Oxidative Stress in Age-Related CNS Degeneration: Modeling, Mechanism and Interventions

1. Introduction

Age-related central nervous system (CNS) degeneration has become a major global health challenge, driven by an aging population, increasing healthcare costs, increased caregiving demands, and limited treatment options. Within the CNS, including the brain and eyes, oxidative stress begins accumulating early in life [1]. As aging progresses, additional stressors such as lifestyle factors and mitochondrial dysfunction can overwhelm cellular defense systems, leading to pathology. Three prominent aging theories highlight the importance of oxidative stress in the aging process: (1) the Free Radical Theory [2], which suggests that aging results from the excessive accumulation of reactive oxygen species (ROS) and reactive nitrogen species (RNS), which promote free radical reactions and cause oxidative damage to cells and tissues; (2) the Mitochondrial Theory [3], which posits that aging is driven by the gradual buildup of mitochondrial dysfunction and mitochondrial ROS accumulation; and, the epigenetic Information Theory of Aging, [4] that we proposed in 2023, which states that aging is driven by the progressive loss of youthful epigenetic information. The link to oxidative stress is raised by our recent work and others which revealed that the epigenetic regulation of oxidative-resilience genes, including GSTA4 [5], GSTM1, GSTM5 [6], becomes disrupted with aging in the eye, a developmental extension of the CNS. CNS cells, particularly dopaminergic neurons in the brain and the retina and RPE in the eye, are particularly vulnerable to oxidative stress due to their high oxygen consumption and lipid-rich content [7]. This intrinsic vulnerability positions oxidative stress as a central driver of age-related neurodegenerative disorders in the brain and eye.

Various oxidative insults have been developed for animal and cell culture models and are sometimes used interchangeably, despite inducing oxidative injury through distinct mechanisms. Yet, comprehensive overviews detailing the mechanisms underlying these different insults remain limited. Existing reviews also tend to focus on either the eye or the brain, despite their shared biology and the potential for experimental and therapeutic insights to translate across these CNS tissues.

Here, we review (i) environmental and lifestyle contributors to oxidative burden (Figure 1), (ii) the molecular mechanisms underlying common oxidative insults (Table 1), and (iii) antioxidant interventions related to CNS age-related disease (Table 2). This framework should guide the selection of appropriate experimental models and inform strategies to counter oxidative damage across a broad spectrum of age-related diseases.

2. In Vivo Models of Oxidative Stress: Environmental and Lifestyle Insults

2.1. High-Fat Diet or Obesity

According to 2022 WHO data, 1 in 8 people worldwide were living with obesity. Excessive fat accumulation in obesity increases mortality from non-communicable diseases, including cardiovascular diseases, diabetes and neurological disorders such as Parkinson’s disease (PD) and Alzheimer’s disease (AD). A key but often overlooked contributor to obesity-related pathology is the oxidative stress driven by lipid accumulation [8]. A recent systematic review further links high-fat diet (HFD) to retinal oxidative stress, inflammation, and neurodegeneration, supporting lipotoxicity as a systemic driver of ocular and CNS aging [9]. Human epidemiological and genetic evidence associates high-fat dietary patterns with increased glaucoma risk, whereas antioxidant-rich diets appear protective, supporting oxidative-stress-mediated retinal ganglion cell (RGC) vulnerability as a plausible mechanism [10].

A HFD increases intestinal chylomicron production, generating free fatty acids (FFAs) that are subsequently taken up by the liver. In hepatocytes, FFAs undergo mitochondrial β-oxidation and feed into the tricarboxylic acid (TCA) cycle [11]. The resulting increase in electron flux through the respiratory chain promotes mitochondrial ROS production and lipid peroxidation, including the generation of 4-hydroxy-2-nonenal (4-HNE) and malondialdehyde (MDA). These processes impair ATP production and mitochondrial respiration, ultimately contributing to mitochondrial dysfunction [12]. Excess ROS and lipotoxicity can then initiate a broader stress response by upregulating NADPH oxidase (NOX) subunits, inducing ER stress, disrupting Ca2+ homeostasis, and depleting antioxidant defenses, thereby establishing a self-amplifying cycle of oxidative damage [13]. FFAs can also be incorporated into triglycerides and very low-density lipoproteins (VLDL) [11]. When oxidized to oxLDL, lipoproteins further exacerbate oxidative stress (see oxLDL section below). Collectively, HFD-induced oxidative stress is initiated by excess lipid flux and mitochondrial β-oxidation and subsequently amplified by mitochondrial ROS production and lipotoxicity (Figure 2).

HFD is therefore a useful model for investigating how dietary lipid excess contributes to CNS and retinal degeneration. For example, an HFD providing 60% of calories from fat for 2 months can induce glial-cell senescence, contributing to neuroinflammation and cognitive decline [14]. Similarly, HFD regimens containing 60% kcal fat for 3 months or 45% kcal fat for 12 months induce oxidative stress in the retinal pigment epithelium (RPE), accompanied by inflammation and features of age-related macular degeneration (AMD) [15,16].

2.2. Radiation

Radiation is the emission or transmission of energy that can either induce atomic and molecular motion that produces heat (non-ionizing radiation) or remove tightly bound electrons from atoms to generate ions (ionizing radiation). Humans have real-world exposure to either of these. Non-ionizing radiation spans a broad range of wavelengths, including ultraviolet (UV), visible, near-infrared, and infrared light. Beyond our planet, astronauts encounter significant ionizing radiation exposures (especially during solar events) on spacecraft and on the International Space Station, creating a unique exposure that maybe related to CNS alterations seen in Spaceflight-Associated Neuro-ocular Syndrome (SANS) [17,18]. Together, ionizing radiation is a potent and long-lasting oxidative stressor to the CNS.

