Introduction

Diabetic retinopathy (DR), a common complication of diabetes, is a leading cause of vision loss [1, 2]. The early non-proliferative stages of DR are characterized by retinal microvascular damage triggering vascular hyperpermeability [2]. The retinal endothelium, which is the intimal lining of blood vessels, forms a barrier between the blood and the interstitium, regulating extravasation of plasma proteins, fluid, and leukocytes [3, 4]. Advances made in recent decades have revealed the complex nature of this semi-permeable membrane, and the key role played by the membrane in the maintenance of vascular homeostasis [5–7]. Appropriate retinal microvascular endothelial function is important in terms of eye homeostasis, and its dysfunction is associated with several pathophysiological conditions, including DR [8–10].

The end-stage of microvascular endothelial dysfunction (MVED) caused by hyperglycemia involves angiogenesis and apoptosis of endothelial cells [11–13]. Angiogenesis, which is the formation of new blood vessels from preexisting vessels, is important in terms of wound healing and in the pathology of many diseases, including tumor development and DR [14]. Angiogenesis is regulated by vascular-specific growth factors that trigger complex intracellular signaling cascades culminating in activation of endothelial cells. The relevant growth factors include vascular endothelial growth factor (VEGF), pigment epithelium-derived factor (PEDF), and RhoA/mammalian Diaphanous homolog 1 (mDia-1)/profilin-1 [15–19].

Recently, an important role of profilin-1 in wound-induced endothelial cell motility has been discovered [20–22]. In mammals, four profilin genes (profilin 1 – 4) have been described to date [23]. Profilin-1 was the first-known family member and is ubiquitously expressed by fungi, plants, certain viruses, and most animal cells except those of skeletal muscle [24, 25]. The roles played by Rho guanosine triphosphatases (GTPases) have been extensively studied in various types of mammalian cells, with the aid of (principally) dominant-negative and constitutively active mutants. The RhoA protein is conserved in evolutionary terms in everything from plants and yeasts to mammals; it acts by binding to and stimulating various downstream targets, including actin nucleators, protein kinases, and phospholipases, (especially mDia-1 and profilin-1) [26]. Although profilin-1 plays important roles in actin polymerization, and RhoA/mDia-1 signaling is involved in endothelial protection, profilin-1 has not yet been clearly shown to engage in angiogenic regulation.

In this review, we summarize the details of the RhoA/mDia-1/profilin-1 signaling involved in MVED caused by hyperglycemia. Our aim is to explore the relationship between DR microvascular endothelial function/dysfunction and the actions of the RhoA/mDia-1/profilin-1 cascade, with a principal focus on cellular and molecular mechanisms as opposed to clinical manifestations.

RhoA

RhoA is the prototypical member of the mammalian Rho subfamily, which has 20 members. The Rho GTPases are 20 to 24 kDa proteins that are essential for appropriate regulation of many cellular functions. The Rho GTPases have been extensively researched, as have Cdc42, Rac1 and other isoforms of G proteins [27–29]. RhoA is a molecular switch that responds to messages from G-protein-coupled receptors and cell surface receptors that bind cytokines, growth factors, and adhesion molecules. The Rho GTPases cycle between inactive guanosine diphosphate (GDP)-bound forms and active guanosine triphosphate (GTP)-bound forms, and their intrinsic hydrolytic activities are affected by various regulators (Fig. 1) [30–33]. Cycling of Rho GTPases between the two states is regulated by three sets of proteins; these are the guanine nucleotide-exchange factors (GEFs), the GTPase-activating proteins (GAPs), and the guanine nucleotide-dissociation inhibitors (GDIs) (Fig. 1) [26].

Fig. 1
figure 1

Focus on the cascade of RhoA/mDia-1/profilin-1 pathway. (1) An inactive RhoA-GDP form switching to an active RhoA-GTP form is regulated by GEF, GAP and GDI, etc. (2) RhoA-GTP binds to the DID part and sequence N-terminal to DID of mDia-1, while profilin-1 binds to the FH1 part of mDia-1. After forming a 1:1 complex with G-actin, profilin-1 modulates actin stress fibers via promoting or preventing actin polymerization. (3) RhoA/ROCK I/II could mediate actin polymerization by regulating profilin-1. Abbreviations: GDP guanosine diphosphate, GTP guanosine triphosphate, GEF guanine nucleotide-exchange factor, GAP GTPase-activating protein, GDI guanine nucleotide-dissociation inhibitor, mDia-1 mammalian Diaphanous homolog 1, DAD diaphanous auto-regulatory domain, FH1/2 formin homology 1/2, CC coil-coiled region, DID Dia-inhibitory domain, ROCK I/II Rho-associated coiled-coil containing protein kinase I/II

