Diabetes is associated with an increased prevalence of hyperlipidemia, hypertension, obesity, and a hypercoagulable state, all of which contribute to higher incidence of coronary, carotid, and peripheral artery diseases as well as associated increases in mortality and morbidity. Some risk factors, like lower HDL as well as hyperglycemia itself, have been independently associated with rapid plaque progression, as seen on imaging studies on patients with stable CAD. Diabetic atherosclerotic lesions include pathologic characteristics associated with atherosclerosis and atherosclerotic complications in general as well as ones more linked specifically to diabetes.

Moreno et al. evaluated coronary atherectomy specimens obtained from 47 patients with T2DM, comparing them with specimens from 48 nondiabetic individuals, and demonstrated that patients with diabetes exhibited a larger content of lipid-rich atheroma (7% ± 2%) than those without diabetes (2 ± 1%). In addition, macrophage infiltration was significantly greater in patients with diabetes (22 ± 3%) than in those without diabetes (12 ± 1%) while the incidence of thrombus was higher in individuals with diabetes (62%) versus those without diabetes (40%). Cipollone and colleagues examined carotid endarterectomy specimens from patients with and without T2DM (n = 30/group) and found that plaques from those with diabetes were richer in macrophages and T lymphocytes while manifesting higher human leukocyte antigen–DR (HLA-DR) expression. Immunohistochemistry revealed greater reactivity of the receptor for advanced glycation endproducts (RAGEs; discussed further below) in patients with diabetes versus those without diabetes, especially in macrophage-rich and angiogenic areas. In the diabetes samples, activity of the master proinflammatory transcription factor nuclear factor kappa B (NF-κB) was increased, which corresponded to RAGE expression. Also, cyclooxygenase 2 (COX-2) membrane-associated protein eicosanoid and glutathione metabolism synthase 1 (mPGES-1), matrix metalloproteinases (MMPs), and gelatinolytic activity, activities associated with plaque destabilization, were increased in patients with diabetes, who also had reduced collagen content and increased lipid and oxidized LDL content, all versus those without diabetes. In this study, RAGE, COX-2/mPGES-1, and MMP expression correlated linearly with plasma hemoglobin A1c (HbA1c) levels. This data supports diabetic atherosclerosis as being characterized pathologically by pathways involving enhanced inflammation, RAGE, macrophage infiltration, and plaque destabilization (see Fig. 7.1 ); other work suggests distinct pathologic characteristics of atherosclerosis in diabetes when specific risk issues are present, as reported with microcalcifications in patients with diabetes, albuminuria, but not yet clinical CV disease. These pathogenic signatures may represent targets whose modification would improve clinical outcomes. For example, statins may decrease levels of RAGE, as seen in patients with T2DM and asymptomatic carotid artery stenosis randomized to diet plus simvastatin (40 mg/day) versus diet alone for 4 months before endarterectomy. Plaques from the simvastatin group showed significant decreases in immunoreactivity for RAGE and advanced glycation endproducts (AGEs) as well as other proinflammatory and proatherosclerotic signals, including myeloperoxidase (MPO); p65; COX-2; mPGES-1; MMP-2; MMP-9; oxidized LDL; and the number of macrophages, T lymphocytes, and HLA-DR–positive cells. In plaque-derived macrophages, simvastatin-mediated decreases in RAGE levels were reversed by AGE stimulation. Thus statin benefits on diabetic atherosclerosis may include reduced RAGE expression and MPO-mediated AGE generation, fostering plaque stabilization.

Additional insights into the pathology of diabetic atherosclerosis derive from autopsy studies. An analysis of morphologic findings in patients with T1DM and T2DM versus age- and sex-matched individuals was performed in a cohort who died suddenly from CAD. A total of 16 patients with T1DM and 50 with T2DM were compared to 66 age- and sex-matched individuals without diabetes who also died from severe CAD ( Table 7.1 ). While the prevalence of smoking and hypertension were similar among all groups, the body mass index (BMI) in individuals with T2DM (30.5 ± 7.4 kg/m ² ) was significantly greater than in those without diabetes (26.6 ± 5.4 kg/m ² , P = 0.001) or T1DM BMI (25.6 ± 6.4 kg/m ² , P = 0.7). The ratio of total cholesterol (TC) to HDL-C was significantly higher in individuals with T2DM than in those without diabetes (7.9 ± 3.9 vs. 6.3 ± 3.4, P = 0.02), with trends for higher TC and lower HDL-C, as often encountered in T2DM. On the contrary, individuals with T1DM had a trend toward lower levels of TC (183 ± 52 mg/dL) and comparable levels of HDL-C (37 ± 14 mg/dL) and TC-to–HDL-C ratio (5.8 ± 2.9) relative to those without diabetes. The percent necrotic core area (necrotic core area divided by plaque area) was greater in individuals with T1DM (12.0% ± 5.7%) and T2DM (11.6% ± 8.4%) than those without diabetes (9.4% ± 9.3%; P = 0.05 vs. T1DM, P = 0.004 vs. T2DM) ( Table 7.2 ). Those with T2DM had greater percent calcified area (12.1% ± 11.2%) than those without diabetes (11.4% ± 13.5%, P = 0.05) while those with T1DM and those without diabetes had a comparable percent calcified area (7.8% ± 9.1%). The number of fibroatheromas was greater in individuals with T2DM (8.8 ± 4.3) than in those without diabetes (6.9 ± 4.7, P = 0.02) or with T1DM (7.1 ± 5.0); healed plaque ruptures but not thin-cap fibroatheromas were also higher in T2DM versus those without diabetes.

By multivariable analysis ( Table 7.3 ), a positive correlation was found between mean percent necrotic core area and glycohemoglobin, independent of HDL-C, ratio of TC to HDL-C, age, smoking, and sex (T = 2.8, P = 0.005). Similarly, the ratio of TC to HDL-C (T = 2.5, P = 0.01) and BMI (T = 3.5, P = 0.006) correlated positively with percent necrotic core area. There was a significant relationship between numbers of fibroatheroma and ratio of TC to HDL-C (T = 3.0, P = 0.0003). Glycohemoglobin correlated positively with number of fibroatheromas, although not with statistical significance (T = 1.7, P = 0.09).

Observations within this cohort also pointed to inflammatory mechanisms in diabetes. Macrophage plaque area and T-cell infiltration were significantly greater in individuals with diabetes versus those without diabetes ( P = 0.03, see Table 7.2 ; Fig. 7.3 ). T-cell infiltration was greater in individuals with T1DM, potentially consistent with T1DM as an autoimmune disease with a common genetic susceptibility to other disorders, like autoimmune thyroiditis, of potential pathophysiological significance in coronary plaque pathology. The incidence of acute thrombi was significantly less in individuals with T1DM (21%) than in those without diabetes (51%, P = 0.03) in sudden coronary death victims (see Table 7.2 ). Individuals with T1DM showed a lower incidence of acute plaque rupture than those without diabetes (6% vs. 27%, P = 0.09) while plaque erosion was significantly less frequent in individuals with diabetes versus those without diabetes (6%, 12% vs. 29%, P = 0.02 and P = 0.04). Acute thrombi were found less in specimens from individuals with T2DM than in those without diabetes, whereas stable severe CAD and chronic total occlusion were more frequently observed in individuals with T2DM than in those without diabetes (see Fig. 7.2C ). The incidence of acute plaque rupture was similar in those with (32%) or without T2DM (27%).

Comparison of inflammatory infiltrate in patients with diabetes versus those without diabetes in autopsy samples.

Bar graph showing quantitative and semiquantitative comparisons of the extent of macrophages, T lymphocytes, and human leukocyte antigen–DR (HLA-DR) expression in coronary arteries from patients with diabetes and those without diabetes. Plaque macrophages and HLA expression were greater in patients with diabetes (type 1 and 2) than in those without diabetes, whereas T-cell infiltration was maximal in patients with type 1 diabetes.

Modified from Burke AP, Kolodgie FD, Zieske A, et al. Morphologic findings of coronary atherosclerotic plaques in diabetics: a postmortem study. Arterioscler Thromb Vasc Biol . 2004;24:1266–1271.

In sudden death victims, approximately half of individuals with diabetes showed triple-vessel disease, whereas those without diabetes more often had single-vessel disease (see Fig. 7.2D ), in line with total plaque burden being significantly greater in those with T2DM than in those without diabetes (358 ± 114 vs. 232 ± 128, P = 0.0001; see Table 7.2 ). Increased plaque burden may involve a higher rate of healed plaque ruptures, which can promote plaque progression. Diabetes is also linked to late complications following coronary artery bypass graft surgery, including acute MI and graft failure. while greater plaque burden in diabetes may underlie evidence suggesting worse outcomes, including mortality, after percutaneous intervention versus bypass surgery in diabetes.