How radiation damages cells have also been characterized. Ionizing radiation, from photons to heavy particles like protons or carbon ions, can damage CNS cells, which are particularly vulnerable, by generating free radicals that harm DNA and induce metabolic stress [19]. Data show that even low-dose (<1Gy) and short-duration (<1 week) exposure to charged particles can trigger persistent oxidative stress for weeks to months, leading to cognitive impairment [20]. In vitro, oxidative stress is elevated for up to 80 cell generations after 10 Gy of X-ray exposure [21].

Then in contrast, non-ionizing radiation generates ROS primarily through energy-transfer mechanisms rather than direct molecular ionization. Ultraviolet radiation, particularly UVB (290–320 nm), contains high-energy photons that promote ROS production through photosensitization mechanisms [22]. Infrared laser radiation can induce oxidative stress by directly converting triplets oxygen (³O₂) to singlet oxygen (¹O₂), a highly reactive form of ROS [23]. Visible light, especially blue light (415-455 nm), has been strongly linked to oxidative stress in the RPE in AMD [24]. This phototoxicity is driven in part by light-induced excitation of N-retinylidene-N-retinylethanolamine (A2E), a major fluorophore of lipofuscin, which generates ROS, particularly ¹O₂ [25,26].

These different wavelengths have consequently been used to model distinct forms of ocular oxidative injury. UVB irradiation (e.g., 35 mJ/cm² daily for 4 days in mice [27]) is a well-established model of photoaging; however, because most UVB is absorbed by anterior ocular structures such as the cornea and lens, its penetration to the retina is limited. In contrast, blue light reaches the retina more efficiently and is commonly used to model RPE phototoxicity and AMD-like damage (e.g., blue LED exposure for 3 days [28]). Laser exposure is also widely used to induce choroidal neovascularization as a model of exudative AMD [29]. However, this approach primarily disrupts Bruch’s membrane and triggers a strong inflammatory and angiogenic response, with oxidative stress arising largely as a secondary consequence.

Importantly, low levels of radiation and particularly of non-ionizing radiation (light) has also been shown to be protective. This has primarily been attributed to the induction of low or chronic levels of cellular stress, at levels the cell can recover from using endogenous repair and antioxidant mechanisms in mitochondria and anti-oxidation pathways, these repair mechanisms, building resilience against future insults. This is discussed further in relevant sections regarding therapeutic interventions below at Section 4.4.

2.3. Cigarette Smoke

According to the WHO, tobacco smoking kills more than 8 million people each year and up to half of its users. Smoking is a well-established risk factor for aging and age-related diseases such as AMD, AD, and PD, with oxidative stress playing a central role in their pathogenesis [30,31].

Cigarette smoke contains a complex mixture of nearly 5000 chemical oxidants [32], including quinone/hydroquinone complexes in the tar phase and reactive oxygen- and carbon-centered radicals in the gas phase [32]. Some of these oxidants enter the bloodstream, causing DNA damage, protein modifications, and lipid peroxidation, which subsequently increase ROS production in the CNS, deplete antioxidants, and trigger inflammation [32].

Smoking-related ROS may facilitate the formation of amyloid-β (Aβ) oligomers and abnormal tau phosphorylation-key hallmarks of AD neuropathology [30]. Genetic susceptibility amplifies this risk: a Finnish study found that apolipoprotein E (APOE) ε4 carriers who smoke heavily in midlife more than double their risk of developing AD [33]. Importantly, smoking cessation reduced AD risk, underscoring the reversibility of smoking-related neurotoxic effects [34].

Smoking also increases the risk of age-related eye diseases, including cataract and AMD [31,35]. This risk arises through multiple mechanisms: smoking-related ROS production, reduction of macular pigment and plasma levels of antioxidants, inadequate peripheral blood flow [36]. Chronic exposure to cigarette smoke leads to impaired cytoprotective mechanisms, mitochondrial dysfunction, oxidative damage and structural degeneration of the RPE and Bruch’s membrane in mice, as well as RPE cell apoptosis in in vitro models. These changes mirror the early features of AMD, especially mitochondrial dysfunction as well as proteostasis and autophagy impairment that leads to RPE dysfunction and cell death [37]. Consistent with gene–environment interactions in AMD, iPSC-RPE studies suggest that the CFH Y402H risk background alters RPE metabolism and mitochondrial responses to chronic cigarette smoke extract [38].

Smoking may also worsen glaucomatous damage through vascular and oxidative mechanisms. Greater cumulative smoking exposure, measured by pack-years, has been associated with lower optic nerve head vessel density and larger choroidal microvasculature dropout, both of which are biomarkers of glaucoma severity and progression [39]. Emerging evidence suggests that smoking may contribute more strongly to glaucoma progression than disease onset, while smoking cessation may help mitigate this risk [40].

Cigarette smoke can effectively generate an oxidative stress model. For instance, exposure to cigarette smoke in mice - 24 hours/day, 5 days/week for 3.5 months, or 5 hours/day, 5 days/week for 6 months - induces oxidative damage to the RPE, leading to AMD-like features [31,41]. In in vitro culture systems, researchers use cigarette smoke extract to establish oxidative stress models [42].

2.4. Air Pollution

WHO data reveal that nearly the entire global population (99%) is exposed to air pollution levels that exceed WHO guideline limits. This polluted air contains harmful substances that contribute to strokes, heart diseases, lung cancer, and both acute and chronic respiratory conditions. Oxidative stress is a primary mechanism underlying these health effects [43,44,45,46]. Key pollutants of significant public health concern include not only fine particulate matter (PM2.5), but also strong oxidants such as nitrogen dioxide (NO₂), ozone (O₃), hydroxyl radicals (OH^•), and transition metal-containing particles (e.g., copper and iron) [47]. These oxidants and metals can generate ROS within cells through redox reactions, while PM2.5 can also trigger inflammation via antigen-presenting pathways [48]. Air pollution (e.g., dust storms, PM2.5, O₃) thus represents a multifaceted environmental hazard for studying age-related CNS degeneration and chronic eye diseases [49,50].