Full size image

Much work performed over the past few decades has suggested that five classical pathways may be closely associated with DR; these are the advanced glycation end products (AGEs), oxidative stress, protein kinase C (PKC), hexosamine, and polyol pathways. In diabetes, RhoA and the receptor for AGEs (RAGE) can form a complex termed RhoA/RAGE, which has been suggested to activate Rho-associated coiled-coil-containing protein kinase (ROCK), resulting in reorganization of the actin cytoskeleton, in turn triggering endothelial cell hyperpermeability [34–36]. A recent study showed that RhoA activity was markedly increased, and endothelial nitric oxide synthase (eNOS) phosphorylation was downregulated by 57 % in retinas of diabetic rats 2 weeks after the onset of diabetes [2, 37]. Interestingly, in other studies, diabetes caused a ROCK-mediated increase in endothelial arginase activity, contributing in part to the impaired nitric oxide (NO) bioavailability characteristic of the disease (Fig. 2) [38, 39].

Fig. 2
figure 2

The crosstalk between RhoA/mDia-1/profilin-1 cascade and five classic pathways that might be closely associated with DR. (1) In diabetic condition, hyperglycemia activates RhoA signaling to interact with a multitude of pathways, e.g., AGEs, PKC, ROS, hexosamine and polyol pathways. (2) RhoA could lead to MVED via junction complexes, such as ZO-1, occludin and claudins, etc. (3) ROCK acts downstream of RhoA to regulate MVED. (4) Some indispensable molecules also participate in the crosstalk, including NF-κB, RAGE, MAPKs, NO, BSA and NADPH, etc. Abbreviations: mDia-1 mammalian Diaphanous homolog 1, ROS reactive oxygen species, AGEs advanced glycation end products, RAGE receptor for AGEs, PKC protein kinase C, ROCK Rho-associated coiled-coil containing protein kinase, NO nitric oxide, phos- phosphorylation, GEF guanine nucleotide-exchange factor, BSA bovine serum albumin, NADPH nicotinamide adenine dinucleotide phosphate, NF-κB nuclear factor-kappa B, MAPKs mitogen-activated protein kinases, MVED microvascular endothelial dysfunction

Full size image

RhoA is involved in other cellular signaling pathways and in a variety of physiological and pathological processes, because it regulates many such processes, including cytoskeletal dynamics, cell polarity, membrane transport, and gene expression [40]. Glucosamine, a product of glucose influx via the hexosamine biosynthesis pathway (HBP) in diabetes, has recently been shown to increase vascular contraction, at least in part via activation of the RhoA/ROCK pathway [41]. Hyperglycemia increases endothelial cell RhoA/ROCK activity in a PKC and reactive oxygen species (ROS)-dependent manner, and activated ROCK mediates glucose-induced expression of the plasminogen activator inhibitor-1 (PAI-1) [37, 42–44]. Furthermore, the polyol pathway is also activated when microvascular permeability increases under hyperglycemic conditions, and the immediate cellular changes observed can be abrogated by inhibiting ROCK [45]. These studies have revealed the critical role played by the RhoA pathway in retinal MVED associated with diabetes (Fig. 2).

RhoA and the endothelium

The endothelium forms the inner lining of blood vessels and is metabolically active [46]. Apart from functioning as a barrier, it senses and responds to environmental factors and has important autocrine and paracrine functions that regulate the contractile state of blood vessels, the hemostatic balance, and other cellular functions [46, 47]. Recently, RhoA has received much attention as a key regulator of cell shape, movement, and proliferation [48].

RhoA and its downstream effector ROCK I/II modulate cell adhesion, migration, proliferation and apoptosis by controlling arrangement of the actin skeleton as well as cell shrinkage [33, 49, 50]. It has been shown both in vivo and in vitro that the Rho pathway plays a critical role in diabetic retinal microvascular pathology, and that hyperglycemia triggers retinal hypertonicity by activating Rho signaling and subsequently increasing RhoA/ROCK activity [2, 37, 51]. The calcium channel blocker fasudil and the lipid-lowering agents ezetimibe and simvastatin protect the retinal microvascular endothelium by inhibiting RhoA/ROCK activity, thus ameliorating endothelial proliferation and hypertonicity in diabetic patients [2, 46, 52, 53].