Plaque hemorrhage is associated with intraplaque angiogenesis. Increased intraplaque hemorrhage, as assessed by glycophorin A staining of red cell membranes, has been linked with plaque progression, enlarging necrotic core, greater macrophage infiltration, and iron deposition within coronary atherosclerotic plaques. In T2DM, plaque pathology has been reported to contain increased angiogenesis, plaque hemorrhage and rupture, with similar findings in abdominal and thoracic aortas in T2DM ( Fig. 7.4 ). The extent of neovascularization correlates with macrophage and T-cell infiltration and plaque hemorrhage, which were greater in individuals with diabetes than in those without diabetes ( Fig. 7.5 ).

Angiogenesis, hemorrhage, iron deposition, and inflammation in diabetic coronary plaques.

Histologic sections from a 48-year-old black man with history of hypertension and diabetes who died suddenly. A , A low-power image shows fibroatheroma with severe luminal narrowing and angiogenesis. B – F , High-power images of the black box in A. B , Note abundant CD31 (platelet endothelial cell adhesion molecule 1 [PECAM-1]) staining that indicates the presence of angiogenesis (arrows). C , The same area shows abundance of iron (blue), suggestive of hemorrhage. D – F , There are also abundant macrophages that are detected by CD68, CD206 (mannose receptor), and CD163 (Hb-haptoglobin receptor) staining.

Inflammation, neovascularization, and intraplaque hemorrhage in aortic atherosclerosis from patients with and without diabetes.

A , Nondiabetic atherosclerotic plaque cap stained with CD68/CD3 in red chromogen shows mild inflammation in high-power field, 40 × (grade 1); less than 25 macrophages in the view. B , Diabetic plaque cap shows severe inflammation in high-power field, 40 × (grade 2); more than 25 macrophages per field are depicted in diabetic plaque. C , Distribution of inflammation grade in nondiabetic and diabetic plaques. D , Double-label immunohistochemistry with neovessels stained by CD34 in blue chromogen and inflammatory cells stained with CD68/CD3 in red chromogen. Nondiabetic plaque shows fewer dense neovessels and inflammatory cells in the high-power field, 40 ×. E , Diabetic plaque showing denser neovessels with tubuloluminal spaces surrounded by inflammatory cells seen in high-power field, 40 ×. F , Mean and 95% confidence interval for plaque neovessel density in nondiabetic and diabetic plaques. G , Nondiabetic plaque with less than 25% hemorrhage seen in high-power field (grade 1). H , Hematoxylin-eosin–stained diabetic plaque showing severe intraplaque hemorrhage occupying more than 75% of the plaque core seen in high-power field, 40 × (grade 3). I , Mean and 95% confidence interval for intraplaque hemorrhage (IPH) grade in nondiabetic and diabetic plaques.

Reproduced with permission from Purushothaman KR, Purushothaman M, Muntner P, et al. Inflammation, neovascularization and intra-plaque hemorrhage are associated with increased reparative collagen content: implication for plaque progression in diabetic atherosclerosis. Vasc Med. 2011;16:103–108.

Advances in coronary imaging now offers this modality as an additional means of characterizing atherosclerotic plaque in diabetes outside of performing autopsies. In the Computed TomogRaphic Evaluation of Atherosclerotic Determinants of Myocardial IsChEmia (CREDENCE) trial, patients (n = 303 subjects) referred for invasive coronary angiography also underwent coronary computed tomographic angiography (CCTA). An analysis of CREDENCE participants with diabetes (n = 95, 31%) found that those with diabetes had greater plaque burden than those without diabetes, with lesions that were often obstructive (58.1%). Patients with diabetes and nonobstructive disease had statistically significant increases in percentage of atheroma volume, PAV, noncalcified plaque, number of diseased vessels, and maximum stenosis compared to those without diabetes—findings that align with pathologic studies discussed previously. Another reported characteristic of coronary arteries in T2DM is a higher coronary calcium score. A three-vessel optical coherence tomography study in nonculprit plaques found that patients with diabetes had a higher prevalence of calcification, along with lipid index and thrombus, than those without diabetes. Imaging studies have also found increasing prevalence and extent of coronary plaque across a range of glucose from euglycemia to prediabetes to diabetes in asymptomatic patients undergoing coronary CT angiography. This glycemia range also aligned in a stepwise manner with increases in the presence of any coronary plaque, noncalcified plaque, and plaque with ≥1 high-risk features. Among participants with diabetes and a coronary calcium score of zero, 30% had coronary plaque. In another study of 898 patients with acute MI (with or without ST-segment elevation) who underwent successful percutaneous coronary intervention (PCI), three-vessel quantitative coronary angiography, and coregistered near-infrared spectroscopy, and intravascular ultrasound imaging were also performed before these subjects were followed for subsequent major adverse cardiovascular event (MACE) occurrence (3.7 years). These subsequent MACEs were adjudicated as to whether the recurrent event involved previously treated lesions and the associated characteristics of the previously obtained coronary imaging. In patients with diabetes (n = 109, 12.1%), recurrent events were more common and primarily involved the initial culprit lesion restenosis or prior nonculprit lesions that underwent spontaneous MI than those without diabetes. However, there was no difference between those with and without diabetes in regard to the extent of high-risk plaque characteristics seen at the start of the study. Although many interpretations of this data exist, it does suggest ongoing limitations in what imaging can inform us of regarding the nature of specific lesions and future events in those with diabetes.

Taken together, these and other studies underscore the complex pathologic state of diabetic atherosclerosis that overlap with nondiabetic atherosclerosis, even if augmented in diabetes, while other inputs, like RAGE pathways, that may be unique to diabetes (see Fig. 7.1 ).

These findings on the nature of atherosclerotic plaque in diabetes prompt the fundamental issue of what mechanisms drive this augmented pathology. Hyperglycemia defines the presence of diabetes, raising the key question of how glucose itself promotes atherosclerosis. Extensive evidence supports the relationship between diabetes and atherosclerosis, even extending to dysglycemic states in the absence of frank diabetes. Such data only fueled the frustration that despite this strong, direct, continuous relationship between hyperglycemia and atherosclerosis, treating hyperglycemia, including the use of many different strategies—earlier intervention, tighter glucose control, insulin-providing versus insulin-sensitizing strategies—all failed to decease atherosclerotic complications even while improving hyperglycemia. The landmark positive clinical CV outcome trials with GLP-1 RAs and SGLT2i in reducing adverse CV outcomes does suggest that the prior therapeutic failures may have derived in part to older antidiabetic agents not targeting mechanisms of potential CV benefit and/or involved offsetting adverse CV effects, as discussed elsewhere in this textbook. This positive data with GLP-1 RAs and SGLT2is also provide new opportunities for insights into modifiable mechanisms in atherosclerotic complications. While evidence that the CV benefit of these agents involved primarily their glucose-lowering properties has been largely lacking, with subsequent data indicating their benefits in those without diabetes, decreases in hyperglycemia may still have an interactive effect with other actions. Despite these questions about hyperglycemia in macrovascular disease, elevated glucose both contributes to vascular pathogenesis and is established as a modifiable factor for improving outcomes in microvascular disease, including neuropathy, retinopathy, and nephropathy. Together these various factors argue for detailed consideration of glucose effects on the vasculature.

Long-term intervention studies have tested if strict glucose control reduces CV consequences in diabetes. In T1DM, as reviewed more extensively elsewhere, the Diabetes Control and Complications Trial (DCCT) randomized adolescents age 15 years (87 patients) or young adults, average age 28 years (191 patients), to either strict versus standard glycemic control. Strict control of hyperglycemia reduced microvascular complications of diabetes compared with standard regimens of glycemic control. No difference in CV events was noted, although study participants were younger. However, in the Epidemiology of Diabetes Interventions and Complications (EDIC) study, a longer-term follow-up of DCCT, both the atherosclerosis surrogate of carotid intimamedia thickness and the clinical combined endpoint of MI, stroke, and CV death were reduced in those who had earlier received strict glucose control as compared to standard regiments. Particularly intriguing was that this reduction in CV complications was seen long after glycosylated hemoglobin levels between control and strict treatment groups became indistinguishable, suggesting a “legacy” effect and raising questions about the mechanisms underlying these results.

In T2DM, the mechanisms underlying the estimated two- to fourfold increase in CV risk are confounded by the overlapping, intertwined nature of risk factors in such patients, including obesity, hypertriglyceridemia, hypercholesterolemia, and hypertension, which coexist with hyperglycemia. The United Kingdom Prospective Diabetes Study (UKPDS) involved 3867 subjects with T2DM randomized to strict versus standard glycemic control. After 10 years, glycosylated hemoglobin levels were significantly lower in the strict control group (7.0%) versus standard (7.9%), with the strict control group having a 16% reduction in risk of MI, which was not statistically significant. Notably, in the posttrial monitoring program, those randomized to standard control group had significantly fewer MIs than the strict glycemic control group, despite their convergence to similar glycosylated hemoglobin levels after the study, another potential legacy effect raising further questions regarding glucose control in CV risk reduction.