2.5. Aging (Mitochondrial Aging, Lysosomal Aging and Inflammaging)

Aging is characterized by mitochondrial dysfunction, lysosomal dysfunction, and chronic low-grade inflammation, which together create a feedback loop that accelerates oxidative stress, cellular senescence, and age-related diseases. With advanced age, mitochondrial function declines partly due to the accumulation of mutations in mitochondrial DNA (mtDNA). These defects impair oxidative phosphorylation, reduce adenosine triphosphate (ATP) production, and increase ROS generation. Mitochondria also serve as key hubs for innate immune signaling. Mitochondrial injury can trigger the release of mtDNA and oxidized mtDNA into the cytoplasm, where they activate cGAS-STING signaling and inflammasome pathways [51]. In the retina, Ferrington and colleagues have provided strong human-donor and iPSC-RPE evidence that AMD is associated with RPE mitochondrial dysfunction, increased mtDNA damage, altered bioenergetics, and impaired responses to oxidative stress, supporting mitochondrial decline as a central driver of retinal aging and AMD [52].

Mitochondrial dynamics, particularly the balance between fission and fusion and mitochondrial transport, are also strongly linked to neurodegeneration in the eye and brain. Fission/fusion genes are identified as specific source mutations in Parkinson’s disease (e.g. parkin), autosomal dominant optic atrophy (e.g. OPA1), and Charcot-Marie-Tooth disease type 2A (e.g. mitofusin-2). In vitro in retinal ganglion cells, the balance between fission and fusion strongly regulates axon growth [53,54,55,56], and in vivo in animal models, the increased intraocular pressure associated with glaucoma risk directly increases mitochondrial fission and stasis [57], while restoring mitochondrial fusion or transport strongly promotes neuroprotection and axon regeneration [58,59,60].

Lysosomal function also deteriorates with age. Our recent longitudinal profiling of lysosomal metabolites in the brain across the lifespan revealed age-dependent accumulation of glycerophosphodiesters and the amino acid derivative cystine. These metabolites are also elevated in lysosomal storage disorders and can predict chronological age [61], raising the possibility that progressive lysosomal metabolite accumulation contributes to aging-associated cellular dysfunction.

Chronic inflammation [62] further amplifies oxidative stress through redox-sensitive pathways such as NF-κB. Although these interconnected processes are difficult to model independently, restoring mitochondrial function, limiting lysosomal metabolite accumulation, and suppressing inflammation may help counter aging and age-related CNS disease. For example, IL-11 inhibition has been shown to extend both health span and lifespan [63].

3. In Vitro Oxidative Stress Models: Mechanistic Insights

3.1. Oxidized Low-Density Lipoprotein (oxLDL)

The in vivo HFD model we discussed before is complemented in vitro by exposure to oxidized oxLDL. oxLDL is generated when free radicals attack the polyunsaturated fatty acids (PUFAs) in LDL, producing reactive lipid peroxides such as 4-HNE [64]. These products modify apolipoprotein B (apoB), converting native LDL into oxLDL, which is recognized by scavenger receptors including LOX-1, CD36, CD68, and SR-A. Because these receptors are differentially expressed across cell and tissue types, oxLDL can trigger distinct downstream responses.

Lipoprotein receptor 1 (LOX-1), highly expressed on vascular endothelial cells, plays a key role in atherosclerosis and related vascular pathology [65,66]. HNE–histidine adducts on oxLDL act as LOX-1 ligands and facilitate receptor-mediated uptake. LOX-1 activation triggers three major responses: (1) ROS production, through p66Shc phosphorylation and PKC-mediated NOX activation [67], generating O₂^•- and H₂O_2. (2) apoptosis, through activation of caspase-3 and caspase-9 and suppression of Bcl-2 and c-IAP-1 [68,69]; (3) inflammation, through NF-κB, NLRP3, and MAPK signaling, inducing VCAM-1, ICAM-1, MCP-1, and pro-inflammatory cytokines [69,70] (Figure 3). SR-A and CD36 are particularly abundant in macrophages [71], where they promote NLRP3 inflammasome activation, ROS production, and foam-cell formation. Their expression can be further reinforced through a positive-feedback SR-A/CD36/PKCα/PPARγ pathway [72]. oxLDL can also activate receptor tyrosine kinase (RTK) signaling, influencing cell proliferation, differentiation, migration, and apoptosis [73]. In RPE cells, OxLDL is also taken up in a CD36-dependent manner [74]. OxLDL is then transported to the lysosomes where it induces lysosomal stabilization and subsequent inflammasome. Inflammasome activation, in turn, leads to the release of activated IL-1ß, a potent inflammatory cytokine and/or pyroptosis.

Malek and colleagues helped establish the AMD relevance of this lipid-stress pathway by showing that cholesterol- and apoB-containing lipoproteins accumulate in Bruch’s membrane, basal deposits, and drusen, supporting mechanistic parallels between AMD and atherosclerosis [75]. Their later work further implicated nuclear-receptor pathways, including PPAR and LXR signaling, in regulating RPE lipid metabolism, inflammation, and AMD-like pathology [76]. Complementing this framework, Miller and colleagues showed that shifting RPE lipid fate toward autophagy/lipophagy and mitochondrial fatty-acid oxidation can reduce apolipoprotein secretion, lipofuscin accumulation, and senescence [77,78]. Together, these studies suggest that impaired RPE lipid handling converts oxLDL/lipoprotein stress into lysosomal dysfunction, inflammation, and drusen-associated oxidative injury. Together, these findings suggest that shifting RPE lipid handling from extracellular lipoprotein secretion toward intracellular lipid degradation and mitochondrial metabolism may reduce drusen-associated lipid burden and oxidative stress.

