Current State of Knowledge on Amiodarone (AMD)-Induced Reactive Oxygen Species (ROS) Production in In Vitro and In Vivo Models

Abstract

Amiodarone (AMD) is an effective antiarrhythmic drug whose long-term use is limited by multi-organ toxicities linked to oxidative stress. This review synthesizes current evidence on how AMD induces reactive oxygen species (ROS) generation in vitro and in vivo, and the mechanistic pathways involved. AMD promotes ROS production through both direct and indirect mechanisms. Directly, AMD accumulates in mitochondria and impairs the electron transport chain, leading to electron leakage and superoxide formation. It also undergoes redox cycling, forming radical intermediates that trigger lipid peroxidation and deplete cellular antioxidants. AMD and its metabolites inhibit antioxidant enzymes (SOD, CAT, GPx) expression and/or activities and reduce glutathione level, compounding oxidative injury. Indirectly, AMD activates signaling pathways that exacerbate ROS generation. This compound can induce pro-inflammatory mediators such as TNF-α and modulate nuclear receptors such as AhR, PXR, CAR, and PPARs, altering the expression of metabolic enzymes and endogenous antioxidants. These processes are time- and dose-dependent: short exposures at low concentrations may transiently scavenge radicals, whereas chronic or higher-dose exposures consistently lead to net ROS accumulation. The oxidative effects of AMD vary by tissue and experimental models. In chronic models, organs such as the lung and liver show pronounced ROS-mediated injury, whereas acute or cell-based systems typically exhibit subtler changes. AMD-induced toxicity arises from multifactorial oxidative stress involving mitochondrial dysfunction, increased radical formation, depletion of antioxidant defenses, and activation of pro-oxidant signaling pathways. Recognizing these pathways suggests that antioxidant and mitochondria-targeted co-therapies could ameliorate the side effects of AMD.

1. Introduction

Amiodarone (AMD, CAS: 19774-82-4) is a highly effective class III antiarrhythmic drug used to treat atrial and ventricular arrhythmias [1]. Its broad pharmacological profile and efficacy in maintaining sinus rhythm make it invaluable in cardiology [2]. Unfortunately, the clinical use of AMD is limited by serious multi-organ toxicities, including, most notably, pulmonary fibrosis, hepatotoxicity, thyroid dysfunction, neuropathy, and ocular complications [3]. These adverse effects are attributable to the physicochemical properties of AMD: it is a highly lipophilic, iodine-rich benzofuran derivative with a long half-life (weeks to months), owing to extensive tissue accumulation [4]. A growing body of evidence indicates that oxidative stress, defined as an imbalance between reactive oxygen species (ROS) production and antioxidant defenses, plays a pivotal role in the pathogenesis of AMD-induced toxicity [4,5]. AMD has been shown to increase the generation of ROS while simultaneously impairing cellular antioxidant systems, leading to oxidative damage in various tissues [4].

Early investigations into AMD toxicity produced conflicting views on the involvement of ROS [6,7]. In the 1990s, some studies questioned whether ROS drive AMD-induced lung injury. For example, Leeder et al. found no significant ROS generation in alveolar macrophages acutely exposed to AMD, and their hamster model suggested that depleting antioxidants did not markedly worsen lung injury, thus not supporting ROS as the primary cause of AMD pulmonary toxicity [6]. Similarly, a chronic rat study by Sarma et al. observed only minimal oxidative changes and concluded that pulmonary/hepatic toxicity was not directly mediated by oxidative stress [7]. However, these authors noted some oxidative perturbations such as increased lipid peroxidation in spleen where AMD accumulated and did not exclude a secondary role for ROS [7]. In contrast, other contemporaneous studies began documenting antioxidant disturbances caused by AMD. Trivier et al. reported that AMD and its active metabolite desethylamiodarone (DEA) overwhelm cellular antioxidant defenses in cultured human cells, hinting that oxidative mechanisms contribute to AMD toxicity [5]. Over the past two decades, a consensus has emerged that ROS generation and oxidative injury are central features of AMD toxic effects in many models.

Therefore, the aim of this review is to provide a detailed examination of how AMD induces ROS production in vitro and in vivo, the impact on antioxidant enzymes, the likely biochemical mechanisms involved, and the current state of knowledge from cell studies, animal models, and clinical observations.

