MicroRNA and mechanisms of impaired angiogenesis in diabetes mellitus

Normal angiogenic processes are disturbed in diabetes mellitus

Diabetes mellitus impairs physiological angiogenesis, which may be manifested as non-healing foot ulcers or refractory angina. Multiple molecular mechanisms have been proposed. Hyperglycemia induces the generation of reactive oxygen species (ROS) that cause endothelial derangements (1), including the reduced synthesis(2) and accelerated degradation(3) of endothelial-derived nitric oxide (NO). The bioactivity of NO is critical for angiogenic processes, such as the survival, proliferation and migration of endothelial cells(4). The impairment in NO bioactivity may also explain in part the reduced expression of a major angiogenic cytokine, vascular endothelial growth factor (VEGF) in hyperglycemic states, as NO and VEGF have a reinforcing and reciprocal relationship(5). Glucose intolerance also reduces the number and function of bone-marrow derived endothelial progenitor cells(6), circulating cells which participate in the angiogenic response. In addition to generating ROS, hyperglycemia may impair cytoprotective mechanisms against oxidative stress. In particular, the thioredoxins play a key role in angiogenic processes by maintaining endothelial redox homeostasis, with favorable effects on protein folding, activity of reductive and metabolic enzymes, energy utilization, and transcription factor activity(7). Emerging evidence indicates that hyperglycemia upsets opposes this cytoprotective mechanism by increasing the expression of the endogenous inhibitor, thioredoxin interacting protein (TXNIP).

Is it possible that these and other disparate mechanisms for the impaired angiogenesis in diabetes mellitus have a common genomic basis? This question logically follows from the work of Caporali and colleagues in the current issue of Circulation {ref this paper}. They have discovered a novel genomic mechanism for hyperglycemia-induced impairment of angiogenesis, i.e. the increased expression of a specific microRNA (miRNA) that appears to orchestrate a pathophysiological response in diabetes mellitus.

MicroRNA and genomic regulation

Accumulating data indicate that non-coding RNA plays a critical role in genomic regulation (8). One of these noncoding RNAs is the so-called microRNA. Discovered in vertebrates less than a decade ago(9), these short (~22 nucleotide), single-stranded, endogenous RNA molecules potently inhibit the translation of specific mRNAs(10). This effect results from the binding of the so-called “seed-sequence” near the 5′ end of the miRNA to its complementary target within the 3′ untranslated region (UTR) of the mRNA molecule(11). Occasionally, perfect Watson-Crick pairing is achieved and the mRNA is degraded(Fig 1). More commonly, imperfect pairing occurs and translation is impeded without destruction of the genetic transcript. Each miRNA may regulate upwards of 500 different genes. The genes of a cluster regulated by single miRNA commonly act together to modulate integrated pathways subserving a biological response (12). Because of the promiscuity of the miRNA system, over 30% of all human genes are predicted to be regulated by fewer than 1000 individual miRNAs in man(12). This intricate and highly conserved class of molecules plays a critical role in many pathological conditions(13), including vascular inflammation(14), arterial remodeling(15), smooth muscle plasticity(16), atherosclerosis(17), stem cell differentiation(18) and endothelial cell apoptosis(19).

Figure 1.

Metabolic alterations in diabetes mellitus increase the generation of microRNA 503. MicroRNA503 causes translational repression or degradation of mRNAs encoding proteins needed for endothelial processes required for angiogenesis.

A microRNA mechanism for impaired angiogenesis in diabetes

In the current issue of Circulation, Caporali and colleagues have augmented our understanding of miRNA biology in the vascular pathophysiology observed in diabetes. They discovered that when endothelial cells were exposed to conditions that mimic hyperglycemia and tissue-hypoxia, the cells expressed increased levels of miRNA-503 {ref this paper}. To determine if there is a causal role for miRNA-503 in the impaired angiogenesis observed in diabetes, they forced the expression of this miRNA in endothelial cells. They observed a dramatic and deleterious effect of this miRNA on several processes central to angiogenesis, including endothelial proliferation, migration and tube formation in the in vitro Matrigel model of angiogenesis. To determine if this miRNA was operative in vivo, they studied its expression in animals and humans. They observed that miRNA-503 expression was increased after surgically inducing ischemia in the hindlimbs of diabetic mice. Notably, miRNA 503 was also elevated in the blood and the calf muscles of diabetic patients with advanced limb ischemia. To definitively show that this particular miRNA directly causes - and is not just associated with – impaired angiogenesis in diabetes, the authors injected into the ischemic hindlimb of diabetic mice a miRNA “decoy” (which contained multiple copies of the target miRNA binding site). Using this technique, miRNA-503 was effectively scavenged, and was no longer available to bind and inhibit its target mRNA. As predicted, they observed that antagonizing miRNA-503 resulted in a dramatic improvement in blood flow and post-ischemic angiogenesis – the first example of a miRNA-based intervention restoring physiological angiogenesis in diabetes.