Overall, oxLDL provides a useful in vitro model of lipid-driven oxidative stress relevant to atherosclerosis, ischemic injury, and age-related CNS degeneration, including AMD in the eye [79] and AD in the brain [80].

3.2. Sodium Iodate (NaIO₃)

Sodium iodate (NaIO₃) was first found to induce fundus pigmentary changes and blindness in 1926 [81], and then developed as a model of oxidative injury because it can induce potent and rapid RPE-initiated retinal degeneration [82]. To investigate retinal degeneration in vivo, NaIO₃ can be given through a single intraperitoneal injection [83], orbital venous plexus injection [84], or tail vein injection in mice [85,86]. However, the extent of RPE damage and the type of cell death differ: apoptosis is induced at 20 mg/kg [86], while necroptosis is induced at 100 mg/kg [83]_. RPE death is followed by photoreceptors apoptosis, blood-retinal barrier breakdown, and retinal neurodegeneration, mirroring the pathological features seen in AMD [83,85,87,88]. In vitro, 2-5 mM NaIO₃ induces apoptosis in ARPE-19 cells, 5-10 mM induces necrosis, and ≥10 mM triggers ferroptosis, all after 24-48 hours of exposure [87,89,90,91,92].

Mechanistically, NaIO₃ oxidizes thiol-containing proteins and polyunsaturated fatty acids, depleting glutathione (GSH), increasing glutathione disulfide (GSSG), and generating lipid hydroperoxides [92]. It can also suppress GSH synthesis by inhibiting system Xc⁻ [93], further promoting ROS accumulation and lipid peroxidation. This oxidative stress activates multiple downstream pathways, including NOX-dependent ER stress and Ca²⁺ overload [89,92], ERK signaling and epithelial-to-mesenchymal transition [89,91,94], NRF2-related antioxidant responses and ferroptotic signaling [94], mitochondrial dysfunction driven in part by lipid peroxidation products such as 4-HNE [95], and NF-κB/NLRP3-mediated inflammation [96] (Figure 4).

Importantly, RPE susceptibility may be further enhanced by melanin-dependent conversion of glycine to glyoxylate, a process increased by NaIO₃ [97]. Glycine depletion and glyoxylate accumulation can promote Ca²⁺ influx, ROS production, lipid peroxidation, inflammatory cytokine release, and macrophage infiltration [98,99].

Overall, NaIO₃ is a robust model of RPE-centered oxidative injury in which direct thiol oxidation, glutathione depletion, lipid peroxidation, mitochondrial stress, and inflammatory signaling converge to drive dose-dependent apoptosis, necrosis, and ferroptosis.

3.3. Amyloid Beta (Aβ)

Amyloid β (Aβ) oligomer accumulation correlates with cognitive impairment, AD and AD-related retinopathy [100,101]. Aβ peptide is produced through the cleavage of amyloid precursor protein (APP) by β- and γ-secretases, with Aβ monomers aggregating into oligomers, protofibrils, and amyloid fibrils, a process accelerated by 4-HNE [102]. Among these species, soluble Aβ oligomers are considered particularly neurotoxic and strongly linked to oxidative injury.

Aβ oligomers promote oxidative stress through several mechanisms. First, they can insert into lipid bilayers and drive lipid peroxidation, generating products such as 4-HNE [103,104]. In turn, 4-HNE promotes formation of cytotoxic protofibrillar Aβ species, impairs Aβ clearance, disrupts mitochondrial and synaptic function, inhibits Na+/K+-ATPase, and elevates intracellular Ca²⁺ [105,106]. Second, Aβ oligomers engage cell-surface receptors, including NMDAR, p75NTR, and LilrB2/PirB, triggering Ca²⁺ dysregulation, wNOX activation, apoptosis, and synaptic damage [107,108,109,110]. Third, Aβ binds redox-active metals such as Cu, Fe, and Zn, promoting aggregation and ROS generation [111,112]. Fourth, Aβ can accumulate in mitochondria [113], where it disrupts cytochrome c oxidase assembly [114,115], interacts with ABAD [116], and cyclophilin D [117], and promotes mitochondrial permeability transition and mtROS production (Figure 5).

Experimentally, soluble Aβ42 oligomers are commonly used to model Aβ-induced neurotoxicity [118]. Because wild-type rodents do not readily develop amyloid plaques, transgenic models such as Tg2576, which overexpresses Swedish-mutant human APP, are widely used to study amyloidosis and test anti-Aβ therapies [119].

3.4. Iron and Iron-Induced Oxidative Species

Aging disrupts iron homeostasis, leading to reduced intestinal iron absorption and systemic iron deficiency while simultaneously promoting iron accumulation in tissues such as the brain and retina, where it contributes to oxidative stress and neurodegeneration [120,121].

Iron enters cells through transferrin-dependent and transferrin-independent pathways. Intracellular ferric iron (Fe³⁺) is reduced to ferrous iron (Fe²⁺) and released into the cytosolic labile iron pool [122]. Fe²⁺ can generate highly reactive hydroxyl radicals (HO•) through Fenton and related redox reactions, causing oxidative damage to lipids, proteins, and DNA and promoting ferroptosis (Figure 6) [123].

