**4. Discussion**

Astaxanthin possesses unique chemical properties derived from its distinctive molecular structure, including two hydroxyl groups, two carbonyl groups, and 11 conjugated ethylenic double bonds. In contrast to other carotenoids, AST can be esterified, has a more polar nature, and displays greater antioxidant capacity, which could be explained by the presence of the hydroxyl and keto moieties on each ionone ring [**?** ]. AST was recently reported to significantly suppress oxidative stress and protect various cells from apoptosis [**?????** ]. In this study, we demonstrated that AST could protect photoreceptor cells from oxidative stress through activation of the PI3K/Akt/Nrf2 signaling pathway in high glucose environments, which further reduced apoptosis and could potentially prevent the progression of DR.

The correlation between oxidative stress and the development of DR has been previously demonstrated [**??????** ]. Increased oxidative stress and production of ROS cause tissue injury through peroxidation of lipids, carbohydrates, proteins, and DNA [**?** ], with the concomitant production of oxidative biomarkers, such as acrolein, carbonylated proteins, nitrotyrosine, and 8-OHdG [**? ?** ]. In this study, we confirmed the harmful effects of high glucose by showing an increase in ROS, oxidative biomarkers 8-OHdG, nitrotyrosine, and acrolein, as well as apoptosis. In addition, our results also demonstrated that the oxidative stress and consequent ROS-induced damage to DNA, lipids, and proteins could be ameliorated by the antioxidative effect of AST. The overall reduction in oxidative stress mediated by AST further led to decreased apoptosis of photoreceptor cells. Baccouche et al. [**?** ] found that short-term use of AST could reduce retinal damage in fat sand rats. In their study, while the serum glucose levels were similar in the AST-treated animals and the control group, the cellular damage in Muller cells and retinal ganglion cells was reduced by AST treatment through increased HO-1 production. Dong et al. [**?** ] also reported a protective effect of AST against the oxidative stress

induced by diabetes in retinal ganglion cells. Another study using streptozotocin-induced diabetic rats demonstrated the antioxidative e ffect of AST in the retina in a high glucose environment [**?** ]. In this study, after AST treatment, the retinal thickness was preserved not only in the inner but also in the outer retinal layers. The cellular damage in DR is not limited to retinal ganglion cells. Photoreceptor loss in DR due to increased apoptosis has also been reported [**?** ]. Our results confirm that the antiapoptotic effect of AST is not limited to ganglion cells and Muller cells but also extends to photoreceptors, indicating the potential universal benefits of AST treatment in DR.

The Nrf2–ARE pathway plays an important role in cellular resistance to oxidative stress [**???** ]. Nrf2 is a transcription factor that binds to the ARE and promotes the expression of Phase II enzymes [**?** ]. In the absence of oxidative damage, Nrf2 interacts with the chaperone Keap1; conversely, in an oxidant environment, Nrf2 dissociates from Keap1 and translocates to the nucleus in its activated form, where it binds to the ARE and induces the expression of Phase II enzymes, such as HO-1 and NQO1 [**?** ]. Several studies have demonstrated the beneficial e ffects of Nrf2 activation in DR. Fenofibrate was reported to activate Nrf2, increase HO-1 and NQO1 expression, and reduce oxidative damage in diabetic mice [**?** ]. Sulforaphane also exhibited similar antioxidative e ffects in streptozotocin-induced diabetic rats and high glucose-treated Muller cells through the activation of the Nrf2 pathway [**?** ]. Furthermore, activation of Nrf2, either through DJ-1 overexpression [**?** ] or through Nrf2 activators [**?** ], could provide additional benefits to improve retinal pericyte survival and reduce vascular endothelial growth factor-induced cell migration in human retinal microvascular endothelial cells, both involved in the development of DR. In the current study, we have further demonstrated that the Nrf2–ARE pathway can be activated by AST, thereby increasing the expression of HO-1 and NQO1, which in turn attenuates ROS-mediated and intracellular oxidative damage and further protects photoreceptor cells from high glucose-induced apoptosis.

AST can ameliorate cell apoptosis via more than one pathway, including ERK, NF-κB, and PI3K/Akt [**????** ]. Among these, the PI3K/Akt pathway is a prominent regulator of numerous proteins involved in cell survival through profound antioxidative and antiapoptotic actions [**?** ]. Studying apoptosis after burn in an animal model, Guo et al. demonstrated a dose-dependent e ffect of AST in increasing p-Akt and decreasing cleaved caspase-3 levels [**?** ]. In diabetic rats, the cognitive functions could also be preserved by AST through activation of the PI3K/Akt pathway and downregulation of caspase-3 expression [**?** ]. In RPE cells, Li et al. [**?** ] reported that oxidative stress was attenuated by AST. They further demonstrated that AST upregulates Nrf2 and Nrf2-related Phase II enzymes through activation of the PI3K/Akt pathway. Collectively, these pieces of evidence support the involvement of the PI3K/Akt pathway in the antiapoptotic e ffect of AST. Nevertheless, the cellular response to AST may di ffer among distinct cell types and with diverse environmental stimuli. In a hamster model of oral cancer, Kavitha et al. [**?** ] found downregulated PI3K and p-Akt in AST-treated animals, which further led to caspase-induced apoptosis. Therefore, we specifically focused here on the response of photoreceptor cells. Our study clearly demonstrated that the PI3K/Akt pathway was upregulated in the presence of AST and that downstream Nrf2 expression was enhanced along with an increased expression of HO-1 and NQO1, which eventually reduced caspase activity and ameliorated apoptosis in high glucose-treated photoreceptor cells. Both PI3K inhibitor (LY294002) and Nrf2 inhibitor (ML385) counteracted the protective e ffects of AST and attenuated the expression of HO-1 and NQO1, indicating the importance of PI3K/Akt/Nrf2 signaling in AST-mediated cellular protection.

Systemic AST supplementation could function in the eye and potentially protect the retina from various diseases by reducing oxidative stress. Oral AST supplementation has been used in previous randomized controlled trials and was found to suppress the aqueous vascular endothelial growth factor levels and peroxide production in humans [**? ?** ]. An earlier trial using AST in combination with lutein and other antioxidants also reported improved central visual function in patients with age-related macular degeneration [**?** ]. The results from our study, along with those from previous studies, imply the potential benefits of AST in DR. Taken together with the benefits of AST in increasing

serum insulin and glucose metabolism [**???** ], the role of oral AST supplementation in the prevention and treatment of DR warrants further evaluation in clinical trials.

There were a few limitations to our study. First, we treated the photoreceptor cells with AST before the high glucose administration; therefore, we evaluated only the effects of AST on the acute response of cells to high glucose but not on the cells that had already developed a certain degree of damage under the high glucose environment. Previous in vivo studies have demonstrated that AST administered after the induction of diabetes in different animal models may reduce the levels of oxidative stress mediators and preserve retinal function [**???** ]. These results indicate that AST may potentially be used for the treatment of high glucose-induced damage, but further studies are needed to confirm the cellular responses of AST administered after the occurrence of cellular damage. Second, we only used a fluorescent probe to detect ROS. Other methods, including spectrophotometry, chromatography, and electron spin resonance, have also been proposed to increase the sensitivity and specificity of ROS detection [**?** ]. Nevertheless, we used other ancillary tests to confirm our detection of ROS, and the results of JC-1 staining and the detection of oxidative stress mediators support our conclusions.