The study left unanswered some questions regarding the mechanism of action by which miRNA-503 has its anti-angiogenic effects. Caporali et al. report that the anti-angiogenic miRNA-503 targets two well known cell-cycle regulating genes, Cyclin E and cdc25 - findings which have been described previously in other cell types(20-21). These cyclin-related factors play an absolutely central role in the cell’s decision to undergo the G1-to-S transition and govern critical cell-fate processes such as differentiation, proliferation and cellular senescence(22). Intriguingly, major genome-wide association studies (GWAS) for cardiovascular disease states have implicated polymorphisms in close proximity to two major cyclin regulators (cyclin-dependent kinase inhibitor 2A and 2B)(23-24). It is possible that these GWAS findings are related to those of the current paper, and underscore the importance of the cell-fate decision pathway in vascular disease. Cell-fate decisions influence whether or not an endothelial cell will proliferate, migrate and incorporate into a new capillary plexus; whether an endothelial precursor cell will differentiate or retain its stem cell characteristics. Cell fate decisions affecting other lineages in the vessel wall influence the progression of disease; e.g. smooth muscle cell migration and proliferation in restenosis. A broader understanding of the processes that govern these choices- genetic, epigenetic or otherwise- will inform new therapeutic targets and greatly advance our understanding of heritable and acquired vascular disease.

Other miRNAs participate in angiogenesis

Prior to the work of Caporali and colleagues, others have found a role for miRNAs in regulating angiogenesis. Early reports revealed that Dicer (a critical microRNA-processing enzyme) led to a dramatic impairment in blood vessel network formation and embryonic lethality- indicating a critical role of miRNAs in vasculogenesis(25). Subsequent studies have identified several potent pro- and anti-angiogenic miRNAs(26).

For example, miR-126 is now known to target SPRED1 and PIK3R2, two critical inhibitors of the angiogenic cytokine, VEGF(27). By reducing the expression of these anti-angiogenic cytokines, miR-126 enhances capillary network stability and flow-induced vascular remodeling in zebrafish models of blood vessel development(28). Other vasculogenic miRNAs (and their target genes) have also been described, including Let-7 (thrombospondin-1)(25), miR-210 (Ephrin A3)(29-30), and the miR-17-92 cluster (thrombospondin-1 and connective tissue growth factor)(31), amongst others. Conversely, several anti-angiogenic miRNAs have also been uncovered, including miR-221/222, miR-92a and miR-509c, which inhibit tube formation, vessel growth and other endothelial cell functions by reducing c-kit, integrin subunit α5 and Hif-1α signaling, respectively(32-34).

How these miRNAs interact to modulate angiogenesis in health and disease is not known. It is likely that some are more prominent than others in specific disease states. Furthermore, it is not clear which if any of these miRNAs underlie other molecular mechanisms featured in diabetic pathophysiology (e.g. generation of ROS and impairment of NO bioactivity). A striking feature of diabetes mellitus is the heterogeneity of angiogenic dysregulation. For example, VEGF is upregulated in the diabetic eye whereas VEGF signaling is impaired in the peripheral vasculature.(35) Could differential tissue regulation of miR503, and/or some other angiomodulating miRNA, explain the paradox of attenuated angiogenesis in the diabetic leg ulcer, co-existing with proliferative angiogenesis observed in the retina of the same diabetic patient?

There remain many questions of scientific interest and clinical relevance to be addressed in this emerging research front. Chief among these is the question of how we will modulate miRNA expression for therapeutic purposes. To be sure, further developments in this field are likely to uncover novel methods of promoting (e.g. vascular regeneration in ischemia) or inhibiting (e.g. tumor growth in metastasis) angiogenesis via manipulation of this epigenetic system.