Iron overload is a feature of several age-related neurodegenerative diseases, including AMD, AD, and PD [124,125], and is therefore widely used to model ferroptotic injury. In vitro, commonly used approaches include ferric ammonium citrate (10–100 mg/L) [126], hemin (5-20 μM) [127], and FeCl₃ (10–500 μM) [128]. In vivo models include high-iron diets (14 g Fe/kg) [129], intracerebral iron injection [130]_, and intravenous iron-sucrose administration [131]. Ferroptosis can also be induced by impairing iron export, for example through hepcidin-mediated inhibition of ferroportin [132].

3.5. Other Oxidative Insults

Byproducts of lipid peroxidation, including 4-HNE and MDA, are widely used markers of oxidative stress. They also act as secondary mediators that amplify oxidative damage and can contribute to cell death. However, their high reactivity and instability limit their use as direct experimental oxidants. 4-HNE rapidly forms adducts with proteins and nucleic acids, making it difficult to quantify, whereas MDA is less reactive, less toxic, and more stable, and is therefore commonly used as a lipid peroxidation marker [133,134].

Hydrogen peroxide (H₂O₂) is widely used to model oxidative stress in age-related diseases [135]. Through Fenton and Haber–Weiss chemistry, H₂O₂ can generate highly reactive hydroxyl radicals. However, H₂O₂ is relatively unstable and rapidly decomposes into water and oxygen, causing rapid changes in effective concentration that can complicate experimental interpretation [136].

Tert-butyl hydroperoxide (t-BHP), a short-chain organic hydroperoxide, provides a more stable alternative to H₂O₂ [137]. t-BHP induces oxidative stress by generating peroxyl and alkoxyl radicals that initiate lipid peroxidation [138]; and by consuming reduced glutathione (GSH) during glutathione peroxidase-mediated detoxification [139]. It can be administered intracerebroventricularly to induce oxidative stress in vivo [140] and is also widely used in cell culture models [141].

1-Methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) is a lipophilic neurotoxin that readily crosses the blood–brain barrier. In the brain, MPTP is converted by astrocytes and other glial cells into 1-methyl-4-phenylpyridinium (MPP+) [142,143]. MPP+ is selectively taken up by dopaminergic neurons, accumulates in mitochondria, and inhibits respiratory complex I, reducing ATP production and increasing ROS generation [144,145]. MPTP is therefore a gold-standard model of dopaminergic neurodegeneration and Parkinson’s disease (PD) [144,146].

Oxidopamine (6-hydroxydopamine, 6-OHDA) is another selective dopaminergic neurotoxin for studying PD [147]. Unlike MPTP, 6-OHDA does not cross the blood–brain barrier and must be delivered by stereotaxic injection. It damages both dopaminergic and noradrenergic neurons, largely through rapid auto-oxidation and ROS generation [148].

N-methyl-D-aspartic acid (NMDA) is not a direct oxidant but induces oxidative stress through excitotoxic NMDA receptor activation. NMDA receptors are Ca²⁺-permeable glutamate-gated ion channels expressed throughout the CNS [149]. Sustained activation causes mitochondrial Ca²⁺ overload and ROS production [150]. Intravitreal NMDA induces excitotoxic degeneration of retinal ganglion and amacrine cells [151], while intrahippocampal injection induces neuronal degeneration in the brain [149,152].

Tau is likewise not a direct oxidant. Although physiological tau stabilizes microtubules, pathological aggregation into neurofibrillary tangles contributes to neuronal dysfunction in AD [153]. Tauopathy can promote oxidative stress by disrupting mitochondrial fission–fusion dynamics, activating NOX signaling, and interacting with redox-active copper to promote H₂O₂ generation [154,155].

4. Therapeutic Oxidant Resistant Interventions: Mechanisms and Clinical Outcomes

4.1. AREDS/AREDS2, Vitamin C, and Vitamin E

The Age-Related Eye Disease Study (AREDS), initiated in 1999, evaluated nutritional interventions for AMD and age-related cataract [156]. The original AREDS formulation contained vitamin C (VC), vitamin E (VE), β-carotene, zinc, and copper and was shown to reduce progression to advanced AMD and vision loss [157], leading to its use in patients with intermediate AMD [158]. Because β-carotene was associated with increased lung cancer risk, particularly in former smokers [159,160], AREDS2 replaced β-carotene with lutein and zeaxanthin. AREDS2 has been shown to reduce progression to advanced AMD and slow geographic atrophy progression toward the fovea [161,162,163].

The protective effects of AREDS/AREDS2 involve free-radical scavenging, support of endogenous antioxidant enzymes, and maintenance of retinal function (Figure 7). VC and VE are direct free-radical scavengers [164]. and have also shown potential benefit in PD [165] and AD [166] (Table 2). Lutein and zeaxanthin are concentrated in the macula, where they filter blue light and provide antioxidant protection [167,168]. Zinc and copper support antioxidant enzymes, including SOD, catalase, and metallothionein, with copper also included to prevent potential zinc-induced copper deficiency [156,169].

4.2. GSH, NAC, GlyNAC

Glutathione (GSH) is the most abundant intracellular antioxidant and a major defense against oxidative damage. GSH directly scavenges reactive species and serves as a reducing substrate for detoxification of lipid hydroperoxides [170]. Glutathione S-transferases (GSTs) further protect cells by conjugating GSH to electrophilic lipid-peroxidation products. In particular, GSTA4 efficiently conjugates 4-HNE to GSH, generating GS-HNE for subsequent export from the cell [171,172,173]. Importantly, our recent study [5] showed that transient expression of the reprogramming factors Oct4–Sox2–Klf4 (OSK) selectively induces GSTA4 in aged retinal pigment epithelium (RPE): GSTA4 knockout abolishes OSK-mediated oxidative protection, whereas AAV-GSTA4 alone recapitulates its cytoprotective and vision-preserving effects in vivo. These findings identify GSTA4 as a key effector of GSH-dependent detoxification of lipid peroxidation products (Figure 8).