Several large clinical trials offer insight into hyperglycemia as a mechanism in atherosclerotic complications and provide a broader context about this interaction. In the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial, stricter glycemic control was associated with higher CV and all-cause mortality versus the standard regimen, prompting early study discontinuation after a mean 3.5 years of follow-up. There was, however, a nonstatistically significant trend toward fewer nonfatal MI, nonfatal stroke, or death from CV causes among those in the tighter glycemic control groups. Subsequent analyses suggested more hypoglycemia in the glycemic control arms may have contributed to the increased mortality. Separately, two important trials—the Action in Diabetes and Vascular Disease: Preterax and Diamicron Modified Release Controlled Evaluation (ADVANCE) and the Veterans Affairs Diabetes Trial (VADT) —also studied glycemic control in T2DM; both trials showed neither CV benefit nor harm. Despite the failure of better glucose control to improve macrovascular complications, considerable work has investigated how hyperglycemia alters microvasculature responses, even if questions about the relationship between hyperglycemia, glucose control, and macrovascular disease persist. In this regard, other more subtle factors are noteworthy as to hyperglycemia as a factor in vascular biology and CV outcomes. Even if glucose control does not decrease MACE directly, improvements in glycemia and/or the glucose-lowering agents in use may help decrease pathologic parameters of a given ischemic event or enhance other effects of the agent. As an illustrative example of this concept, considerable preclinical and clinical data indicate that cerebrovascular events while on statin treatment are smaller than without statin treatment, a potentially distinct outcome than preventing the event completely. Separately, among glucose-lowering agents with established CV benefits, improved glucose levels may help augment the agents beneficial mechanisms and/or decrease pathologic parameters. Taking these issues together, reviewing the data for mechanisms that connect glucose metabolism and hyperglycemia to vascular responses is important.

POLYOL PATHWAY

The two major enzymes of the polyol pathway are aldose reductase (AR), the first and rate-limiting step that metabolizes glucose to sorbitol, and sorbitol dehydrogenase (SDH), which converts glucose to fructose. AR actions involve converting nicotinamide adenine dinucleotide phosphate, reduced form (NADPH) to nicotinamide adenine dinucleotide phosphate (NADP ⁺ ) with SDH consuming nicotinamide adenine dinucleotide (NAD ⁺ ) to yield NADH. Globally AR-overexpressing transgenic mice, which have human-relevant AR levels, worsen atherosclerosis in the Apo E-deficient proatherosclerotic mouse model, with effects that may also involve inflammation ( Fig. 7.6A and B ) ; similar data are seen in the low-density lipoprotein (LDL) receptor-deficient atherosclerosis model with streptozotocin-induced diabetes, effects not due to altered lipid parameters. Other work implicates endothelial hAR in promoting atherosclerosis. Of note, these transgenic hAR models of the polyol pathway increased proatherosclerotic changes in the presence of diabetes while administration of the AR inhibitor zopolrestat decreased the augmented atherosclerosis in the diabetic transgenic hAR/apoE-deficient mice ( Fig. 7.6C ). Together, these data implicate glucose flux via the polyol pathway in promoting diabetic atherosclerosis.

Impact of diabetes and aldose reductase (AR) expression on atherosclerosis at 14 weeks after induction of diabetes.

Representative images of aortic root sections stained with oil red O ( A ) and aorta specimens stained en face with Sudan IV ( B ). Hearts were retrieved from nondiabetic and diabetic apoE −/− (n = 10 and 9, respectively) mice, nondiabetic and diabetic apoE −/− hAR + (n = 10 in each group) mice, and diabetic apoE −/− hAR + mice treated with and without aldose reductase inhibitor (ARI) (n = 10 and 10, respectively) ( C ), and mean atherosclerotic lesion areas were determined.

Reprinted from Vedantham S, Ananthakrishnan R, Schmidt AM, Ramasamy R. Aldose reductase, oxidative stress and diabetic cardiovascular complications. Cardiovasc Hematol Agents Med Chem . 2012;10:234–240.

Of note, in contrast to these transgenic studies, an earlier report found that distinct AR inhibitors (ARIs, tolrestat and sorbinil) and genetic ablation of AR in diabetic apoE-deficient mice increased early lesion formation through increased levels of toxic aldehydes in lipid particles, suggesting AR might limit atherosclerosis. Experimental differences, specifically genetic overexpression of AR to human-relevant levels, compensatory responses to complete genetic AR deletion, as well as potential ARI off-target effects may underlie these discrepancies. In humans with diabetes and neuropathy, 1 year of ARI zopolrestat treatment improved cardiac function, as measured by echocardiography while vehicle-treated patients with diabetes demonstrated ongoing reduction in cardiac function. This study, which did not directly address diabetic atherosclerosis, suggested pharmacologic AR inhibition by zopolrestat did not worsen diabetic CV complications. Perhaps more potent and specific ARIs may hold promise for treating atherosclerosis in diabetes. Finally, it is worth noting that increased polyol pathway activity may increase oxidative stress. NADPH is a cofactor of glutathione production; consumption of glutathione through polyol pathway activity may reduce the availability of this antioxidant mechanism. These findings may align with data from mouse models and human macrophages that AR activity increases oxidative stress after high glucose and oxidized LDL exposure.

GLYCATION: RECEPTOR-DEPENDENT AND INDEPENDENT MECHANISMS IN DIABETIC ATHEROSCLEROSIS

Indirect consequences of hyperglycemia include the nonenzymatic glycation and oxidation of proteins and lipids to form AGEs, which also include key “intermediates” of dicarbonyl compounds (e.g., methylglyoxal [MG], glyoxal, and 3-deoxyglucone [3-DG]) ( Fig. 7.7 ). AGE formation involves multiple mechanisms: (1) aldehydic group of reducing sugars interacting with proteins or lipids, forming Schiff bases and Amadori products; (2) glucose flux via the polyol pathway; (3) lipid and sugar oxidation, with resulting dicarbonyl intermediates that ultimately yield AGEs after further modification. AGEs are heterogeneous compounds that include the highly cross-linked “brown” fluorescent AGEs (e.g., pentosidine and crosslines), nonfluorescent cross-linking AGEs (e.g., arginine-lysine imidazole), and noncross-linking AGEs (e.g., carboxymethyl lysine [CML]–AGEs).

Advanced glycation endproduct (AGE)/receptor for advanced glycation endproduct (RAGE): generation and interactions.

( A ) Chemical reactions leading to hyperglycemia-induced AGE generation are shown. Reactive intermediates include methylglyoxal (MG) and 3-deoxyglucone (3-DG). ( B ) Glucose-driven increases in intermediates and ultimately RAGE can increase reactive oxidative species (ROS), resulting in glutathione depletion and RAGE-dependent repression of the transcription of glyoxylase 1 (Glo1).

AGEs may form in diabetic tissues, exacerbating complications. For example, aging may promote AGE formation, particularly on long-lived proteins whose exposure to even normal glucose levels may gradually increase AGE formation. Hypoxia and ischemia/reperfusion (I/R) may generate AGEs, thereby increasing AGE damage in settings such as MI, stroke, or PAD. Renal failure significantly accelerates AGE formation; in patients with diabetes and severe nephropathy, increased AGE formation added to diabetes-associated glycation may significantly augment production and accumulation of these damaging species. In other settings, myeloperoxidase enzymatic activity reportedly generates CML-AGEs. Hence, in infectious or inflamed environments, AGE formation via inflammatory cell myeloperoxidase may further increase tissue stress, perhaps impairing effective wound healing, a major issue in diabetes.

Interestingly, food-derived AGEs may form during high temperatures while environmental pollutants (e.g., fly ash particles) may promote AGE formation. Of note, endogenous AGE detoxification mechanisms may exist, as suggested for the toxic AGE precursor MG dicarbonyl through the glutathione-dependent glyoxalase 1 (Glo1) enzyme system. Glo1 blocks MG formation into AGEs, producing lactate instead. In RAGE-deficient mice, Glo1 mRNA and protein levels are significantly higher in the kidneys versus those in diabetic wild-type RAGE-expressing mice, perhaps due to (1) decreased RAGE-dependent generation of ROS, which depletes glutathione and (2) RAGE-dependent transcriptional regulation of Glo1 ( Fig. 7.7 ).

Receptor-Independent Pathways

One significant consequence of cross-linking AGEs is the formation of intermolecular bonds between extracellular matrix (ECM) elements. In the vasculature, such interactions may increase arterial stiffness and molecular trapping, for example, of oxidized lipoproteins, promoting atherogenesis in diabetic macrovessels. Oxidized LDL contains significant amounts of AGE. Furthermore, AGE-induced ECM formation in micro- and/or macrovessels may increase vascular permeability, thereby facilitating inflammatory or other cells migration into the perturbed vessel wall.