2. Mechanisms of AMD-Induced ROS Generation

Multiple, interconnected mechanisms underlie ROS production by AMD in cells (Figure 1). A primary route of AMD action leads to mitochondrial dysfunction [8]. AMD is known to accumulate in mitochondria (due to its amphiphilic cationic nature) and disrupt the electron transport chain, leading to electron leak and superoxide formation [4,9]. In human lung epithelial cells, Nicolescu et al. observed that AMD rapidly depolarizes mitochondrial membranes and releases cytochrome c, and that mitochondrial damage precedes detectable ROS increase, suggesting that ROS are generated as a downstream consequence of mitochondrial injury [9]. This mitochondrial ROS pathway is supported by in vivo liver studies [10]. Serviddio et al. showed that AMD administration in rats causes a surge in mitochondrial hydrogen peroxide (H₂O₂) production, increased lipid peroxidation, and hepatic cell damage, whereas an analogue (dronedarone) that causes less mitochondrial ROS exhibited reduced toxicity [10]. Thus, AMD direct impairment of mitochondrial respiration and ATP synthesis triggers ROS generation in both lung and liver models. Notably, mitochondria-targeted antioxidants like coenzyme Q₁₀ and α-tocopherol provided significant protection against AMD toxicity in vitro, more so than generic antioxidants, reinforcing that mitochondria are a key ROS source [9].

Another major mechanism is the impairment of cellular antioxidant defenses by AMD. Both in vitro and in vivo studies demonstrate that AMD depresses the activity or expression of key antioxidant enzymes, tipping the redox balance towards oxidation. For example, in human cell lines, AMD and DEA caused significant decreases in glutathione-dependent enzymes and superoxide dismutase (SOD) activity [5]. Similarly, in rat organs, chronic AMD exposure led to marked reductions in catalase (CAT), SOD, and glutathione peroxidase (GPx) levels along with depletion of reduced glutathione (GSH) [11,12]. In rodent study, 4 weeks of AMD treatment caused a significant drop in all lung antioxidant enzymes (SOD, CAT, GPx), after initial partial changes at earlier time points [12]. The authors concluded that inhibition of the antioxidative protective enzymatic system by AMD permits excess ROS formation, establishing oxidative stress as a driving factor in AMD-induced pulmonary injury [12]. This loss of antioxidant defense likely occurs via multiple pathways such as direct enzyme inactivation by ROS or reactive AMD metabolites, depletion of enzyme cofactors, e.g., GSH, and downregulation of antioxidant gene expression [12]. Indeed, AMD treatment has been associated with oxidant-mediated consumption of GSH (e.g., ~50–100% GSH depletion in rat lungs over 1–4 weeks) and increased oxidative modification of proteins [12]. This leads to a self-reinforcing positive feedback loop. AMD-induced ROS can damage antioxidant enzymes and reduce GSH, which in turn further weakens ROS scavenging and allows for reactive species to accumulate unchecked [5,11].

The chemical structure and metabolism of AMD also contribute to ROS generation. AMD is a benzofuran derivative containing iodine [13]. Its metabolism—primarily by CYP3A4 and CYP2C8, yields the active metabolite DEA as well as potentially reactive intermediates [14]. There is evidence that AMD can undergo redox cycling or radical formation [15]. Nicolescu et al. detected aryl radical species in AMD-exposed lung cells using spin-trapping, implicating direct free radical formation in AMD toxicity [9]. Such radicals can initiate lipid peroxidation chain reactions in membranes, compounding oxidative injury [16]. Lipid-rich organelles such as lysosomes and mitochondrial membranes are particularly vulnerable—in fact, AMD causes phospholipid accumulation in these organelles, which may sensitize them to peroxidative damage [16]. Oxidative breakdown of accumulated lipids yields malondialdehyde (MDA) and other harmful aldehydes, further damaging cellular components [17]. Lysosomal dysfunction from AMD with lamellar lipid inclusions has been observed in multiple tissues such as lungs, liver, and optic nerve, and is thought to impair autophagic clearance, indirectly promoting oxidative stress by allowing for damaged mitochondria and peroxidized lipids to accumulate [3]. This ties into the concept that enhancing autophagy can alleviate AMD toxicity [18]. To date, it has been described that activation of autophagy rescues AMD-induced apoptosis of lung epithelial cells and pulmonary toxicity in rats, probably by removing dysfunctional, ROS-generating mitochondria [19].