Intravenous GSH showed minimal clinical benefit in PD (Table 2) [166], potentially because of limited blood–brain barrier penetration. Subsequent studies therefore explored intranasal delivery, which can access the CNS through olfactory and trigeminal pathways. Intranasal GSH has been reported to be safe and well tolerated, with some studies suggesting symptomatic improvement in PD [174,175].

To further optimize intracellular GSH levels, the combination supplement GlyNAC- composed of the GSH precursors N-acetylcysteine (NAC) and glycine-was developed to support and enhance the body’s natural GSH synthesis. This approach enables cells to efficiently produce GSH based on their metabolic needs, especially during periods of increased oxidative stress or GSH deficiency. Additionally, GlyNAC can efficiently cross the blood-brain barrier. In mice, GlyNAC supplementation reduced oxidative stress and improved age-related phenotypes, with one study reporting lifespan extension [176]. Clinical studies in older adults further found GlyNAC to be safe and associated with improved mitochondrial function, GSH availability, and markers of oxidative stress and aging [177,178]. In human RPE models, NAC has also been shown to protect against oxidative damage, supporting glutathione-precursor supplementation as a rational retinal antioxidant strategy [179].

4.3. Coenzyme Q10 and Its Analogues

Unlike VC, VE, and GSH/GlyNAC, which directly scavenge reactive species, Coenzyme Q10 (CoQ10) refers to a mitochondrial redox-active coenzyme that cycles between oxidized ubiquinone and reduced ubiquinol forms. In its oxidized ubiquinone form, CoQ10 functions in the electron-transport chain by shuttling electrons from complexes I/II to complex III. In its reduced ubiquinol form, CoQ10 acts as a lipophilic antioxidant, limiting membrane lipid peroxidation by quenching lipid peroxyl radicals [180,181].

Clinical trials of CoQ10 in Parkinson’s disease (PD) began in the late 1990s. Early studies found CoQ10 to be safe but largely ineffective at lower doses [182], while later trials using up to 1,200 mg/day suggested it slows functional decline [183]. However, subsequent studies failed to confirm meaningful clinical benefit in PD [184,185,186]. Similarly, although CoQ10 improved cognitive outcomes in animal models of Alzheimer’s disease, clinical efficacy has not been demonstrated [187]. Its therapeutic potential is limited in part by poor bioavailability and inefficient mitochondrial uptake [188].

Preclinical studies in glaucoma further support CoQ10 as a mitochondrial-protective antioxidant in the visual CNS. In a mouse model of glaucoma, CoQ10 reduced glutamate excitotoxicity, oxidative stress, and mitochondrial alteration, thereby improving retinal ganglion cell survival [189]. A later study showed that ubiquinol, the reduced form of CoQ10 with improved bioavailability, also attenuated oxidative stress and preserved retinal ganglion cell survival and visual function. These findings suggest that CoQ10-based interventions may be particularly relevant in optic neuropathies where mitochondrial dysfunction and oxidative injury contribute to retinal ganglion cell degeneration.

To improve mitochondrial delivery, analogues such as mitoquinone (MitoQ) were developed. MitoQ links ubiquinone to the lipophilic triphenylphosphonium (TPP) cation, enabling preferential mitochondrial accumulation [190]. Additionally, MitoQ is absorbed into cells faster than CoQ10 due to its smaller size, water solubility, and positive charge, allowing for rapid uptake from circulation without the need for fat or food [136].

Despite improved uptake, a double-blind, placebo-controlled trial found no effect of MitoQ on PD progression [191]. Further studies are needed to determine whether mitochondrial-targeted CoQ analogues can provide benefit in other neurodegenerative settings.

4.4. Other Noteworthy Oxidation-Resistant Interventions

Sterically hindered phenols have been shown in models of ox-LDL induced injury to provide potent protection of RPE from oxidative stress [192]. In RPE cultures, hindered phenol–containing antioxidants such as troglitazone, PMC (2,2,5,7,8-pentamethyl-6-chromanol), and trolox prevent ~90% of ox-LDL–induced cell death, whereas several non-phenolic antioxidants with strong general radical-scavenging capacity fail to confer similar protection, indicating a structure-specific effect. These compounds accumulate in lysosomes where they suppress ROS, stabilize lysosomal membranes, and facilitate the clearance of ox-LDL, in part by increasing levels of NCP1, a cholesterol-export protein [192]. Mechanistically, PMC blunts ox-LDL induced upregulation of oxidative stress-responsive genes such as HMOX-1 and other genes downstream of NRF2 [193]. Collectively, these studies support hindered phenol as promising candidates for pharmacologic protection in dry AMD.

Alpha-lipoic acid (ALA) is a mitochondrial antioxidant containing two thiol groups that support redox reactions and directly scavenge hydroxyl radicals, hypochlorous acid, and singlet oxygen [194]. Although ALA did not protect against expansion of the geographic atrophy form of AMD in a small clinical trial [195], and its clinical efficacy remains debated when used for diabetic peripheral neuropathy [196,197,198]. However, small clinical studies have suggested cognitive stabilization in mild AD [199] and benefit from combining ALA with omega-3 fatty acids [200].

Deprenyl (selegiline) irreversibly inhibits monoamine oxidase B (MAO-B), reducing dopamine breakdown and associated H₂O₂ generation. Clinical trials suggested that deprenyl can delay disability and the need for levodopa in Parkinson’s disease (PD) [201,202,203], although later reviews found no clear evidence that MAO-B inhibitors slow disease progression and noted more treatment withdrawals due to adverse events [204].