The evidence that RhoA is ubiquitously expressed in various types of endothelial cells is overwhelming, and it is clear that RhoA activation induces the breakdown of the endothelial barriers of microvascular mesenteric endothelial cells, human dermal microvascular endothelial cells, and macrovascular pulmonary artery endothelial cells, but not microvascular myocardial endothelial cells [54]. The primary negative regulator of Rho, RhoGDI-1, represses RhoA activity in the lung microvessel endothelium, and thus preserves endothelial barrier function in vivo. Inhibition of the RhoA pathway by GDIs can reverse the increase in microvascular permeability induced by acute stimulation with the PAR1 peptide or prolonged stimulation of RhoGDI-1-/- mice [55].

RhoA is also involved in the induction of endothelial hyperpermeability by certain agents, including thrombin, VEGF, angiopoietin-2 (Ang-2), and lysophosphatidic acid (LPA) [56–58]. RhoA and PI3 kinase mediate certain processes, and specific inhibitors prevent ROS-induced monocyte migration across an in vitro model of the blood brain barrier (BBB). Interestingly, such processes are also mediated by protein kinase B (PKB/Akt), previously unrecognized as a player in cytoskeleton and tight junction (TJ) dynamics; PKB acts downstream of both RhoA and PI3 kinase [59].

RhoA also regulates the activities of inter-endothelial junctions, affecting cell motility, proliferation, survival, and permeability [60–63]. In both DR and macular edema (ME), the TJ proteins occludin and zonula occluden-1 (ZO-1), and the adherens junction protein cadherin-5, are critical for maintenance of the endothelial barrier and for modulating the paracellular transport of large vessel endothelia [64–67]. Recently, many in vivo and in vitro studies have shown that RhoA/ROCK signaling affects the activities of junction complexes [51, 68–70]. Rho inhibition reduces localization of ZO-1 and occludin to cell junctions. Notably, constitutive Rho signaling conversely causes ZO-1 and occludin to accumulate at cell junctions [71].

The observed improvement in endothelial function upon inhibition of Rho-kinase activity by ezetimibe, an inhibitor of intestinal cholesterol absorption, suggests that the agent might have novel anti-atherogenic effects in humans [53]. In addition, inhibitors of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase, such as statins, may improve endothelial function and decrease vascular inflammation and atherosclerosis by inhibiting the Rho/ROCK pathway [46].

RhoA and angiogenesis

RhoA coupled to Gα13 modulates cell migration [72, 73]; this is accomplished by the direct interaction of activated Gα13 with a family of GEFs binding to RhoA. This family is termed the “regulator of G-protein signaling (RGS)-homology domain (RH)-containing GEFs” (RH-RhoGEFs) [73–75]. Disruption of the Gα13 gene in mice impaires the ability of endothelial cells to develop into an organized vascular system, resulting in intrauterine death [76–78]. The biological functions of RhoA depend principally on the associated RhoGEFs, which both control the RhoA GDP/GTP binding state and directly influence the development of angiogenesis [79, 80].

Rho signaling is also involved in sphingosine-1-phosphate (SPP)-induced angiogenesis, which is greatly inhibited by C3 transferase, a type of RhoA inhibitor [81]. A great deal of in vivo and in vitro work has shown that RhoA is required for angiogenesis [82–85]. Conversely, angiogenesis progression is restrained by blocking Rho or RhoA/ROCK signaling [86–88].

In terms of pharmacology, cerivastatin, a cholesterol-lowering agent, has been shown to inhibit in vitro microvascular endothelial cell proliferation induced by growth factors; this is reversed by treatment with geranylgeranyl pyrophosphate (GGPP). This mechanism is associated with the transition of RhoA from the cell membrane to the cytoplasm and depolymerization of actin fibers (which is also prevented by GGPP treatment). RhoA-dependent inhibition of cell proliferation is mediated by inhibition of focal adhesion kinase and Akt activation [89]. In addition, it has been proposed that the HMG-CoA reductase inhibitor simvastatin may interfere with angiogenesis by inhibiting the geranylgeranylation and membrane localization of RhoA. Furthermore, tube formation is inhibited by GGTI (a specific inhibitor of Rho geranylgeranylation), C3 exotoxin (which inactivates Rho), and adenovirus-mediated expression of a dominant-negative form of RhoA (which reverses the effect of simvastatin on tube formation). Finally, inhibitors of HMG-CoA reductase also inhibit signaling by VEGF, Akt, and focal adhesion kinase (three RhoA-dependent pathways involved in angiogenesis) [90].