Receptor-Dependent Pathways

Given the heterogeneous nature and diverse effects of AGEs, it is not surprising that multiple different AGE “receptors” have been identified, such as AGE-R1 (an antiinflammatory AGE receptor), members of the scavenger receptor families such as CD36, and the macrophage scavenger receptor. Among AGE receptors, the receptor for AGEs (RAGE) is a well-characterized signal transduction receptor of the immunoglobulin superfamily. RAGE binds AGEs such as CML-AGE and possibly hydroimidazolone AGEs. It is likely that other distinct AGEs exist and can also bind to RAGE.

RAGE is characterized by the presence of three extracellular domains, an N-terminal V-type Ig domain and two distinct C-type Ig domains, with the V-C1 domain reported as the chief ligand binding unit. The extracellular portion of RAGE is composed of a large hydrophobic patch and a large negative patch, which modulate RAGE ligand binding patterns.

In addition to AGEs, RAGE also binds other distinct ligands, as established for the S100/calgranulin family members. Although RAGE was first described as a S100A12 receptor, other work reveals S100B, S100P, S100A8/A9, S100A4, and S100A6, as examples, may bind to and signal via RAGE. S100 family members exert multiple pathogenic effects, including inducing and sustaining inflammatory reactions, promoting tumor cell proliferation, migration, induction of MMP expression and activity, and regulating cell survival. RAGE also transduces signals for high-mobility group box 1 (HMGB1). Like many RAGE ligand families, HMGB1 is also promiscuous, binding to RAGE and also certain toll-like receptor (TLR) signaling family members. HMGB1, like S100/calgranulins, exerts both proinflammatory and protumor properties. In tumor cells, HMGB1 binding to RAGE mediates increased pancreatic tumor cell autophagy and decreased apoptosis, which together enhances tumor cell survival. RAGE is also a receptor for amyloid-β peptide, among other amyloidogenic polypeptides, complement-related factor C1q and lysophosphatidic acid (LPA). These findings underscore RAGE complexity, defining it as not a “one ligand–one disease” molecule. Rather, multiple ligands of RAGE may converge in distinct settings and across different timelines, mediating pathogenic inputs like hyperglycemia and inflammation, effecting chronic disease pathogenesis, like diabetic atherosclerosis and cancer.

Considerable evidence links RAGE ligands to diabetes and atherosclerosis in animal models and humans. Immunohistochemical studies in human atherosclerosis reveal RAGE expression in atherosclerotic plaques from carotid and coronary artery lesions, with higher levels in samples from those with diabetes versus nondiabetic patients. RAGE expression in lesions in those with diabetes showed increased inflammation (higher macrophages and T-cell numbers), increased NF-κB activation and expression of COX-2/mPGES-1, increased MMP expression and activity, more apoptotic SMCs, and higher RAGE ligand S100A12 levels. Notably, RAGE expression in diabetic carotid plaques correlated with glycosylated hemoglobin values. Indeed, at the level of the RAGE ( AGER ) gene, RAGE ligands such as AGEs induce RAGE expression itself, at least in part via NF-κB binding elements within the RAGE promoter. RAGE is present in multiple cell types relevant to atherosclerosis, including ECs, monocytes/ macrophages, SMCs, and T lymphocytes. In these cell types, RAGE ligands mediate the induction of inflammatory signals and key transcription factors, such as NF-κB and Egr-1, and proatherogenic pathways, including in diabetes.

Preclinical in vivo studies have used various approaches to investigate RAGE in diabetic atherosclerosis. ApoE-deficient mice with streptozotocin-induced diabetes demonstrated increased atherosclerosis and increased vascular inflammation versus vehicle-treated normoglycemic mice. The role of RAGE was initially tested using soluble RAGE (sRAGE), the extracellular ligand-binding domain of RAGE. Administering sRAGE to diabetic apoE null mice decreased early atherosclerosis in a concentration-dependent manner and in other studies, atherosclerosis progression. Although streptozotocin-treated demonstrated higher plasma cholesterol, sRAGE administration reduced aortic inflammation, even in tissues without vascular lesions, but without altering cholesterol levels. Similar findings were observed in the T2DM-like db / db mouse model in the apoE null background, with sRAGE administration reducing atherosclerosis. In additional approaches, mice with global RAGE deficiency or specific endothelial deletion of the RAGE cytoplasmic domain (and other cell types in which the preproendothelin-1 promoter is active) demonstrated significantly reduced atherosclerosis, including in diabetes, with cholesterol or lipid-independent effects.

Studies reveal that the RAGE cytoplasmic domain binds to the formin family molecule diaphanous-1 (DIAPH1, also known as mDia1). The formin family regulates cellular signaling via effectors of Rho GTPase protein while also influencing cellular migration and cytokinesis. Studies to date in transformed cells, SMCs, and macrophages indicate that RAGE ligand-dependent signaling in these cell types is blocked after siRNA-knockdown of DIAPH1 or in DIAPH1 null cells. In vivo, nondiabetic mice undergoing guidewire-induced femoral artery endothelial injury displayed significant DIAPH1 induction, especially in SMCs ( Fig. 7.8A–I ). Furthermore, DIAPH1-deficient mice were protected from aberrant neointimal expansion ( Fig. 7.8J ). In keeping with DIAPH1 transducing RAGE signaling in SMCs, DIAPH1 null-injured vessels and isolated aortic SMCs displayed reduced oxidative stress, cell signaling via GSK-3β, and cellular migration compared with wild-type counterparts with intact DIAPH1 ( Fig. 7.8K ), data implicating DIAPH1 in transducing RAGE ligand effects on key signaling mediators, like activation of P-c-Src, Rac1, and Nox1, with subsequent AKT/GSK3β ser 9 phosphorylation ( Fig. 7.8L ), essential steps for RAGE ligand-induced vascular SMC migration in neointimal formation and atherosclerosis. Indeed, recent studies reveal that global deletion of DIAPH1 significantly attenuated atherosclerosis in male and female mice the LDL receptor null background.

Diaphanous-1 (mDia1) in vascular responses after endothelial denudation injury.

Wild-type (WT), Drf1 ^−/− , and RAGE ^−/− mice subjected to femoral artery endothelial denudation or sham, and tissues were analyzed at the indicated times. A and D , Assessment of neointimal expansion by elastic–van Gieson (E-VG) staining on day 21 after injury in WT mice ( A , sham and D , injury). B, C, E , and F , Immunostaining for mDia1 or isotype IgG control in WT mice on day 21 after injury ( E and F ) or sham ( B and C ). G– J , Immunofluoresence staining for mDia1 and α–smooth muscle cell actin (SMA) in sections of injured vessels revealed these two molecules colocalize in the neointima on day 21. K , Intima/Media (I/M) ratio measurement based on morphometric analysis of the vessels of WT and aged-sex-matched Drf1 ^−/− mice (n = 11/group) was performed 21 days after guidewire-induced femoral artery denudation. Representative images are shown. P < 0.001. L , Proposed mechanism of the role of mDia1 in receptor for advanced glycation endproduct (RAGE)-induced redox signaling smooth muscle cell (SMC) migration and neointimal expansion. The data supports a critical role for mDia1 in transducing the effects of RAGE ligands on P-c-Src, Rac1, and Nox1 activation and consequent phosphorylation of AKT/GSK3β ser 9 processes essential for RAGE ligand-induced vascular SMC migration.

Reprinted from Toure F, Fritz G, Li Q, et al. Formin mDia1 mediates vascular remodeling via integration of oxidative and signal transduction pathways. Circ Res . 2012;110:1279–1293.

Although RAGE antagonists have not yet been tested in humans with diabetic atherosclerosis, ongoing evidence links RAGE to this disease. Single-nucleotide polymorphisms (SNPs) of RAGE are associated with human diabetic atherosclerosis. Multiple studies report sRAGE levels correlate with human diabetic CV disease in humans. Hence, in diabetic atherosclerosis, RAGE may serve as both a biomarker and a therapeutic target.

Importantly, in addition to glycation, RAGE involvement in diabetic atherosclerosis also involves inflammatory mechanisms. Given that multiple RAGE ligands are expressed in diabetic macrovessels and largely converge on this specific receptor, identifying specific effects of different ligand classes for RAGE is challenging. Clinical translation involving RAGE may well require identification and study of more specific RAGE inhibitors.