Deuterated docosahexaenoic acid (D-DHA) has emerged as a promising oxidation-resistant supplement because deuterium substitution markedly slows the initiation of DHA oxidation and subsequent lipid-peroxidation chain reactions. Photoreceptor outer-segment membranes are exceptionally rich in native DHA - accounting for ~60% of their polyunsaturated lipid content - making them especially susceptible to iron-catalyzed peroxidative damage that underlies age-related retinal degeneration. Joshua Dunaief’s group demonstrated that orally delivered D-DHA efficiently incorporates into retinal membranes and, in an iron-overload mouse model of geographic atrophy, virtually abolished carboxyethlypyrrole adduct formation, preserved photoreceptor and RPE architecture, and maintained electroretinographic function - benefits not achieved with unmodified DHA [205]. Similarly, D-DHA protects RPE from degeneration in a mouse model of chronic oxidative stress [206]. These findings establish D-DHA as a mechanism-based antioxidant that acts upstream to curb lipid peroxidation, holding translational promise for dry AMD and other oxidative retinopathies. A clinical trial of D-DHA for geographic atrophy has been initiated [NCT07029945].

Memantine, an uncompetitive NMDA receptor antagonist, was evaluated for glaucoma neuroprotection by targeting glutamate-mediated RGC excitotoxicity. In non-human primate glaucoma models, memantine reduced structural RGC/optic nerve damage and protected lateral geniculate nucleus neurons from shrinkage, supporting glaucoma as a broader CNS neurodegenerative process [207,208]. However, two large Phase 3 trials in open-angle glaucoma failed to meet their primary endpoints, highlighting the translational challenges of neuroprotection trials, including endpoint selection and trial-execution limitations [209].

Several metabolic interventions may also enhance oxidative resilience. Caloric restriction and lithocholic acid (LCA) can reduce mitochondrial ROS and extend lifespan in model organisms [210,211,212]. Alpha-ketoglutarate (AKG) can imitate a state of energy scarcity by regulate tricarboxylic acid cycle and limit ATP synthase activity [213]. Besides, AKG can directly consume ROS by reacting with ammonia and generates glutamine [214] or H₂O₂ decomposition [215]. AKG has been reported to extend lifespan and exhibit anti-aging effects in worms, flies, and mice [216,217,218].

Photobiomodulation (PBM) using red and near-infrared (NIR) light has emerged as a non-invasive strategy to enhance mitochondrial resilience. Red/NIR wavelengths (~600–1000 nm) are thought to act primarily through mitochondrial cytochrome c oxidase, enhancing electron transport and ATP production while activating antioxidant and pro-survival pathways [219,220];. In vitro and in vivo studies show that PBM can improve mitochondrial function and reduce ROS, lipid peroxidation, and retinal degeneration following oxidative injury [221]. Glaucoma-relevant studies further demonstrate that 670-nm red light protects retinal ganglion cell dendrites following axonal injury and increases the redox-protective protein DJ-1, suggesting activation of endogenous oxidative resilience [222]. More broadly, PBM has shown therapeutic potential across retinal diseases, including AMD [223]. Most recently, the FDA authorized the Valeda Light Delivery System, combining 590-, 660-, and 850-nm wavelengths, for treatment of dry AMD, marking an important clinical translation of PBM. These support red/NIR light as a drug-free strategy to reinforce mitochondrial function and endogenous antioxidant defenses.

Finally, controlled hypoxia has emerged as a strategy for mitochondrial disease [224]. Supporting this concept, the small-molecule compound HypoxyStat - which alters hemoglobin-oxygen dissociation dynamics to reduce tissue oxygen delivery - demonstrated remarkable efficacy in a preclinical model of mitochondrial disease. Daily oral administration of HypoxyStat induced systemic hypoxia under normoxic conditions and prolonged survival in Ndufs4 knockout mice, while ameliorating neurobehavioral deficits and pathological hallmarks [225].

4.5. Rejuvenating Endogenous Oxidative Resilience with Gene Therapy

An emerging strategy is to use AAV gene therapy to rejuvenate endogenous oxidative resilience that deteriorates with age, rather than relying solely on repeated administration of exogenous antioxidants. NRF2 is a central regulator of antioxidant and detoxification programs, and AAV-mediated restoration of NRF2 signaling could reinforce multiple endogenous defense pathways simultaneously [226]. We discovered that epigenetic restoration provides a broader approach: AAV delivery of epigenetic reprogramming factors Oct4–Sox2–Klf4 (OSK) can restore youthful cellular programs in retinal neurons [227], benefiting animal models of glaucoma [227,228], multiple sclerosis [229], NAION [230], and AD [231], and even extending mouse lifespan [232]. AAV-OSK has now entered into clinical trial to examine efficacy in restoring vision of glaucoma and NAION patients (NCT07290244). Notably, whereas the effects of OSK in neurons are largely mediated through epigenetic remodeling, including mechanisms involving the DNA demethylase TET2, a distinct non-canonical function of OSK has recently been identified in RPE: enhancing cellular resilience to diverse oxidative insults, including NaIO₃, ox-LDL, and H₂O₂ [5]. Mechanistic dissection of this response identified GSTA4 as a direct downstream effector of OSK, whose promoter is activated by OSK and whose expression enhances oxidative resilience in the RPE by detoxifying the lipid peroxidation product 4-HNE [5]. These findings suggest that individual cytoprotective effectors downstream of reprogramming factors may capture selected therapeutic benefits of rejuvenation while avoiding risks associated with prolonged expression of the reprogramming factors themselves. Together, NRF2, OSK, and GSTA4 illustrate a developing paradigm in which gene therapy is used not simply to neutralize ROS, but to restore the intrinsic stress-response networks that enable young cells to withstand oxidative injury.

