DR affects patients with diagnosed or undiagnosed DM. The most significant risk factors for developing DR are older age, longer duration of DM, high glycemic levels, and hypertension (2). A higher HbA1C is significantly associated with advancement of DR, whereas intense glycemic control significantly correlates with a reduction in incidence and progression of DR (7, 8). Recent studies have shown high variability in glucose levels also contributes to severity of DR; therefore, controlling post-prandial glucose levels is important in reducing the progression of DR as well (9, 10). Tight blood pressure control is another important risk factor modification that has been shown to reduce the damage inflicted on the retina from DR (11). Other modifiable risk factors that are not as critical in DR but still contribute to the development and advancement of disease include: nephropathy, hyperlipidemia, smoking, and obesity (12, 13). Non-modifiable risk factors include pregnancy, age, and genetics (1, 14).

Despite the above risk factors, current studies find a variation between the development and severity of DR that cannot be fully explained by the beforementioned risk factors. Therefore, identifying newer biomarkers that can help stratify patients and determine their response to therapeutics is important (2). Systemic biomarkers that have been found to strongly correlate with severity of DR include markers for inflammation, such as C-reactive protein (CRP), homocysteine, and advanced glycation end products (AGEs) (15–17). Newer markers discovered include apolipoprotein, vitamin D, leptin, as well as various genetic markers (1). Ocular biomarkers from sampling vitreous and tears include VEGF and platelet-derived growth factor (PDGF), but these markers need more validation studies for predicting DR severity (18, 19).

DME is the most common cause of visual loss in DR patients occurring in both NPDR and PDR at various levels of disease (3, 20). DME results from leakage of fluid and proteins from retinal vessels into the macula due to breakdown of BRB (20). Hard exudates and edema can be seen on fundoscopy (21). Tractional retinal detachment (TRD) with vitreous hemorrhage is the second form of vision loss resulting from DR (20). Unlike DME, TRD only occurs in PDR due to the fibrovascular scarring generated during neovascularization being pulled on by the vitreous (3). For TRD, a ring of white scar tissue can be seen on fundoscopy (22). Fibrovascular proliferation can also cause vitreous hemorrhage, contributing to vision loss from PDR (20).

Endothelial cell to Mesenchymal Transition (EndMT) occurs during the various stages of NPDR and PDR (23). The transformation can contribute to both mechanisms of vision loss mentioned above. EndMT contributes to DME by disrupting the BRB, increasing vascular permeability due to the loss of endothelial characteristics (24). This disruption leads not only to leakage of fluid into the retina, but also of inflammatory mediators (24). EndMT contributes to fibrosis within the retina due to the gain of mesenchymal characteristics and eventually increases the likelihood of TRD during PDR (23).

There are several metabolic pathways involved in the pathogenesis of DR, such as the polyol, protein kinase c (PKC), and hexosamine pathways (25). Activation of these pathways cause upregulation of cytokines and growth factors that leads to increased vascular permeability and occlusion of blood vessels (1). Initially, the blood vessels within the retina dilate in response to the hyperglycemic insults to increase retinal metabolism (20).

Pericytes are specialized, contractile mesenchymal stem cells (MSCs) that provide vascular tone and perfusion pressure within the capillaries of the retina (26). Hyperglycemia causes apoptosis of pericyte and endothelial cells leads to microaneurysm formation, breakdown of BRB, and development of retinal ischemia (20). Insufficient circulation leads to upregulation of VEGF (27), the primary angiogenic factor that is associated with the vascular permeability and angiogenesis seen in DR (20).

The high metabolic rate of the neural retina requires the regulation of a distinct supply of specific nutrients by the BRB (28). It regulates the transport of ions, water, and nutrients and prevents immune cells and antibodies from passing into the retina tissue (29). The BRB includes the inner BRB (iBRB) and the outer BRB (oBRB) (30). The iBRB consists of endothelium, Müller cells, and pericytes (30). The capillaries of the iBRB provide oxygen, glucose, and other nutrients to neurons and prevent other molecules from entering the retina for protection (31). The oBRB is composed of retinal pigment epithelium (RPE) and functions to regulate the transport between the choriocapillaris and the retina (32). Various growth factors and cytokines that are upregulated in DR, such as VEGF, HIF-1, and IL-1 β, induce BRB breakdown through various mechanisms (33). Damage to the iBRB and oBRB leads to fluid and protein extravasation into the macula, as well as the degeneration of retinal capillaries, playing a critical role in the development of DME and severe visual loss in DR (32).