Platelets of patients with T2DM manifest dysregulation of several signaling pathways, resulting in enhanced platelet adhesion, aggregation, and activation ( Fig. 7.9 ). Hyperglycemia, insulin resistance, dyslipidemia, inflammation, and increased oxidation may all increase diabetic platelet hyperactivity. Hyperglycemia increases platelet reactivity by altering different biochemical pathways, including PKC activation, with subsequent increased platelet granule release and aggregation. Glucose may also increase platelet reactivity through direct osmotic effects. In addition, by inducing nonenzymatic glycation of platelet surface proteins, hyperglycemia may decrease membrane fluidity while increasing platelet adhesion and activation. Consistent with these findings, acute hyperglycemia can increase platelet activation markers like P-selectin and CD40 ligand (CD40L), while improved glycemic control may decrease platelet reactivity.

Abnormal thrombosis and coagulation in diabetes.

Many pathologic inputs in diabetes promote platelet dysfunction and hypercoagulability, all thus contributing to a prothrombotic phenotype in patients with diabetes. Hyperglycemia and insulin resistance, a fundamental pathophysiologic feature of T2DM, drives inflammation, dyslipidemia, endothelial dysfunction, and oxidative stress. Each of these stimuli activates platelets by increasing expression of surface receptors for aggregation, increasing production of vasoactive molecules, reducing nitric oxide bioavailability. Simultaneously, production of coagulation factors by ECs including von Willebrand factor (vWF) and tissue factor, along with fibrinogen and factor VII from other sources, enhances coagulation. Lastly, a defect in endogenous fibrinolysis through increased PAI-1 expression tPA all conspire to heighten thrombosis in diabetes. AGEs , Advanced glycation end products; GP , glycoprotein; IGF-1 , insulin-like growth factor 1; IRS-1 , insulin substrate receptor 1; NO , nitric oxide; PAI-1 , plasminogen activator inhibitor 1; PGI ₂ , prostaglandin I ₂ ; PKC , protein kinase C; ROS , reactive oxygen species; TAFI , thrombin-activatable fibrinolysis inhibitor; TFPI , tissue factor pathway inhibitor; t-PA , tissue plasminogen activator.

Platelet aggregation is mediated by platelet surface receptors and adhesive proteins such as glycoproteins GPIIb/IIIa, GPIb, and P2Y12, each of which is altered in T2DM. Platelet turnover in T2DM appears accelerated. Hyperglycemia increases the release of reticulated, larger, and thus more reactive platelets, including a higher production of the potent vasoconstrictor, proaggregant thromboxane. Diabetic platelets may also have altered P2Y12 signaling, which increases adhesion, aggregation, and procoagulant activity. Increased levels of circulating microparticles, derived from platelets and various stimulated cell types, may also promote coagulation in diabetes. Microparticle size is larger in those with T2DM than in normal subjects, with an increased microparticle number being associated with increased diabetic complications.

Intracellular calcium flux helps regulate platelet function. Platelets in patients with diabetes contain lower cyclic adenosine monophosphate (cAMP) levels and higher intracellular calcium levels than nondiabetic patients, contributing to hyperreactivity, increased aggregation and activation, and stimulation of thromboxane synthesis. Altered calcium homeostasis may involve changes in the activity of calcium ATPases, which are highly sensitive to oxidative damage. The activity of calcium-activated proteases known as calpains is increased in platelets from those with diabetes, contributing to dysregulated platelet calcium signaling and hyperreactivity.

Insulin resistance and insulin deficiency can both alter platelet reactivity. Insulin opposes the effects of platelet agonists through activation of an inhibitory G protein by insulin receptor substrate 1 (IRS-1); during insulin resistance, impaired insulin receptor signaling attenuates insulin-mediated antagonism of platelet activation, thus increasing platelet reactivity. Insulin-like growth factor 1 (IGF-1), which is present in platelet granules while IGF-1 receptors are present on the platelet surface, stimulates tyrosine phosphorylation of IRS, potentiating platelet activation. Reduced insulin sensitivity in platelets lowers cAMP levels and increases intracellular calcium levels, enhancing platelets degranulation and aggregation, with platelets from insulin-resistant patients displaying diminished nitric oxide (NO), prostacyclin sensitivity, and platelet NO-synthase activity.

As noted, some other systemic abnormalities in diabetes may also alter platelet biology. Hypertriglyceridemia increases platelet reactivity, perhaps in part through apo E. Glycation of LDL particles may also impair NO production and increased intraplatelet calcium concentration, with subsequent increased platelet hyperreactivity and microparticle formation. Central obesity is linked to platelet dysfunction, with reduced responses to insulin, nitrates, and prostacyclin; elevated platelet count and volume; increased cytosolic calcium concentration; and increased oxidative stress. Interestingly, weight loss reverses some of these changes, decreasing platelet activation, as may be relevant to the CV benefits of GLP-1 RAs, as discussed elsewhere. Increased oxidative stress found in T2DM increases platelet reactivity. Lipid peroxidation and protein glycation may also affect platelet activation. Inflammation may promote platelet reactivity by increasing the expression of platelet activation mediators, such as CD40L, which also has proinflammatory properties and whose plasma-soluble levels are increased in T2DM.

T2DM has been associated with increased activity of the complex coagulation cascade that helps generate thrombin, converts fibrinogen to fibrin, and enables fibrin clot formation ( Fig. 10.2 ). For example, in healthy individuals, synthesis of the potent procoagulant protein tissue factor, expressed by ECs and vascular smooth muscle cells (VSMCs), is inhibited by insulin; platelets from T2DM patients have a significant increase in tissue factor levels versus matched controls, a shift promoted by both hyperglycemia and hyperinsulinemia. AGEs can also contribute to the activation of surface clotting factors. AGEs and ROS can promote tissue factor expression by activating NF-κB transcription factors.

In addition to tissue factor, many other coagulant proteins are implicated in T2DM’s prothrombotic state. Factor VII, linked to increased CV mortality, is elevated with hyperglycemia, insulin resistance, and T2DM. Hyertriglyceridemia is independently associated with Factor VII activity levels in patients with diabetes. Factor XIII, activated by thrombin, produces multiple cross-links in the fibrin clot, increasing resistance to lysis. Factor XIII subunit levels correlate with metabolic syndrome components and insulin resistance with evidence for an association between Factor XIII polymorphisms and thrombotic CVD risk. Von Willebrand factor (vWF), which promotes platelet adhesion by binding to the platelet glycoprotein GPIb receptor, has been linked to endothelial damage, atherosclerosis, and future CV events as well as specifically insulin resistance and T2DM. Increased platelet thrombin, which converts fibrinogen to fibrin, is associated with hyperglycemia. Thrombin is increased in patients with diabetes, including as a function of glucose control. Given the connection of hyperglycemia and many of these responses, improved glucose control may help reduce thrombogenicity. Fibrinogen, an acute-phase protein that independently predicts future CV events, is elevated in patients with diabetes and associated with microvascular and macrovascular complications. Glycemia and insulin resistance correlate with increased fibrinogen levels, perhaps through enhanced fibrinogen production stimulated by diabetes-associated hyperinsulinemia and inflammation. IL-6 cytokine levels, which are elevated in T2DM, can stimulate hepatic fibrinogen production, further connecting inflammation to prothrombotic states. Despite these myriad findings, questions persist regarding the extent to which glycemic control is involved in coagulation balance, especially given the question of CV risk reduction with prior clinical trials in diabetes; perhaps studies with the new agents with CV benefit will change this.

The prothrombotic state of T2DM involves not only increases in procoagulant factors, but also changes in endogenous anticoagulants, such as antithrombin, tissue factor pathway inhibitor (TFPI), protein C, and thrombomodulin, all part of physiologic coagulation balance. Antithrombin forms a stable complex with thrombin, limiting its action, as it does with other coagulation factors, like inhibiting Factor VII bound to tissue factor. Patients with diabetes reportedly have reduced antithrombin anticoagulant activity. Hyperglycemia may also induce conformational changes in antithrombin, prompting its retention and aggregation, perhaps through nonenzymatic glycation or endoplasmic reticulum (ER) stress. The endogenous anticoagulant TFPI, produced mainly by ECs and associated with atherosclerosis, inhibits tissue factor–initiated coagulation by binding with activated Factor X and modifying Factor VII–tissue factor catalytic complex activity. TFPI circulates primarily bound to circulating lipoproteins, with increased atherogenic lipoprotein levels associated with shifting tissue factor–TFPI balance toward plaque thrombogenicity. Other identified noncoagulant roles of TFPI, including inflammation, angiogenesis, and lipid metabolism, may also occur with diabetic vascular damage.

Activated protein C (APC), converted from protein C through endothelial thrombin-thrombomodulin complex action, inactivates the coagulation Factors V and VIII as well as PAI-1, thus promoting fibrinolysis and limiting inflammation, oxidation, and cellular damage. Low protein C levels are a reported risk factor for incident ischemic stroke but not CHD, while decreased APC generation is associated with progressive atherosclerosis in T2DM.