AMD is a prototypical cationic amphiphilic drug that induces drug-induced phospholipidosis (DIPL) across multiple tissues, characterized by lysosomal lamellar inclusions and “foamy” lipid-laden cells [3,16]. Mechanistically, AMD accumulates within acidic compartments, binds anionic phospholipids, and functionally impairs lysosomal phospholipases, which in turn hinders autophagic flux and mitophagy. The resulting backlog of damaged, ROS-generating mitochondria provides a sustained source of oxidative stress; accordingly, genetic or pharmacologic enhancement of autophagy alleviates AMD toxicity in liver and lung models [18,19]. Beyond autophagy blockade, DIPL expands the pool of peroxidation-prone lipids within lysosomes and mitochondria, thereby amplifying chain lipid peroxidation and secondary aldehyde formation (e.g., MDA), which is observed in vivo alongside depletion of GSH and activities of SOD, CAT, and GPx [10,11,12].

AMD also promotes hepatic steatosis, and steatohepatitis-like injury has been recapitulated experimentally with mitochondrial dysfunction and lipid peroxidation as early events [10,17]. In mouse liver, AMD—but not its less tissue-accumulating analogue dronedarone—triggers robust mitochondrial ROS production and respiratory chain impairment, with antioxidant rescue by N-acety-L-cysteine (NAC), directly tying steatotic remodeling to redox imbalance [10]. Nuclear-receptor crosstalk appears to modulate these lipid-redox phenotypes. Peroxisome proliferator-activated receptor alpha (PPARα) activity mitigates AMD hepatotoxicity (PPARα deficiency worsens injury, while fenofibrate is protective), consistent with restoration of β-oxidation and antioxidant tone [20]. In contrast, in airway epithelia, AMD upregulates PPARγ and lipogenic genes (e.g., SCD, FADS2), which aligns with phospholipid accumulation; notably, PPARγ can simultaneously induce antioxidant enzymes (CAT, GPx3), suggesting a mixed adaptive–maladaptive response wherein lipid remodeling and ROS defenses are co-activated but ultimately overwhelmed under chronic exposure [21,22,23]. Together, these data support a model in which AMD-induced DIPL and steatosis are not mere histologic bystanders but active amplifiers of oxidative stress: lysosomal–mitochondrial traffic jams impair organelle quality control, expand peroxidizable substrates, and entrain nuclear-receptor programs that initially compensate yet progressively fail, culminating in tissue-specific ROS injury.

Inflammatory signaling is another mechanism by which AMD induces ROS. AMD has been noted to increase pro-inflammatory cytokines like tumor necrosis factor-alpha (TNF-α) in vivo [11]. Inflammation can stimulate ROS production via activated leukocytes and cytokine-induced intracellular pathways [24]. Lu et al. found that co-treatment of murine liver cells with subtoxic TNF-α and AMD led to increased ROS generation and lipid peroxidation, far greater than with AMD alone [25]. The oxidative stress was accompanied by heightened caspases activation and cell death [25]. Notably, TNF-α/AMD toxicity was substantially mitigated by α-tocopherol (a lipid-soluble antioxidant), which reduced ROS and MDA levels, whereas water-soluble antioxidants such as NAC, GSH or ascorbate, were less effective [25]. This suggests that AMD-related ROS damage occurs largely in hydrophobic compartments such as membranes or lipids which is consistent with AMD localizing to membranes, and highlights how inflammation amplifies oxidative injury. Clinically, this mechanism may explain why systemic inflammation or high oxygen therapy can precipitate acute lung toxicity in patients on AMD. Indeed, AMD has been reported to elevate lung tissue TNF-α in rats, and antioxidants like L-carnitine can significantly reduce these cytokine levels along with ROS markers [11].

Finally, neurohormonal and secondary pathways may contribute to ROS generation by AMD. There is some evidence that AMD can activate the renin–angiotensin system (RAS) in the lung, possibly by causing lung injury [26]. Angiotensin II signaling in turn promotes ROS formation via NADPH oxidases in pulmonary cells. Consistent with this, concomitant use of angiotensin converting enzyme (ACE) inhibitors or angiotensin receptor blockers has been associated with lower incidence of AMD pulmonary toxicity in patient series, implying that blocking RAS and its downstream oxidative stress is protective [27]. Moreover, oxidative nitrosative stress may occur. AMD-treated rats exhibit increased lung nitric oxide (NO) metabolites, and since NO readily reacts with superoxide to form peroxynitrite, this could exacerbate tissue injury [12]. The interplay of ROS and reactive nitrogen species (RNS) further complicates the oxidative landscape in AMD toxicity.