5. Measuring Oxidative Stress - Biomarkers Aiding Translation to Human Therapies

Particularly in the retina, we have a unique opportunity to measure oxidative stress and related biomarkers through imaging modalities and liquid biopsy sampling [233]. Imaging is particularly motivating through hyperspectral and fluorescent wavelengths that return reliable data on mitochondrial redox stress balance, as well as blood flow and oxygenation, both at rest and in response to visual/light stimuli [234,235]. In the eye, aqueous humor liquid-biopsy approaches such as TEMPO enable proteomic and metabolomic profiling from small-volume samples, providing a more direct molecular readout of the retinal environment for inflammation, cellular stress, and biological aging [236]. Novel visual stimuli combined with advanced approaches to the analysis of visually evoked potentials in the brain can parse out the eye-brain connection in glaucoma [237]. Such biomarkers in the eye are also proving useful as biomarkers for neurodegenerative disease in the brain and even healthspan more broadly [238,239], and are discussed at length elsewhere.

More broadly, we propose a multilayered oxidative-stress profiling framework (Table 3; Figure 9) combining markers of lipid, protein, and DNA oxidation; iron homeostasis; endogenous antioxidant capacity; and inflammation with retinal imaging, MRI/MRS, genetics, and longitudinal wearable data. Integration of these measurements through AI could help define individual redox vulnerabilities, stratify disease risk, and guide selection and monitoring of antioxidant or cytoprotective interventions.

6. Summary and Future Directions — From Antioxidant Therapy to Restoring Endogenous Redox Resilience

Oxidative stress is a major driver of age-related CNS degeneration. Here, we categorize commonly used oxidative insults by mechanism and application to facilitate selection of appropriate experimental models. In vivo models—including high-fat diet, radiation, cigarette smoke, air pollution, and aging itself—capture interactions among environmental exposures, mitochondrial dysfunction, inflammation, and neurodegeneration. Complementary in vitro models provide greater mechanistic precision: NaIO₃ models RPE-centered oxidative injury; NMDA induces calcium overload and secondary oxidative damage; iron dysregulation and oxLDL model ferroptotic and vascular stress; Aβ and tau reproduce components of Alzheimer’s disease pathology; and MPTP and 6-OHDA model distinct mechanisms of dopaminergic degeneration. Despite this mechanistic diversity, a recent study identified ROS-induced disulfide bond formation at surface-exposed cysteines and consequent loss of protein mobility as a potentially shared mechanism of oxidative damage [240], raising the possibility that diverse oxidative insults ultimately converge on common cellular vulnerabilities.

Antioxidant interventions are similarly diverse, ranging from direct radical scavengers such as vitamins C and E to approaches that augment endogenous antioxidant systems, including NAC, GlyNAC, alpha-lipoic acid, and mitochondrial-targeted interventions such as CoQ10 and MitoQ. Clinical studies have provided evidence of benefit in selected settings, most notably AREDS/AREDS2 for AMD, but overall efficacy across age-related CNS diseases remains variable. A major challenge is therefore to determine who is experiencing oxidative stress, which form of oxidative damage predominates, and which antioxidant pathway is insufficient. As discussed above, multimodal oxidative-stress profiling may help identify which redox pathways are impaired in individual patients and thereby guide more rational therapeutic selection.

Beyond supplying exogenous antioxidants, an emerging therapeutic opportunity is to rejuvenate the endogenous oxidative resilience that deteriorates with age. NRF2 exemplifies such a cytoprotective hub, coordinating antioxidant and detoxification programs while supporting mitochondrial function, lysosomal homeostasis, and proteostasis [226]. Its dysregulation during aging and diseases such as AMD may compromise the ability of cells to adapt to oxidative stress. Similarly, our recent work identified GSTA4 as a downstream effector of OSK-mediated rejuvenation to enhance oxidative resilience in the aged RPE [5], demonstrating that rejuvenation programs can be mechanistically dissected to identify individual cytoprotective effectors. Such approaches may circumvent limitations of conventional antioxidants, including poor delivery to the relevant cells and subcellular compartments, by restoring or augmenting intrinsic defense programs.

Along this line, a further challenge is delivering redox-modulating therapies to the appropriate CNS cell type and subcellular compartment. This is particularly relevant in the eye, where systemic therapies such as oral or intravenous drugs often achieve limited intraocular exposure because of the blood–retinal barrier, while topical formulations primarily reach anterior tissues and penetrate poorly to the posterior retina. For diseases involving the retina, such as glaucoma, where retinal ganglion cells are the primary therapeutic target, and AMD, effective delivery may therefore require local approaches such as intravitreal administration of drugs, intraocular drug delivery platforms, or AAV vectors for gene therapy approaches. Alternatively, sustained-release ocular implants [241] could provide prolonged local exposure while reducing the need for repeated dosing. For brain delivery, intranasal administration has also emerged as a promising strategy to access the CNS while bypassing the blood–brain barrier, with studies in PD showing improved delivery and encouraging clinical outcomes [174,175]. Thus, therapeutic efficacy will depend not only on identifying the appropriate antioxidant or cytoprotective pathway, but also on achieving sufficient and sustained delivery to the vulnerable cell population.

Together, the next generation of redox medicine may move beyond broad antioxidant use toward defining disease-specific oxidative vulnerabilities, rejuvenating endogenous defenses, and delivering therapies to the right cells. Combining mechanistically appropriate models, multidimensional biomarkers, rejuvenation of cytoprotective networks, and tissue-targeted delivery could enable more precise interventions for age-related CNS degeneration while creating opportunities for discoveries in the eye to inform the brain, and vice versa.