To the best of our knowledge, vascular angiogenesis is regulated by several cytokines, of which VEGF-A and its receptor, VEGF receptor 2 (VEGFR-2), play indisputably important roles [91–94]. VEGF induces RhoA activation and its recruitment to the membrane of human endothelial cells. ROCK inhibition prevents VEGF-enhanced endothelial cell migration that follows mechanical wounding, but has no effect on basal endothelial cell migration. These findings indicate that the VEGF-induced cytoskeletal changes in endothelial cells require both RhoA and Rho kinase, and activation of signaling by these materials is involved in the VEGF-induced in vitro migration and angiogenesis of endothelial cells [95]. Furthermore, transient overexpression of the dominant-active RhoA mutant also increases tyrosine phosphorylation of VEGFR-2, whereas overexpression of a dominant-inactive form of the protein has no such effect. Together, these results indicate that the Rho proteins play important roles in angiogenesis by modulating the tyrosine phosphorylation status of VEGFR-2 [96].

Controversially, loss-of-function experiments with endothelial cells have revealed that inhibition of ROCK I/II by the pharmacological inhibitor H-1152 and ROCK I/II-specific small-interfering RNAs (siRNAs) increases VEGF-driven retinal neovascularization and sprouting angiogenesis [97].

Recently, a regulatory role for VEGF-C in initiation and potentiation of angiogenesis has been described [98–100]. VEGF-C knockdown decreases RhoA expression. Furthermore, RhoA knockdown, even upon supplementation with VEGF-C or VEGF-A, decreases endothelial cell proliferation and stress fiber formation, indicating that VEGF-C promotes angiogenesis via a RhoA-mediated pathway [101].

RhoA and apoptosis

RhoA was first described as an inhibitor of endothelial cell death [102]. An earlier study suggested that inhibition of the RhoA/ROCK pathway by a ROCK inhibitor, Y27632, attenuates glucose-induced apoptosis to an extreme degree [103]. More recently, RhoA signaling has been shown to trigger mitochondrial proximal tubule cell apoptosis in response to mechanical stretching, which is inhibited by phosphorylation of Erk1/2 and p38 MAPK [104].

On the other hand, inhibition of RhoA/ROCK1 signaling promotes apoptosis of leukemia cells by enhancing phosphorylation of Erk1/2 in an Mek1/2-independent manner [105]. Intriguingly, it had recently been shown that inhibition of RhoA and ROCK I activation by the C3 exoenzyme and Y27632, respectively, attenuates apoptosis of human leukemia cells [106].

RhoA involvement in cancer cell apoptosis is a major field of research. RhoA activation, induced by CNFy, triggers intrinsic apoptosis of the prostate cancer cell line LNCaP [107], and inhibition of RhoA/ROCK signaling promotes apoptosis of gastric cancer cells [108]. These data reveal a novel RhoA activity, which may aid in a comprehensive understanding of DR, but which currently remains enigmatic.

mDia-1

mDia-1 is an isoform of the formin family, which contains potent dynamic regulators. The formin family is defined by the presence of the formin homology 2 (FH2) domain [109], and is further classified in terms of the presence and arrangement of additional domains [110, 111]. Most eukaryocytic cells contain proteins with a great diversity of FH2 domains. Each such domain consists of approximately 400 amino acids that directly control how actin is modified by the formin; the FH1 domain affects the function of the FH2 domain by binding to profilin [109, 112, 113]. Thus, the FH1 domain binds profilin, which is required for actin chain elongation by diaphanous (Dia) (Fig. 1) [114–116].

Phylogenetic analyses of the FH2 domain have shown that mouse formins can be divided into seven subfamilies, as follows [109]: Dia, dishevelled-associated activator of morphogenesis (DAAM), formin-related gene in leukocytes (FRL), formin homology domain-containing protein (FHOD), inverted formin (INF), formin (FMN), and delphilin.

