Update on the Effects of Antioxidants on Diabetic Retinopathy: In Vitro Experiments, Animal Studies and Clinical Trials

Current therapies for diabetic retinopathy (DR) incorporate blood glucose and blood pressure control, vitrectomy, photocoagulation, and intravitreal injections of anti-vascular endothelial growth factors or corticosteroids. Nonetheless, these techniques have not been demonstrated to completely stop the evolution of this disorder. The pathophysiology of DR is not fully known, but there is more and more evidence indicating that oxidative stress is an important mechanism in the progression of DR. In this sense, antioxidants have been suggested as a possible therapy to reduce the complications of DR. In this review we aim to assemble updated information in relation to in vitro experiments, animal studies and clinical trials dealing with the effect of the antioxidants on DR.


Introduction
Diabetic retinopathy (DR) is one of the most frequent causes of blindness in the adult and elderly population worldwide [1]. Control of blood glucose, blood pressure and lipidemia have been shown to alleviate the appearance and evolution of DR [2]. Nonetheless, some patients might exhibit a progression of DR with appropriate blood glucose and pressure control.
DR is primarily a microvascular disorder related to the loss of pericytes, endothelial cell proliferation, the disruption of tight junctions between endothelial cells, the thickening of the basement The results obtained so far have allowed us to conclude that several of the microvascular changes observed in DR can be prevented or mitigated by different antioxidant medicinal derivatives of plant origin. The properties of resveratrol-a powerful antioxidant, which is anti-inflammatory and neuroprotective-are widely known in in vitro studies, which have subsequently been confirmed in different clinical trials [5]. Other flavonoid polyphenols (e.g., naringin, anthocyanins) and non-flavonoids (e.g., curcumin), are characterised by a group of natural active compounds, of which the antioxidant and anti-inflammatory capacity was verified in human cell models of retinal endothelial cells (HREC) and retinal pigment epithelial cells , under conditions of sustained high concentrations of glucose, mimicking a diabetic environment [18,19].
Non-provitamin A carotenoids (lutein, zeaxanthin, lycopene, and astaxanthin) were shown to be effective, not only in reducing in vitro inflammation, but also in modulating gene expression, eliminating ROS as a consequence [20]. Furthermore, there is strong evidence for xanthophyl carotenoids, such as lutein and zeaxanthin, which are selectively absorbed by RPE cells, and accumulate preferentially in the macular region, conferring protection on photoreceptors. Recently, it was found that this beneficial effect on the retina can be increased synergistically when lutein is administered in association with Ω-3 long-chain polyunsaturated fatty acids. An antioxidant and anti-inflammatory effect were found by reducing ROS production and inhibiting the expression of inflammatory mediators [21]. In this sense, the fish oil emulsion, rich in Ω-3 polyunsaturated fatty acids, has anti-inflammatory and antioxidant in vitro properties, with the capability to inhibit the production of pro-inflammatory growth factors and cytokines [22]. Table 1 summarizes the main in vitro studies on antioxidants and their results as possible therapeutic agents for DR treatment . Table 1. In vitro studies on diabetic retinopathy (DR) and hyperglycemic conditions (↑ = increase of; ↓ = decrease of).

Antioxidant(s) Studied. Culture Cell Line Type Main Outcomes Reference
Alpha-linolenic acid (ALA), zinc and linoleic acid Choroid-retina endothelial cells (monkey) Modulation of endothelial proliferation. ALA ↓ reactive oxygen species (ROS) production and vascular endothelial growth factors (VEGF) secretion, and ↑ superoxide dismutase (SOD) activity.