As noted, endothelial thrombomodulin also reduces procoagulant activities such as fibrinogen clotting, Factor V, and platelets while also modulating cellular proliferation, adhesion, and inflammation, and revealing endothelial damage. However, the association between circulating thrombomodulin levels and incident CHD is controversial. In healthy individuals, an inverse association between soluble thrombomodulin levels and risk for future T2DM is reported, while in patients with T2DM, plasma thrombomodulin levels were increased and positively correlated with metabolic syndrome components. Elevated plasma soluble thrombomodulin in T2DM levels may reflect enhanced hypercoagulability and altered fibrinolysis.

Blood flow involves a coordinated balance between thrombus formation and its dissolution. Fibrinolysis, the process of clot dissolution and removal, involves multiple interacting proenzymes and enzymes while inhibition of fibrinolytic pathways promotes thrombus formation; shifts in this fibrinolytic balance are strongly implicated in atherothrombosis. Impaired fibrinolysis has been noted in T2DM and is cited as a risk for CV complications. Altered glucose concentrations can induce fibrin network modifications that promote thrombosis. Fibrin clots in patients with diabetes are more compact, with decreased clot matrix pore size and more resistance to fibrinolysis, as compared to healthy controls. Indeed, improving glycemic control in T2DM may foster a more benign clot structure. The balance between thrombus formation and dissolution involves tissue plasminogen activator (t-PA) and PAI-1 as offsetting mediators. t-PA, produced by ECs, mediates plasminogen to plasmin conversion, initiating fibrinolysis. By binding to t-PA, PAI-1 blocks plasminogen conversion into active plasmin, thus inhibiting fibrinolysis. Adipocytes also produce PAI-1, a potential link between obesity in diabetes and CVD. In general, both t-PA and PAI-1 are reported to predict an increased CV risk and a worse prognosis, although still debated. PAI-1 is elevated in insulin resistance, correlates strongly with metabolic syndrome components, and may predict future T2D.

Hyperglycemia and hyperinsulinemia, by increasing expression and activation of proinflammatory mediators (e.g., NF-κB and PAI-1), reduce t-PA activity, shifting fibrinolytic balance toward thrombosis while glucose lowering may reduce PAI-1 levels.

Although t-PA, and its relationship with PAI-1, are key for thrombosis, other endogenous anticoagulants are implicated in diabetic atherothrombosis. For example, thrombin-activatable fibrinolysis inhibitor (TAFI) inhibits fibrinolysis by cleaving lysine residues on fibrin, interfering with t-PA and plasminogen binding. Increased plasma TAFI levels were reported in insulin resistance and T2DM patients. However, inconsistent results exist regarding TAFI levels and activation in thrombosis, especially in CAD. Alpha ₂ -antiplasmin, the main physiologic inhibitor of plasmin, may correlate with MI risk while the plasmin–alpha ₂ -antiplasmin complex formation, which indicates reactive fibrinolysis, may associate with subclinical atherosclerosis and CAD, with some questions as to gender-specific effects. Specific involvement of alpha ₂ -antiplasmin in diabetes-related pathology has been suggested but remains unresolved. More global assessment of whole-plasma fibrinolytic potential may better link fibrinolysis to arterial thrombosis than studies on specific fibrinolytic factors, especially in T2DM ( Table 7.4 ).

Endothelial dysfunction also involves increased endothelial adhesiveness. Increased vascular cell adhesion molecule 1 (VCAM1) protein levels are among the earliest molecular events seen in atherosclerotic experimental models such as the Watanabe heritable hyperlipidemic rabbit. In humans, adhesion molecules are detected in atherosclerotic plaque, and circulating levels of soluble adhesion molecules such as VCAM-1 and ICAM-1 predict the future risk of CVD, including in T2DM. Aortic endothelium from genetic models of murine hypercholesterolemia like the LDL receptor null mouse demonstrate increased leukocyte rolling and firm adhesion versus aortas from animals with normal lipid levels. Diabetes is associated with endothelial activation and enhanced leukocyte adhesion, driven by many forces, including reduced NO bioavailability and chronic vascular inflammation. TLRs are EC and macrophage surface proteins that bind circulating FFAs and propagate proinflammatory signal transduction cascades. The master transcription factor NF-κB mediates multiple proinflammatory responses, including those in the endothelium, enhancing the expression of adhesion molecules and chemoattractant cytokines (chemokines) to recruit monocytes to sites of injury. Chemokines, like MCP-1, are strongly implicated as key signals promoting atherosclerosis and diabetic atherosclerosis. Together, endothelial activation can be understood as a series of progressive steps in response to injury—whether involving hyperglycemia, elevated FFAs, hypertension, smoking, and/or other noxious stimuli—that promote multiple steps of leukocyte trafficking into the arterial wall. Such responses can be integral to host defenses and healing, even if maladaptive in atherosclerosis. Leukocyte influx, including monocytes and lymphocytes, increases plaque cellularity. Lipid-laden macrophages, termed foam cells , phagocytose necrotic cells and free cholesterol in the vessel wall, forming the characteristic atherosclerotic plaque and also help promote plaque disruption. Aside from these classic models of atherogenesis/atherosclerosis, superficial erosion , involving loss of ECs, is another pathologic mechanism involved in atherosclerotic plaque formation, progression, clinical ischemic events, and sudden cardiac death, which involves novel mechanisms like neutrophil extracellular traps, with potential relevance to diabetes.

Early atherosclerotic lesions, known as “fatty streaks,” typically form aorta branch points, are regions characterized by a disturbed blood flow profile distinct from the physiologic laminar shear stress other aortic regions encounter. Aortic ECs at these branch points are irregularly shaped, differing from ECs that align in the direction of laminar blood flow. Importantly, blood flow models indicate that shear stress forces not only alter EC shape and function but also gene expression through flow-responsive gene regulatory regions. Exposing static EC monolayers to physiologic shear stress levels induces genes that suppress atherogenesis (e.g., eNOS, superoxide dismutase, catalase, and TGF-β signaling molecules). ECs exposed to shear stress fail to express these atheroprotective transcriptional programs while also shifting their NO bioavailability, ROS generation, and levels as well as activity of NF-κB levels. Confocal microscopy reveals that lesion-prone aortic regions, including branch points, with their structural shifts in mechanical flow patterns, manifest higher endothelial NF-κB activity levels. Moreover, the NF-κB–dependent transcriptional responses at these branch points are significantly higher in the presence of low-level, proinflammatory stimuli, as often seen in diabetes. Enhanced inflammatory signaling through NF-κB activation and loss of NO heightens endothelial activation and early diabetic endothelial dysfunction. Thus hemodynamic forces help modulate endothelial function and inflammation involved in early atherosclerosis. Hypertension can also augment these flow- or stress-related programs, as often encountered in T2DM.

Inflammatory signals help recruit monocytes into the arterial wall. Monocyte differentiation into macrophages enables these phagocytes to begin engulfing cholesterol, forming foam cells and fatty streaks, prompting foam cell formation and further vascular inflammation, amplifying proatherogenic signals from ECs, circulating monocytes, and lesional macrophages. In humans with T1DM, circulating levels of monocyte-derived, proinflammatory cytokines are elevated, as seen with TNF-α, IL-6, IL-1β, and IL-1α versus controls, while other proinflammatory biomarkers including hsCRP, soluble (s) ICAM-1, sE-selectin, and sP-selectin are also elevated. Monocytes isolated from human patients with T1DM spontaneously secrete the proinflammatory cytokines IL-1β, IL-6, and TNF-α, which aligns with their higher mRNA levels while coculturing monocytes isolated from those with diabetes with lymphocytes increase IL-17–positive lymphocytes, cells implicated in vascular inflammation.

Although monocytes typically constitute only 5% to 10% of circulating leukocytes, they are considered critical determinants of atherosclerosis, which includes the significant phenotypic heterogeneity within this cell population. The data for specific monocyte/macrophage subtypes remains under active debate and exploration. A brief consideration of this topic of monocyte/macrophage polarity provides a framing construct as more data continue to emerge. In humans, classically activated monocytes (M1 cells) are positive for the surface marker CD14 and negative for CD16 (also known as FcγRIII). M1 cells represent 90% of the entire monocyte pool. Alternatively activated macrophages (M2 cells) are both CD14 and CD16 positive and serve posited antiinflammatory roles in tissue patrolling and inflammation resolution. These two monocyte populations also differ in other ways, including chemokine receptor expression. Certain chemokine patterned monocytes, like Ly-6C high monocytes, hone to atherosclerotic plaque and are also associated with adipose tissue infiltration in obesity. In humans with diabetes and known vascular complications, the circulating CD16 positive monocyte levels are reduced; issues regarding monocyte-specific subsets in diabetic atherosclerosis and their clinical utility continue to be pursued.

