipoprotein cholesterol (VLDL-C), LDL-C, or high-density lipoprotein cholesterol (HDL-C) in the two groups of diabetic mice (those overexpressing or not transgenic for human AR [hAR]) ( Fig. 9-1 ). Similar roles for hAR in acceleration of atherosclerosis in diabetic LDL receptor null mice fed a high-cholesterol diet at 8 or 12 weeks were found, and when mice were fed a cholic acid–containing diet, atherosclerosis increased in the diabetic tg hAR animals versus the nondiabetic LDL receptor null mice. Of note, there were no differences observed in nondiabetic mice overexpressing transgenic hAR or not in the LDL receptor null background, thereby suggesting that glucose flux via the polyol pathway was specific to the diabetic state in atherosclerosis. In parallel with increased atherosclerosis in the diabetic transgenic hAR-overexpressing mouse, macrophages retrieved from these animals revealed increased expression of inflammatory mediators and greater uptake of modified lipoproteins. In addition to these findings in mice, Gleissner and colleagues discovered increased expression and activity of AR in human monocyte-derived macrophages during foam cell formation stimulated by oxidized LDL (oxLDL), a process that was further exacerbated when macrophages were grown in hyperglycemic conditions (30 mM d -glucose) compared with osmotic control conditions.

Recent studies by Vedantham and colleagues demonstrated that when tg hAR mice were bred into the apoE null background and rendered diabetic with streptozotocin, increased atherosclerosis ensued compared with the nontransgenic diabetic apoE null mice (see Fig. 9-1A, B ). As in the case of LDL receptor null mice, there was no effect of transgenic hAR expression in the nondiabetic apoE null mice. Furthermore, they showed that administration of an AR inhibitor, zopolrestat, was effective in reducing accelerated atherosclerosis in the diabetic transgenic hAR mice in the apoE null background (see Fig. 9-1C ). Important roles for endothelial cell hAR in diabetic atherosclerosis were demonstrated in that work. When tg Tie2-hAR mice in the apolipoprotein (apo) E null background were rendered diabetic with streptozotocin, atherosclerotic lesion size was increased, suggesting that endothelial cell AR contributes importantly to acceleration of atherosclerosis in diabetes.

It is important to note that an earlier study suggested that distinct inhibitors of AR (ARIs; tolrestat and sorbinil) and genetic ablation of AR in diabetic apoE null mice increased early lesion formation as a result of increased levels of toxic aldehydes in the lipid particles. Differences in the mouse models, specifically genetic overexpression of AR to human relevant levels or complete genetic deletion, as well as potential distinct off-target effects of the different ARIs may underlie these findings. In human patients with diabetes and neuropathy, however, 1 year of treatment with the ARI zopolrestat resulted in improved cardiac function, not worsened function, as measured by echocardiography. In contrast, the vehicle-treated diabetic patients continued to display reduction in their cardiac function. This study, which did not directly address diabetic atherosclerosis, did however suggest that pharmacologic inhibition of AR by zopolrestat did not worsen cardiovascular complications of diabetes. Hence, it is possible that more potent ARIs with less off-target effects may hold promise for the treatment of atherosclerosis in diabetes.

Finally, it is important to note that increased oxidative stress may result from the overactivity of the polyol pathway. NADPH is a cofactor of glutathione production; consumption of glutathione by action of the polyol pathway may result in reduced availability of this antioxidant mechanism. These considerations are consistent with the observations in mouse models and human macrophages that AR activity increases oxidative stress on high glucose and oxLDL exposure.

Hexosamine Pathway

When excess levels of glucose are shunted into the hexosamine biosynthetic pathway (HBP), products emerge that have been shown to cause endoplasmic reticulum stress and to alter transcriptional activity of key molecules implicated in atherosclerosis. In this pathway, fructose-6-phosphate is converted to glucosamine-6-phosphate and uridine diphosphate (UDP)– N -acetyl glucosamine via the actions of the rate-limiting enzyme of the hexosamine pathway, l -glutamine: d -fructose-6-phosphate amidotransferase (GFAT).

Examples of how the HBP may contribute to conditions that exacerbate atherosclerosis in diabetes include the following. First, in a manner dependent on mitochondrial superoxide production, hyperglycemia increases hexosamine biosynthesis and O -glycosylation of the transcription factor Sp1 in bovine aortic endothelial cells. Consequences of increased modification of Sp1 include increased expression of plasminogen activator inhibitor type 1 (PAI-1) and transforming growth factor beta 1 (TGF-β1). Second, findings similar to the effects of high glucose and the HBP on PAI-1 expression were also shown in adipose tissue. Third, in bovine aortic endothelial cells, endothelial nitric oxide synthase (eNOS) activity was inhibited by HBP-mediated increases in O -linked N -acetylglucosamine modification of eNOS and a decrease in O -linked serine phosphorylation at residue 1177. In the aortas of diabetic mice, similar changes in eNOS activity and these post-translational modifications were also observed. Because reduced eNOS activity is observed in diabetes and linked to endothelial dysfunction, HBP-mediated reductions in eNOS activity spurred by hyperglycemia may contribute to endothelial cell dysfunction, which presages accelerated atherosclerosis.

Protein Kinase C

Hyperglycemia stimulates the generation of diacylglycerol (DAG), which is an activator of at least certain isoforms of protein kinase C (PKC). The PKC family of enzymes consists of at least 12 members. PKCs are involved in a diverse array of cellular functions, many of which may be considered to play roles in diabetic atherosclerosis, such as cellular proliferation, signal transduction, cellular fate, and transcription factor modulation (e.g., Egr1, NF-κB, and Sp1), cytokine expression, and oxidative stress in cells such as endothelial cells, smooth muscle cells, and monocytes and macrophages, all of which contribute to atherosclerosis mechanisms.

In atherosclerosis, isoforms of PKC have been implicated in the pathogenesis of this disorder. First, work by Harja and colleagues showed that global deletion of the PKCβ isoform resulted in significant reduction in atherosclerosis in apoE null mice, even without diabetes. In parallel, these researchers showed that a chief mechanism by which deletion of this PKC isoform was protective was by reduction in the vascular expression of the key transcription factor, Egr1. Egr1, previously shown to influence proinflammatory and prothrombotic genes in atherosclerosis, is regulated by PKCβ. Furthermore, treatment of the apoE null mice with the PKCβ inhibitor LY333531 (or ruboxistaurin) resulted in decreased atherosclerosis. This work, although not performed in diabetic animals, nevertheless may suggest that this PKC isoform may play key roles in diabetic atherosclerosis. Supportive of this conclusion is the report showing that administration of ruboxistaurin to type 2 diabetic patients improved brachial artery flow-mediated dilation compared with vehicle treatment. In addition to PKCβ, possible protective roles for PKCδ in atherosclerosis have been suggested, particularly in smooth muscle cell survival. In a model of vein graft atherosclerosis in nondiabetic mice, deletion of PKCδ resulted in more severe atherosclerosis. As in the case of PKCβ, further studies are essential to determine potential implications in diabetes.

Other studies have suggested that advanced glycation endproduct (AGE) pathways may contribute to activation of PKC isoforms—for example, studies reported in bovine retinal endothelial cells.