Animal Studies
There is an increasing body of evidence about the role of antioxidants in the control of diabetic retinopathy in animal models (Table 2)  .
Vitamins have a determining task in the control of the oxidative cascade involved in the development of diabetic retinopathy [102][103][104][105]. Vitamins C (ascorbic acid) and E [α-tocopherol] have been shown to have the ability of avoiding abnormalities of ocular blood hemodynamics and leukostasis in streptozotocin-induced diabetic rats [102,103]. Moreover, they prevent the advancement of acellular capillaries [102][103][104][105].
Combinations of vitamins have a synergistic effect. In this manner, vitamin E + selenium and taurine diminish retinal conjugated dienes (CD) in the early stage of the diabetic disease. This combination can also reduce lipid hydroperoxides (LP), but with a 500 IU dose of vitamin E [97].
In addition, vitamins C and E, supplemented with other antioxidants (Trolox, N-acetyl cysteine, β-carotene and selenium), improve the survival of retinal cells more obviously. This combination prevents a decrease in SOD, glutathione reductase (GR) and catalase, and lowers the activation of retinal protein kinase C and diminishes lipid peroxidation indicators. [102,103].
Trolox is an analog of vitamin E that allows the regeneration of diminished pericytes in retinal vessels in diabetic mice [33].
Age-related eye disease study (AREDS)-based micronutrients (ascorbic acid, vitamin E, β-carotene, zinc, and copper) diminish oxidative and nitrative damage to retinas in diabetic rats and prevent the development of retinal acellular capillaries [52].
Nicanartine (also known as l-carnitine) is responsible for transporting fatty acids into the mitochondria. It also acts as an antioxidant and lipid-lowering compound. Nicarnitine slows down pericyte loss, but has no effect on the formation of capillaries or micro-aneurysms in diabetic rats [86]. Another study showed that l-carnitine can reduce glycemic levels. Administrated daily, it has an important role as an antioxidant factor and avoids protein degradation [77].
High levels of VEGF protein concentration are observed in diabetic rats if compared with control non diabetic rats. Taurine and α-lipoic acid alleviate VEGF levels in diabetic rats [36] and α-lipoic acid maintains a constant number of pericytes and foil the formation of acellular capillaries [54,55,57]. These molecules obtain the inhibition of lipid peroxidation, increase glutathione peroxidase and produce an activation of AMP-activated protein kinase (AMPK) [57]. The normalization of nuclear transcriptional factor and the restoration of antioxidant defences are observed in the retinas of diabetic rats [54][55][56][57]. Moreover, oral supplementation with lipoic acid allows one to re-establish the electroretinogram b-wave amplitude of diabetic animals, to control values [56].
It has been known for a long time that food plays a crucial function in the control of diabetes. However, it is not only necessary to avoid foods with a high glycemic index. Several studies are being carried out to prove the beneficial effect for the control of pathogenesis of structural damage caused by diabetes.
Hydroxytyrosol is the main polyphenol compound present in olive oil. It has a neuroprotective effect on the retina. In diabetic rats treated with hydroxityroxol, the total retinal thickness and the cellular size were similar to non-diabetic animals. Retinal ganglion cell loss is smaller in rats treated with hydroxytyrosol, compared with non-treated diabetic rats [76].
Curcumin, a natural yellow pigment used commonly to color foods and cosmetics, decreases oxidative stress, diminishes levels of VEGF, NF-KB AND IL-1 β, and seems to have an anti-inflammatory effect on DR [68,69].
Resveratrol is present in the skin of grapes, blueberries, raspberries and blackberries. It is also found in high concentrations in red wine.   A daily dose of Resveratrol suppresses endothelial nitric oxide synthase production [89], the expression of oxidative biomarkers and superoxide dismutase capacity in the retina and blood of diabetic rats [90]. Furthermore, it improves the expression of the PON1 gene that regulates inflammatory response and microvascular complications in diabetes [91] and can restore the transcription of the proteins of the retinoic acid metabolism pathway [92]. Resveratrol coated gold nanoparticles redevelop the equilibrium between the stimulators and inhibitors of antiangiogenesis, increasing the retinal expression of PEDF and decreasing VEGF-1. Retinal expressions of TNFα, IL-1 β, MCP-1, ICAM-1 and IL-6 are also diminished [91][92][93]. Caffeic acid hexyl (CAF6) and Dodecyl (CAF12) amides also have a neuroprotective effect [57,60].
Carotenoids, such as lutein or zeaxanthin, also seem to have protective effects on DR in diabetic rats. Zeaxanthin inhibits oxidative damage and performs anti-inflammatory action [106]. Plus, lutein normalizes the markers of retinal oxidative stress, avoids ganglion cell loss and decreases apoptosis markers (caspase-3) [79]. Furthermore, it attenuates the loss of ganglion cells and prevents impairment in ERGs [79].
Tempol, a superoxide dismutase (SOD) mimetic and pleiotropic intracellular antioxidant, can neutralize the excess of superoxide radical expression in DR. Tempol protects cells of the vessels, especially ameliorating hemodynamics in the retina of diabetic rats [100]. Moreover, tempol avoids the gathering of fibronectin and glial fibrillary acidic protein in a diabetic murine model [100].
Eriodictyol, a flavonoid obtained from California Yerba Santa, preserves the blood-retinal barrier, decreasing pro-oxidative biomarkers as well [72]. Rutin has shown antiapoptotic activity (decreasing levels of caspase-3) [94]. It has activity at the systemic level, decreasing glucose levels and improving insulin concentration. Plus, other flavonoids with beneficial effects are naningerin [85] and taxifolin, a flavonone found in onion, milk thistle (carduus marianus), douglas fir bark and frech maritime (pinus pinaster) fir bark. Taxifolin is useful to prevent diabetic retinal damage. It can normalize the Il-6, TNF, IL1 and tGSH levels in blood serum [98]. Other flavonoids such as hesperetine [75] have been studied, obtaining promising results.
Green tea, rich in polyphenols, decreases the number of pericyte ghosts and acellular capillaries and forces a lower concentration of VEFG and TNF-α in the retina of animal models [73,74].
Melatonin is a strong antioxidant naturally secreted by the pineal gland. When orally administered, it presents the amelioration of cytokine, nitrotyrosine, and malondialdehyde concentrations, [80] and depletion in concentrations of VEGF, MMP9, and oxidation protein products (AOPP) [81]. Merhazadi et al. also found that oral melatonin administered to diabetic rats decreased retinal cell size and the fluorescein leakage of retinal vessels [82].