In murine models of T1DM, peritoneal macrophages elicited in response to thioglycolate injections demonstrate increased mRNA expression of proinflammatory mediators, including TNF-α, IL-1β, prostaglandin-endoperoxide synthase 2 (TPGS2), and COX-2. Indeed, an important relationship between fatty acid signaling and monocyte activation exists in diabetes. TLRs are primitive pattern recognition receptors that activate proinflammatory signal transduction cascades through NF-κB, as seen in response to long-chain fatty acids, among other inputs. Cell surface TLR expression in peritoneal macrophages increases twofold after the induction of diabetes in mice, with lesser effects on TLR4 expression but increased NF-κB activation. TLR2 deletion essentially abrogates this increased NF-κB activity. Similarly, levels of proinflammatory cytokines IL-1β, IL-6, MCP-1, and TNF-α are elevated in diabetic macrophages versus nondiabetic cells; TLR2 deficiency significantly attenuates this induction. Altered PTGS2 expression, along with the enzyme long-chain acetyl-CoA synthetase (ACSL1) from mouse monocytes with T1DM, correlates with increased levels of prostaglandin E ₂ (PGE ₂ ). In addition, CD14 ⁺ monocytes from human patients with T1DM also demonstrated higher ACSL1 mRNA levels. In M1-activated murine macrophages derived from bone marrow–derived monocytes, or human monocyte–derived macrophages (induced with LPS and interferon gamma [IFN-γ]), ACSL1 gene expression is significantly increased as well. Notably, ACSL1 deficiency reduced proinflammatory cytokine release from LPS-stimulated macrophages isolated from diabetic mice. ACSL1 deficiency in bone marrow reduced diabetes-associated atherosclerosis and monocyte accumulation in the vessel wall in low-density lipoprotein receptor (LDLr) deficient mice, implicating ACSL1 in monocyte recruitment and activation in experimental diabetic atherosclerosis. Whether this effect relates to altered PGE ₂ production remains unknown but suggests altered eicosanoid handling in diabetic atherosclerosis through altered monocyte inflammatory activation.

Multiple other drivers of inflammation and atherosclerosis in diabetes are also postulated. Oxidatively modified lipoproteins and oxidized lipid constituents, and their impact on monocyte-macrophage biology and circulating antibodies, continue to receive attention, including in diabetes. Autophagy—the process through which cells engulf and consume themselves—has received increasing attention as an inflammatory mechanism in macrophages, with effects on monocyte subtypes and is also suggested in cardiomyopathy. ER stress has been invoked as contributing to diabetes and atherosclerosis. The ER is integral to the metabolism of proteins, lipids, and glucose, playing a part in lipoprotein secretion and other basic cellular processes, including mitochondrial function and the unfolded protein response, all with direct relevance to diabetes and atherosclerosis. ER stress in diabetes and diabetic atherosclerosis has been linked to changes in apoptosis, inflammation, hepatic dysfunction, among other settings, offering another way in the complex intersection among cardio-metabolic issues can be further considered and understood.

Both B and T lymphocytes are implicated in atherogenesis in both the absence and presence of diabetes. By immunohistochemistry, most lymphocytes in atherosclerotic plaque are CD4 ⁺ cells, which can differentiate along Th1 or Th2 lineages, as directed by cytokines produced by other lymphocytes, ECs, and macrophages. Disruption of the Th1 lineage reduces atherosclerosis in murine models of disease. The role of Th2 cells continues to be explored and is coming into focus, with reported involvement in T1DM, gestational diabetes, MASLD, and other related conditions. Smaller subsets of T cells, including T regulatory cells (Tregs) and Th17 lymphocytes, exert local control on plaque inflammation and expansion. In patients with diabetes, a lymphocytosis has been observed with expansion of a less common, proatherogenic CD4 ⁺ CD28 null T lymphocyte; moreover, in acute coronary syndrome patients, this specific lymphocyte population’s frequency was 12.7% in T2DM versus 3.8% without T2DM, a pattern also independently associated with glycosylated hemoglobin. Increased visceral fat is associated with a loss of local Treg cells, one of many connections between T cells, inflammation, obesity, and CV disease.

During atherosclerosis, VSMCs proliferate in the media and also migrate from the media to the subintimal space, thus enlarging the neointimal lesion, processes also reported as being part of diabetes and its associated CV complications. VSMC reactivity is increased in isolated human arteries exposed to norepinephrine or phenylephrine, an effect that correlates with reduced subplasmalemmal Ca ^{2 +} , which helps control K ⁺ channels and VSMC relaxation. VSMCs grown under high-glucose conditions proliferate, migrate, undergo hypertrophy, and produce ECM more than VSMCs grown in low-glucose media. Aortic VSMCs isolated from db/db mice—an established model of aggressive T2DM—manifest increased proinflammatory gene expression (e.g., MCP-1 and IL-6), while other preclinical evidence points to enhanced VSMC responsiveness to stimuli like platelet-derived growth factor. These observations, together with other data, suggest that the diabetic environment, including AGEs, “preactivates” VSMCs and thus promotes pathogenic responses involving VSMCs.

Although cell biology and animal models have provided key scientific insights into atherogenesis, ongoing and extensive efforts have aimed at translating these findings to human disease. The identification of stable, circulating biomarkers of inflammation has allowed consideration of inflammation as an independent CV risk factors and targeting inflammation to decrease atherosclerotic complications, including in diabetes.

PPARs—the family of ligand activated transcription factors/nuclear receptors known to control gene expression in energy balance and metabolism—were also established as being expressed and active in vascular and inflammatory cells, in addition to their role in metabolic tissues and organs. The PPAR family consists of PPAR-α, PPAR-γ, and PPAR-β/δ, with each of these isotypes having unique profiles. The subsequent unfolding of the PPAR data and efforts to use PPAR agonists to improve CV outcomes in diabetes and dyslipidemia provides an example of the challenges in extending scientific findings to clinical benefit. While scientific and therapeutic efforts in diabetic atherosclerosis may have moved on from PPARs, as driven by clinical issues, this obfuscates the clear, important role of PPARs in biology and what unfolded in clinical trials.

Thiazolidinediones (rosiglitazone, pioglitazone) are potent insulin sensitizers that exert their effects by activating PPAR-γ, whereas fibrates lower hypertriglyceridemia and increase HDL by activating PPAR-α. Interestingly, these agents and their effects were identified even before PPARs were identified or well understood. The subsequent extensive evidence for PPAR action in vascular and inflammatory cells, including repressing inflammatory and proatherosclerotic gene expression in ECs, VSMCs, and macrophages, drive interest in these agents for treating diabetic atherosclerosis and diabetes associated dyslipidemia. Treating hypercholesterolemic mice with PPAR-α and PPAR-γ activating drugs in vivo also suppressed lesion formation in different models with and without diabetes. PPAR effects correlated with alterations in macrophage foam cell formation in vitro. While PPAR-δ was shown to be involved in these processes, no PPAR-δ agonist ever reached clinical approval. In contrast, considerable evidence argued that both PPAR-γ and PPAR-α activating agents exerted antiatherosclerotic effects, including early data in humans, like PPAR-γ drugs reducing biomarkers like hsCRP and decreasing surrogate endpoints like carotid intimamedia thickness while fibrates had early positive trials. Despite these prospects, PPAR activators yielded mixed results in translation to humans for different reasons, as discussed elsewhere.

Briefly, for PPAR-γ, a meta-analysis of smaller studies raised concerns about a possible increase in CV events with rosiglitazone. A prospective, placebo-controlled study with pioglitazone—PRoACTIVE (Prospective Pioglitazone Clinical Trial in Macrovascular Events)—did demonstrate a 20% reduction in the secondary endpoint of MACE in patients with diabetes but with a potentially flawed primary combined endpoint, which incorporated multiple outcomes, including the notoriously refractory outcome of improving peripheral vascular disease. If in fact the more traditional MACE outcome had been the primary endpoint in PRoACTIVE, the field may have unfolded differently. Importantly, pioglitazone did not increase adverse events in this study as had been suggested for rosiglitazone, one of several lines of evidence that establishes how PPAR agonists can have quite distinct transcriptional and functional effects, perhaps as a result of the especially large ligand- binding domain present in PPARs. For example, rosiglitazone does not lower triglycerides while pioglitazone does, perhaps through activity on PPAR-α. In contrast to PRoACTIVE, in the IRIS (Insulin Resistance Intervention after Stroke), pioglitazone decreased stroke as well as MIs, in line with the prior preclinical data. Despite this, multiple issues limited the clinical use of PPAR-γ, including ongoing concerns about CV safety for rosiglitazone, fluid retention that could cause issues for patients with heart failure, questions about other untoward side effects like bone fractures and bladder cancer, and weight gain. Interesting, PPAR-γ agonists can cause weight gain, which appears primarily subcutaneous, while also decreasing hsCRP, supporting different adipose depots as having different effects on inflammation.