The mDia subfamily contains three isoforms termed mDia-1 (also known as Diap-1), mDia-2 and mDia-3 (also known as Diap-2) [111, 117]. mDia proteins contain an RBD/FH3 sequence in their N-terminal regions, which in turn contain a Rho-binding domain (RBD), four Arm repeats (also termed the Dia-inhibitory domain, DID), a dimerization domain (DD), and a putative coil-coiled region (CC) [117, 118]. mDia-1 can be activated only by Rho (RhoA-C), whereas mDia-2 and mDia-3 can be also activated by Rac and Cdc42 [119, 120]. Members of the Dia protein family are key regulators of fundamental actin-driven cellular processes, which are conserved from yeast to humans [111]. Cellular studies have suggested that RhoA competes with the diaphanous auto-regulatory domain (DAD) to bind to the mDia-1 N-terminus, relieving the auto-inhibitory interaction, and thus enabling mDia-1 to influence actin dynamics [121–123].

Many studies have shown that mDia-1 is an Rho-regulated actin nucleator that acts downstream of RhoA [117, 118, 124]. It is a multimodular protein that interacts with numerous actin regulators, adapters and signaling components such as profilin-1 [111]. To date, at least 12 formins have been shown to interact with Rho family GTPases, the best-studied interaction is that between mDia-1 and RhoA [113, 125–127].

mDia-1 and angiogenesis

As emphasized above, VEGF and Ang-1 play essential (complementary) roles in vascular development during embryogenesis. VEGF is required for the formation of the initial vascular plexus early in development, and Ang-1 is necessary for subsequent vascular remodeling into mature blood vessels [128–130].

Gavard et al. [128] showed that Ang-1 counteracted VEGF-induced endothelial permeability by triggering the RhoA pathway and the consequent association of mDia with Src, thereby preventing activation of Src by VEGFR-2. Knockdown of endothelial RhoA by siRNA restored the VEGF-induced S665 VE-cadherin phosphorylation blocked by Ang-1, and knockdown of either mDia-1 or mDia-2 removed that protein from the heterocomplex and eliminated the ability of Ang-1 to counteract the VEGF-induced endothelial permeability of mouse endothelial cells. This proved that mDia was required downstream of Ang-1 to block VEGF-dependent permeability. Once activated, mDia-1 played a key role when Ang-1 controlled endothelial barrier function; expression of active mutant mDia-1 blocked VEGF-induced permeability. Thus, by limiting the access of Src to VEGFR-2, mDia-1 may restrict activation of the SFK-initiated pathway; ultimately controlling the interplay between Ang-1 and VEGF, and the biological outcome of this interaction.

mDia-1 and apoptosis

Kamasani et al. [131] showed that the Rho effector mDia-1 is a critical downstream player in farnesyl transferase inhibitor (FTI)-induced apoptosis. Dominant inhibition of mDia-1 action ablated FTI-induced apoptosis, but not actin reorganization or growth inhibition, the latter of which may be associated with a Rho effector kinase pathway interaction that downregulates c-Myc. In nude mice, dominant inhibition of mDia-1 promoted tumor formation and ablated the anti-tumor action of FTI. These findings suggest that the Rho/mDia-1 pathway plays a critical role in the cell death mechanism engaged by FTI, and that mDia-1 may be important in terms of the Rho-dependent survival of oncogenically transformed cells, perhaps influenced by oncogenic RhoGEF AKAP13/Lbc.

Profilin-1

To the best of our knowledge, actin is a highly dynamic protein network containing many actin-associated proteins [132, 133]. Of these, one key regulatory protein, profilin-1, binds to actin monomers in the skeletal body (at the barbed ends), and contributes to many biological activities by assembling and disassembling actin filaments [134–136]. Profilin-1 consists of 140 amino acids, and has a molecular weight of 12 – 15 kDa. It is a ubiquitously expressed protein that binds to G-actin [137–139] and is associated with many cellular activities ranging from control of actin polymerization to gene transcription [23].

Traditionally, profilin-1 has been considered an essential control element for actin polymerization and cell migration. Originally identified as an actin-sequestering protein that formed a 1:1 complex with G-actin, it was thought to prevent actin polymerization (Fig. 1) [140]. However, subsequent studies showed that it promotes actin polymerization by catalyzing the exchange of actin-bound ADP for ATP and transporting ATP-G-actin to the barbed end of actin [141, 142]. In addition, it liberates actin monomers from the sequestering protein thymosin-β4 [24], thus contributing indispensably to both physiological and pathological cell proliferation and migration [143]. In summary, profilin-1 has a dual effect on actin polymerization, depending on its concentration relative to those of G-actin and the free barbed ends of actin filaments [23].