Clinical Studies
There are not as many clinical studies as in vitro or animal investigations. There is an increased interest in antioxidant supplementation for human disorders, because they can be administered orally and they are easily available and affordable. Difficulties in analyzing these studies are that end-points and results are variables and, in some cases, controversial. Some studies focus their objectives on functional results (best-corrected visual acuity-BCVA-; contrast sensitivity; glare sensitivity), others in anatomical results (central macular thickness-CMT-), in analyzing oxidative stress status, or even in a combination of the anterior parameters. Plus, in the majority of these studies, the population studied include type 2 diabetic mellitus (T2DM) patients. Some studies deal with the first stages of the disease and other studies consider typical complications of late stages. Furthermore, as previously noted, tangled and not well-known mechanisms in the pathogenesis of diabetic retinopathy make early and new combined therapies of antioxidants desirable. The most relevant studies are listed in Table 3.    Table 3 shows three studies with ALA. Single administrations in 467 T2DM patients over the course of two years found no results preventing diabetic macular edema (DME) or improving BCVA. [107,108]. ALA combined with genistein and vitamins C, E and B in 32 pre-retinophatic patients over the course of 30 days showed some increases in plasma antioxidant levels and eletroretinogram (ERG) oscillatory potential values [119].
Calcium dobesilate (Ca Dob) is a vasculoprotector agent that has been widely studied in the context of DR. [138]. However, the results of clinical trials in DR are controversial. Farsa [139] described some Ca Dob mechanism of actions, such as antioxidant and anti-ROS activity, defensive activity at the endothelial level, antiapoptoic properties and angiogenesis inhibition. Other mechanisms of Ca Dob improving microcirculation and reducing micro-vascular injury have been described recently [140]. There are nine studies listed with Ca Dob supplementation ( Table 3). The study with the highest number of patients recruited (635) and the longest duration showed no effects on the reduction of the development of clinically significant diabetic macular edema (CSDME) in T2DM [116]. Likewise, another study with 40 non-proliferative DR (NPDR) patients with macular edema failed to find beneficial effects by adding Ca Dob to laser in the treatment of macular edema [117]. However, other studies have positive results analyzing microaneurisms and retinal hemorrhages [115], exudates [113], and blood viscosity [111,112].
There is an increasing interest in combined antioxidant therapy (CAT), as shown by the latest investigations. A long-lasting study (60 months) by Garcia-Medina et al. [118] evaluated antioxidant supplementation (lutein, VC, alpha-tocopherol, niacin, beta-carotene, Zn and Se) in 105 T2DM with NPDR. This long follow-up allowed one to demonstrate a retardation of the DR progression and the maintenance of antioxidant plasma status in the treated group. However, no effect on visual acuity was reported. In another work, the supplementation of lutein and zeaxanthin for three months in type 1 diabetic mellitus (T1DM) and T2DM showed an improvement of visual acuity, contrast sensivity and foveal thickness [127]. Likewise, Chous et al. reported an improvement in BCVA, but no changes in retinal thickness in his study. They administered a DiVFuSS formula that consists of vitamins C, D3 and E (d-α tocopherol), zinc oxide, eicosapentaenoic acid, docosahexaenoic acid, α-lipoic acid (racemic mixture), coenzyme Q10, mixed tocotrienols/tocopherols, zeaxanthin, lutein, benfotiamine, N-acetyl cysteine, grape seed extract, resveratrol, turmeric root extract, green tea leaf, and Pycnogenol [122].
Other studies focus their results on the interpretation of oxidative/nitrosative stress and antioxidant defenses, such as Rodriguez-Carrizalez in a trial with 60 T2DM patients over the course of months [123]. Furthermore, Roig Revert et al. studied 208 T2DM for 18 months and found a reduction of MDA and a significant increment of total antioxidant status (TAS), a reduction of MDA and a significant increase of TAS in the treated retinopathy group [121].
The association of CAT with the latest therapies such as anti-VEFG therapy is very interesting. Lafuente et al. [124] studied 55 T2DM for three years. They measured the effect of adding a combined antioxidant treatment to ranibizumab in the treatment of diabetic macular edema (DME). They reported lower macular thickness in the supplement group when compared to the control group.
The antioxidants, anti-inflammatory and neuroprotective effects of crocin have also been studied by Sepahi et al., in 60 DM pattens with refractive DME showing promising results [125].
Ginkgo biloba supplementation has been tested by Huang in 25 T2DM patients for three months. They found a decrease in MDA and fibrinogen levels and an improvement in other blood parameters and retinal blood flow rate [126].
Pycnogenol ® , a French maritime pine bark extract rich on flavonoids, has been dosed alone or combined since the 1980s, in several studies. Two studies administering this antioxidant alone (Table 3) are referenced. One is a study of 32 diabetic patients for six months, five of which displayed a reduction in exudates in both eyes [113]. The other study reported the functional and anatomic results of 46 T2DM for three months [131] Domanico et al. have studied patients with NPDR. Patients were divided into two groups: receiving supplementation or placebo. The treatment was made up of an antioxidant combination (pycnogenol, vitamin E and coenzyme Q10). Treated group showed that central macular thickness was significantly reduced during the study period (six months) [120].
In 2020, Ali et al. published an exploratory, two-month investigation, to evaluate the efficacy and safety of resveratrol in T1DM patients. A total of 13 patients were studied. Between the baseline and the endpoint resveratrol, treatment was associated with a significant decrease in the level of MDA and a significant increase in the level of total antioxidant capacity. There was no change in the serum levels of TNF-alpha and IL-1beta [132].
Tabatabaei-Malazy et al. reviewed 10 observational studies that reported lower vitamin C levels in subjects with DR compared to those without DR [141]. Park et al. studied the association of vitreous vitamin C depletion with diabetic macular ischemia in proliferative diabetic retinopathy (PDR). This is a study with 40 patients, 20 with PDR and 20 with idiopathic epiretinal membrane (control group), that underwent a pars plane vitrectomy. They found that a vitreous level of vitamin C in PDR patients showed a decrease, that was correlated with the degree of macular ischemia. Levels of vitamin C in the serum, aqueous humor and vitreous humor were lower in patients with PDR than in the control group. However, there was no correlation among the serum, aqueous humor and vitreous humor levels of vitamin C in the PDR group. These findings suggest that ocular factors could be relevant in the pathogenesis of PDR and open new strategies and routes of administration of the treatment [133].
Two clinical trials administering vitamin E (α-tocopherol) are included (Table 3). Bursell et al. studied high-dose vitamin E supplementation in a T1DM group and in a control group. The authors reported that diabetic patients showed a decreased retinal blood flow that improved after treatment, similarly to control cases [134]. Chatziralli et al. administered 300 mg of vitamin D to 282 insulin dependent T2DM with DR. They found a significant diminution of MDA in all DR stages [135].
Vitamin E supplementation was compared to ALA or selenium supplementation or placebo, in a study performed by Kähler in 80 diabetics during three months. They found that oxidative status was reduced in all treated groups compared to the placebo group [108].
Finally, a recent study by Kheirour analyzed serum VEFG levels with zinc supplementation. The authors did not observe any changes after three months of treatment [137].