Similarly, PPAR-α agonists also had a complicated story despite promising science. Fibrates as PPAR-α activating agents that decrease triglycerides and increased HDL, reduced CV events with gemfibrozil, in the Veterans Administration-HDL Intervention Trial (“VA-HIT”). Subsequent studies had less definitive evidence for primary endpoint benefit with fibrates, including fenofibrate in diabetes and in combination with statins in patients with T2DM. In these studies, the subgroup of patients with higher TG and lower HDL levels seemed most likely to benefit. However, subsequent data with the PPAR-α agonist pemafibrate failed to show a major adverse CV event benefit in PROMINENT, despite being performed in patients with higher TG levels. A substudy analysis of PROMINENT did suggest a benefit in PAD in patients with diabetes. Omega-3 fatty acids are also considered to be PPAR-α activating agents, in keeping with their effects on lowering triglycerides, again raising the prospect that the benefit of icosapent ethyl seen in REDUCE-IT may have involved PPAR-α activation. The basis for other omega-3 fatty acid trials not showing CV benefit is not clear but may align with the complexity of PPAR biology.

While PPARs are clearly key transcriptional regulators positioned at the intersection of metabolism, inflammation, and atherosclerosis, a distinction exists between any biologic target and the therapeutic agent(s) aimed at that target. While subsequent efforts to develop other PPAR modulators, including ones that would subtract adverse effects or have greater effectiveness, have generally failed, other endogenous mechanisms shown to generate natural PPAR agonists are associated with decreased atherosclerotic risk, such as genetic variants with increased LPL action and apoC3 loss of action.

In contrast to the complexities of PPAR therapeutics, HMG-CoA reductase inhibitors or statins have consistently demonstrated decreased CV risk in patients with diabetes, as also discussed elsewhere. Mechanistically, statin benefits involve lowering the LDL-C, with other pleiotropic effects also suggested, including antiinflammatory action. Statins lowered circulating hsCRP levels as compared to a placebo, as seen in randomized, prospective primary and secondary CVD outcomes trials. In the Pravastatin or Atorvastatin Evaluation and Infection Therapy—Thrombolysis in Myocardial Infarction 22 (PROVE IT–TIMI 22) trial, the subjects benefiting the most from high-dose atorvastatin therapy after acute coronary syndrome events were patients who achieved both an LDL-C level below 70 mg/dL and an hsCRP level below 2.0 mg/L. In JUPITER (Justification for the Use of Statins in Prevention: An Intervention Trial Evaluating Rosuvastatin), 17,802 individuals without known CVD and average LDL-C levels (<130 mg/dL, median 108 mg/dL) but hsCRP levels of 2 mg/L or higher were randomized to either a placebo or rosuvastatin 20 mg/day. The study was terminated after only 3 years when the interim analysis revealed a 44% reduction in the primary CV endpoint for those on rosuvastatin, who had significant reductions in hsCRP (37%). Those patients on rosuvastatin also had a major reduction in LDL-C (50%), precluding a definitive conclusion about a specific, causal antiinflammatory effect being the basis for the CV benefit. Preclinical studies continue to suggest various mechanisms through which statins may decrease inflammation independent of LDL lowering, including changes in modification of proteins, induction of other targets such as the kruppel-like factor (KLF) transcription factors, and changes in mRNA stability, as reported for eNOS.

Given that statins decrease both hsCRP and LDL, thus confounding the issue, consideration was given to other therapies that might decrease inflammation without altering LDL, both as proof-of-concept and as an additional potential therapeutic strategy, as relevant to diabetes. The Nod-like receptor protein 3 (NLRP3) inflammasome is a signaling pathway that generates the active form of IL-1β, which is strongly implicated in atherosclerosis and thrombosis, including NLRP3 activation by intracellular cholesterol crystals. The Cardiovascular Risk Reduction Trial (CANTOS; ClinicalTrials.gov identifier NCT013278) was a randomized, double-blind, placebo-controlled, event-driven trial testing if canakinumab, a human IL-1β neutralizing monoclonal antibody, reduces recurrent MI, stroke, and CVD death versus a placebo in patients at least 30 days after MI and who had hsCRP levels of 2 mg/L or higher. In CANTOS, the primary outcome (CV death, MI, or stroke), occurred in 4.11/100 person-years of the 50 mg group versus 3.86/100 person-years of the 150 mg group ( P = 0.02 for the canakinumab 150 mg group vs. placebo [other doses nonsignificant]). These benefits were seen despite no change in LDL-C, offering proof-of-concept that directly modulating inflammation can reduce CV events. While canakinumab is not being pursued for a CV indication, other direct modulators of inflammation are under investigation, including antibodies to IL-6, now in clinical CV outcome trials.

Other evidence regarding inflammation in CV risk derives from studies using agents for treating rheumatologic conditions. The Cardiovascular Inflammation Reduction Trial (CIRT; NCT01594333), sponsored by the National Institutes of Health, was a multicenter, placebo-controlled trial in subjects with stable CAD on standard care and metabolic syndrome or T2DM (n = 7000) randomized to low-dose methotrexate (15–20 mg weekly) with folate supplementation. Methotrexate, which lowers hsCRP without affecting lipid levels, had seven nonrandomized observational studies of patients with rheumatoid arthritis or psoriatic arthritis that suggested significant CV event reduction among those taking low-dose methotrexate. In CIRT, enrolled subjects, which excluded those with chronic inflammatory conditions (e.g., lupus, rheumatoid arthritis, or inflammatory bowel disease), low-dose methotrexate did not reduce CV events versus a placebo among patients with established CAD and either diabetes or metabolic syndrome or both. Importantly, levels of IL-1β, IL-6, and hsCRP were not decreased in the trial while those receiving methotrexate had increased side effects, including transaminitis, leucopenia, anemia, and infections.

In contrast to methotrexate, more positive data on modulating inflammation has been seen with colchicine, an oral tubulin binding and polymerization inhibitory agent that accumulates in neutrophils, exerting antiinflammatory effects on these and other relevant cells, including ECs and macrophages. Aside from studies in pericarditis, atrial fibrillation, and postcardiac surgery, colchicine was reported to reduce atherosclerotic complications in the open label LoDoCo pilot study (ACS, out-of-hospital cardiac arrest, or noncardioembolic ischemic stroke, 5.3% vs. 16%; P < 0.001) and LoDoCo2, a multicenter, double-blind, placebo-controlled randomized trial in stable CAD patients. The 5522 patients who tolerated colchicine were randomized and after 29 months (median), colchicine reduced the primary composite CV outcome (CV death, MI, ischemic stroke, or ischemia-driven coronary revascularization) by 31% (6.8% vs. 9.6%; P < 0.001) as compared to a placebo. In LoDoCo2, among the 18.2% with baseline T2DM, similar benefits were seen (HR, primary endpoint 0.87 [0.61–1.25] in T2DM, 0.64 [0.51–0.80] without diabetes, P interaction = 0.14). Interestingly, the incidence of new-onset T2DM was less in the colchicine group, although not significant (1.5% colchicine group, 2.2% in placebo group, P = 0.10). Although various caveats exist, LoDoCo2 supported inflammation as a potential therapeutic target, with relevance to diabetes.

The large, double-blind, placebo-controlled Colchicine Cardiovascular Outcomes Trial (COLCOT) investigated low-dose colchicine in patients within 30 days after MI. After 23 months, colchicine decreased the primary composite endpoint (CV death, resuscitated cardiac arrest, MI, stroke, or urgent hospitalization for angina requiring coronary revascularization) by 23% (HR, 0.77; 95% CI, 0.61–0.96; P = 0.02). Colchicine was in general well tolerated with minimal adverse events, aside from a small increase in nonfatal pneumonia hospitalization in those on colchicine. Of note, enrollment in COLCOT was not based on any specific inflammatory marker entry criteria. COLCOT enrolled 959 patients with T2DM, who benefited similarly to those without diabetes, with a primary endpoint event of 8.7% in the colchicine group and 13.1% in the placebo group (HR 0.65; 95% CI, 0.44–0.96; P = 0.03) although with more nausea (2.7% vs. 0.8%, P = 0.03, and pneumonia, 2.4% vs. 0.4%, P = 0.008). More recently, the CLEAR SYNERGY (OASIS 9) reported results in a 2 × 2 randomized controlled trial of low-dose colchicine 0.5 mg daily versus a placebo and spironolactone 25 mg daily versus a placebo in 7062 post-MI participants (mean age 60.5 years; 20% female) treated within 72 hours of the index percutaneous coronary intervention. In this trial, which included 18% of patients with diabetes, colchicine showed no benefit, contrasting with the COLCOT and LoDoCo2 results, which had enrolled different patient populations.