One of the advanced stages of DR, known as PDR, is characterized by neovascularization, as well as an increase in the presence of myofibroblasts (34). This process can lead to the development of fibrovascular tissue, which can result in traction on the retina, leading to TRD and eventual blindness (2). There is a switch at some point in PDR between neovascularization and fibrosis; this is known as the angio-fibrotic switch (34). As previously stated, VEGF is majorly responsible for neovascularization (20). Connective tissue growth factor (CTGF) is another factor upregulated during DR that contributes mainly to fibrosis within the eyes of patients (34). The decrease in VEGF and increase in CTGF is associated with the switch from angiogenesis to fibrosis in PDR ( Figure 1 ) (34). CTGF is a Transforming Growth Factor-Beta (TGF-β) effector that induces tissue fibrosis during PDR (35). In addition to fibrosis during PDR, TGF-β is also a main contributor to EndMT earlier on in DR (23).

As previously stated, the early vascular loss in DR is a major contributor to the clinical manifestations of DR, and endothelial cell and pericyte apoptosis is only partly responsible for this vascular loss (20). Another part of the vascular loss found in DR is the transformation of endothelial cells to mesenchymal cells through EndMT (23). As a consequence of EndMT, endothelial cells undergo a shift in their cell markers and phenotype, transitioning from their original characteristics to acquire the cell markers and phenotype typically associated with mesenchymal cells (36–38). This process leads to the breakdown of the BRB and contributes to the microvascular pathogenesis of DR (23).

In addition to optimizing blood glucose, lipids, and blood pressure in diabetic patients, there are various intraocular management strategies that have become standard practice in DR patients (39, 40). For instance, the management of DME has been significantly altered with the introduction of intravitreal anti-VEGF injection therapies, such as ranibizumab, bevacizumab, and aflibercept (2). Several randomized control trials since 2010 have shown these agents significantly reduce DME and improve vision (41–45). Aflibercept has been found to be the most efficacious agent in patients with initial poor visual acuity (46). Despite all the current research on anti-VEGF treatments, the optimal frequency of injections and total time for treatment courses is still unknown (2). The current practice is to administer multiple injections within the first year of treatment and then a gradual decrease in the following years to maintain remission (39). Intravitreal anti-VEGF injections have also been shown to benefit patients with PDR (47). These medicines do, however, have some limitations and adverse effects (20). The short half-life of these agents means that initially, bimonthly injections are necessary to ensure efficacy (20). Elevation of intraocular pressure, vitreous hemorrhage, and inflammation are uncommon adverse effects (20). A rare adverse effect due to the high number of injections necessary is endophthalmitis (20, 48). Additionally, high cost and poor patient adherence are of concern when it comes to this treatment modality (20). Currently, there is research being done on other various anti-angiogenic agents that inhibit vasoconstriction and vascular occlusion within retinal capillaries multiple angiogenic factors in addition to or other than VEGF, such as Squalamine and Nesvacumab, respectively (20, 49).

In cases of refractory DME that are unresponsive to anti-VEGF treatments, intravitreal corticosteroid injections can be used (50). These refractory cases are thought to be driven by the effects of multiple cytokines (20). Corticosteroids target inflammatory mediators known to contribute to the pathogenesis of DME (20). There have been multiple clinical trials done for DME treatment using triamcinolone acetonide, dexamethasone (DEX) intravitreal implant, and fluocinolone intravitreal implant (48, 51–53). The use of intravitreal corticosteroids has been associated with a lower number of injections, lower cost, and higher compliance (20). However, given the higher incidence of adverse effects, such as cataracts, glaucoma, and vitreous hemorrhage, and no proven benefit in PDR, intravitreal corticosteroid injections are considered second-line to anti-VEGF injections (20, 48, 51–53).

On the other hand, in patients that have progressed to PDR, panretinal laser photocoagulation (PRP) is considered first-line due to its effectiveness in reducing vision loss in this patient population, especially in patients with vitreous hemorrhage complications (54–56). Aflibercept, an anti-VEGF agent previously mentioned, has been shown to result in better visual outcomes and is a safer alternative to PRP in select patients (47). Due to laser-induced retinal damage from PRP, adverse effects include mild central visual acuity loss and reduced night vision (57). However, PRP remains important as an adjuvant and rescue therapy for PDR patients with high-risk complications by reducing the rate of severe visual loss and halting the progression of retinopathy (20). In modern times, there have been efforts to develop newer laser approaches that reduce side effects (20) e.g. the pattern scanning laser (PASCAL) (58), the subthreshold micropulse diode laser (D-MPL) (59), and the navigated laser system (NAVILAS) (60).