Conclusions
The studies examined in this survey suggest that a number of antioxidants may prevent or alleviate the deleterious effects of hyperglycaemic environment in cultured tissues, or the complications of DR in animal models of diabetes. The results in human studies are more controversial, but some of them are quite promising, particularly when combinations of antioxidants are considered. It is necessary to keep in mind that studies in humans cannot be compared to those performed in cultured tissues with a perfect control of the environment or performed in laboratory animals in which diets and the rest of the living habits are supervised. In addition, metabolism, senescence rate or administered doses in animal models are not identical to those in human individuals [142].
The heterogenicity of the studies included in this review, in terms of antioxidant types, methods, doses and results, is remarkable. It is difficult to bring together all this information, but we can infer that there is a tendency to use antioxidants in combination. However, we have to keep in mind that pro-oxidative conditions and their consequences are still not well-known in a chronic disorder such as DR. Plus, a combination of antioxidant may interact not only with different metabolic pathways, but also among themselves. In this context, it is hard to define the exact effect of each antioxidant that may even change when administered in combination.
Another important concern is the safety of administering antioxidants for years, when the clinical studies performed so far are not longer than five years.
Despite the fact that encouraging outcomes have already been obtained in this field, future studies must be conducted considering the efficacy and safety of different doses of antioxidants, in order to get better results to prevent ocular damages in this blinding disease. Funding: This research received no external funding.

Conflicts of Interest:
The authors declare no conflict of interest.