In general, profilin-1 binds strongly to three major classes of ligands; in order of strength, these are: actin monomers [144], phosphatidylinositol 4, 5-bisphosphate (PIP2) [145] and proteins containing poly L-proline (PLP) (including vasodilator-stimulated phosphoprotein, or VASP; Wiskott-Aldrich Syndrome Protein, or WASP; and Dia) [116, 146, 147]. Notably, profilin-1 binds to (and regulates the action of) retinal cadherin (R-cadherin), downstream of the Rho GTPases [148].

Based on in vivo experiments with transgenic mice overexpression of profilin-1 in smooth muscle cells increases actin polymerization and subsequently activates the Rho/ROCK pathway [149]. In another work, ROCK and Dia-1 together mediated actin polymerization by regulating the activity of profilin-1 [150]. RhoA controls the actions of R-cadherin, a member of the classical cadherin family, through the Dia-1/profilin-1 signaling pathway [148].

Profilin-1 and the endothelium

Romeo et al. [151] showed that profilin-1 acted downstream of low-density lipoprotein (LDL) to mediate diabetic MVED. Profilin-1 overexpressed in rat aortic endothelial cells triggered three indicators of endothelial dysfunction: an increase in apoptosis, elevated expression of intracellular adhesion molecule 1 (ICAM-1), and decreased phosphorylation of VASP (a marker for NO signaling). In addition, loss of profilin-1 was associated with reduced cell-cell adhesion and inhibition of cell migration. Furthermore, such loss inhibited cell growth without compromising cell survival, at least in the short term, thus suggesting that profilin-1 plays an important role in endothelial proliferation. In another study, silencing of profilin-1 expression suppressed the matrigel-induced early cord morphogenesis of endothelial cells [20].

Profilin-1 and angiogenesis

Fan et al. [22] recently showed that VEGF-A-inducible phosphorylation of profilin-1 at Tyr 129 was critical in terms of endothelial cell migration and angiogenesis. Chemotactic activation of VEGFR-2 and Src induced profilin-1 phosphorylation at the leading edge of the cell, promoting the binding of profilin-1 to actin and actin polymerization. Subsequently, use of a conditional endothelial knockin of phosphorylation-deficient profilin-1Y129F in mice revealed that profilin-1 phosphorylation was critical to allow angiogenesis after wounding and ischemic injury, but not developmental angiogenesis. Thus, the VEGFR-2/Src-mediated phosphorylation of profilin-1 bypasses canonical, multistep, intracellular signaling events to initiate endothelial cell migration and angiogenesis in some other manner, and may serve as a highly selective and nontoxic target of therapeutic interventions seeking to minimize pathological angiogenesis.

Profilin-1 and apoptosis

Of all conditions in which profilin-1 affects apoptosis, breast cancer has received the most attention. Yao et al. [152] found that stable expression of ectopic profilin-1 sensitized the breast cancer cell line MDA-MB-468 to apoptosis. Thus, profilin-1, which functions primarily to promote the formation of local superstructures from actin filaments and integrin, may contribute to the promotion of apoptosis. A previously unknown activity of profilin-1 was discovered; the protein mediates staurosporine (STS)-induced apoptosis in breast cancer cells by upregulating integrin α5β1 synthesis, presenting a new target for breast cancer therapy. A subsequent study [153] showed that profilin-1 overexpression sensitized cancer cells to apoptosis of the typical intrinsic mitochondrial pathway triggered by STS. Again, this revealed a new function/action of profilin-1: it combines synergistically with apoptotic agents to increase apoptosis.

Conclusion

In summary, we have explored the roles played by the RhoA/mDia-1/profilin-1 signaling pathway during MVED progression in DR. Many studies have shown that endothelial function is critical in terms of eye homeostasis, and its dysfunction is closely associated with DR. Angiogenesis and apoptosis are the most common end-stage symptoms of advanced retinal MVED caused by the hyperglycemia of DR. However, DR is much more than simple chronic hyperglycemia in one eye. The understanding and management of MVED is a major focus of research seeking to prevent microvascular complications associated with all stages of DR. Our review of angiogenesis and apoptosis during DR, together with previous studies, reinforces the concept that MVED predisposes toward DR.