Other treatments for DR under clinical evaluation include cardiolipin inhibitors, mitochondria-specific antioxidants, as well as other classes of antioxidants (61–63). All these treatments for DR are focused on preventing late complications of the disease. An alternate approach being investigated currently is treating the root causes and early developments within DR (23). We researched the current literature on EndMT, an early pathological event in DR and potential treatments to prevent it and subsequent complications.

DR, a leading cause of visual impairment in adults, can be categorized into NPDR and PDR stages (20). EndMT occurs throughout these stages, contributing to increased vascular permeability and fibrovascular scarring (23). Endothelial cells within retinal capillaries help maintain the BRB, which consists of cells joined together to prevent certain substances from entering the retina from the circulatory system (80). EndMT disrupts the integrity of the BRB, permitting plasma and proteins to enter the retina, which can give rise to complications like DME (80, 81).

In DR, EndMT leads to thickening of the basement membrane and increased deposition of matrix proteins, contributing to the breakdown of the BRB (24, 82). Additionally, EndMT increases the number of myofibroblasts, which can lead to TRD because of an increase in fibrosis within the retina (79). Hyperglycemia, suppression of long non-coding RNAs (lncRNAs), upregulation of Notch2, and various circRNAs and miRNAs are all implicated in EndMT induction through TGF-β signaling in DR (76, 81). Overexpression of H19 and MEG3, effectively prevents EndMT (24, 75, 81). Inhibition of Notch2 and suppression of various circRNAs are also potential therapeutic targets for preventing EndMT in DR (76).

We conducted our own study on EndMT within DR. The goal of our study was to evaluate if there is a significant difference in the amount of EndMT at different stages of the disease. EndMT and fibroproliferative transformation were characterized by de novo cellular expression of α-SMA. We used Ins2Akita mice as an experimental model of type 1 diabetes. The mice eyes were dissected to retrieve their retinas. We then performed immunofluorescence (IF) on the retinas of 23 diabetic mice at 4-weeks (8 mice), 12-weeks (8 mice), and 32-weeks (7 mice) old. We also performed IF on the retinas of 9 non-diabetic mice that were 12-weeks-old to use as a control. The antibodies used to stain the retina slides were for α-SMA, CD31, and DAPI. These were used as markers for mesenchymal tissue, endothelial cells, and nuclear markers, respectively. Images of these slides were taken at 20x using fluorescence microscopy. The intensity of α-SMA was analyzed using Image J software, and statistical analysis was performed using One-way ANOVA Multiple Comparisons Test with a p-value <0.05 considered significant. An increase in α-SMA intensity correlated with an increase in EndMT. Our results indicated that there was a significant increase in EndMT in the retina of diabetic mice compared to the control group ( Figures 2A, C ). Moreover, EndMT was significantly increased in the 12-week-old diabetic mice when compared to the 4-week-old and 32-week-old diabetic mice ( Figures 2B , C ). . There was no significant difference between the amount of EndMT when comparing the 4-week-old diabetic mice with the 32-week-old diabetic mice ( Figure 2C ). The significant increase of EndMT, at 12-weeks in comparison to 4-weeks may be due to a temporal requirement for diabetes to initiate the process of EndMT. The significant decline of EndMT, at 32-weeks in comparison to 12-weeks, could be attributed to the loss of pericytes and endothelial cells by 24 weeks in DR (83). Pericytes are mesenchymal stem cells, so they contribute to the increase in mesenchymal cell population found with EndMT (83).

Our findings are consistent with the suggestion that EndMT plays a crucial role in the pathogenesis of microvascular dysfunction in DR. This transformation could lead to loss of endothelial cells, capillary degeneration, vascular leakage, and retinal hyperpermeability, leading to complications like DME and TRD. Our study provides valuable insight into the temporal changes of EndMT in the context of diabetic retinopathy and highlights the importance of further research in this area. Our research group has started investigating EndMT in postmortem human retina that were obtained from Georgia Eye Bank and processed to prepare paraffin-embedded sections. Slides from 3 diabetic human samples and 3 non-diabetic human samples were used in an IF experiment, with α-SMA as a mesenchymal marker, DAPI as a marker for the nucleus, and VEGFR2 marking endothelial cells. Our preliminary findings suggest that diabetic human retinas exhibit a higher amount of mesenchymal tissue compared to control groups as shown by a marked increase in α-SMA immunoreactivity in retinal vessels of diabetic human subjects compared to the non-diabetic group ( Figure 3 ). Further research is needed to confirm these results and to explore the temporal changes of EndMT in diabetic human retinas.