In summary, AMD induces ROS production through a multifactorial mechanism involving mitochondrial electron transport impairment, direct free radical formation, antioxidant system suppression, and inflammatory oxidative cascades. These processes act in concert to produce oxidative stress in cells exposed to AMD. The next sections review the experimental evidence for ROS and antioxidant changes in vitro and in vivo, providing concrete data on how these mechanisms manifest in different models.

3. In Vitro Studies on AMD, ROS, and Antioxidant Enzymes

A variety of cell culture models such as pulmonary cells, hepatic cells, cardiac myocytes, and neurons cells have been used to analyze the pro-oxidant effects of AMD under controlled conditions [28,29]. Due to the numerous manuscripts showing the impact of AMD on ROS, Table 1 contains key and selected publications that summarize key findings from in vitro studies. Early evidence of AMD-induced oxidative stress was reported in cultured mammalian cell lines in the 1990s. Trivier et al. treated human hepatoma (Hep3B) cells and lung epithelial (L132) cells with AMD or DEA and measured antioxidant defenses [5]. The authors observed significant alterations in enzyme activities in Hep3B cells. Notably, glutathione S-transferase (GST) activity was strongly inhibited by AMD, whereas glutathione reductase activity was suppressed by DEA. In L132 cells, SOD activity dropped markedly after exposure to either compound [5]. At the same time, the oxidized-to-reduced glutathione ratio (GSSG:GSH) increased in Hep3B cells, indicating a shift toward a more oxidized redox state; this change was not observed in L132 cells [5]. The study concluded that AMD can overwhelm cellular antioxidant defenses, making oxidative injury a likely early event in cell damage [5]. These findings highlight cell type-specific responses: in lung-derived L132 cells, the predominant effect was loss of SOD activity, suggesting superoxide accumulation, whereas liver-derived Hep3B cells exhibited impairment of the glutathione system. Interestingly, Szychowski et al. reported that in the human hepatoma (HepG2) cells, exposure to a non-toxic concentration of 10 µM AMD led to an increase in SOD activity, while CAT activity and GSH levels remained unchanged [30]. Notably, the CAT protein level decreased, whereas the SOD1 protein level was unaffected [30]. These findings suggest that AMD toxicity may involve multiple oxidative pathways, potentially depending on cell-specific metabolic profiles and baseline antioxidant defenses. Furthermore, discrepancies between different studies may result from variations in the cell culture models used, exposure times, or AMD concentrations.

Interestingly, in acute settings, AMD has demonstrated some antioxidant properties in heart cells [31]. Ide et al. showed that AMD at 0.1–100 µM scavenged exogenous hydroxyl radicals (·OH) in a cell-free system and protected adult canine cardiomyocytes from H₂O₂-induced oxidative injury in vitro [31]. AMD reduced electron-spin resonance signals of ·OH radicals in a dose-dependent manner and prevented myocyte hypercontracture which is a hallmark of oxidative damage during exposure to H₂O₂ and Fe³⁺ [31]. The authors concluded that the AMD molecule can directly scavenge highly reactive oxygen radicals, a property that might contribute to its beneficial effects in ischemic heart disease [31]. This finding underscores a paradox that AMD can exhibit antioxidant effects under certain conditions such as low or short-term exposure in some cell types, even though it tends to provoke net oxidative stress with prolonged exposure or in other tissues. One possible explanation is that the AMD iodine-containing benzofuran ring can donate electrons to neutralize radicals making it a radical sponge in chemical terms, but when AMD accumulates in organelles or is metabolized, the balance shifts toward ROS generation. Thus, acute cardiomyocyte experiments revealed an antioxidant facet of AMD, whereas chronic exposure models generally demonstrate pro-oxidant outcomes.

The link between mitochondrial dysfunction and ROS in vitro was solidified by the work of Nicolescu et al. [9]. Using human lung epithelial (HPL1A) cells which is a model relevant to alveolar type I cells in the lung, they found that AMD (100 µM) significantly increased intracellular ROS levels, but notably only after a prolonged incubation (6–24 h) [9]. A short 2 h exposure to AMD did not raise ROS in those cells, despite causing severe mitochondrial depolarization and initiating apoptosis measured by cytochrome c release or caspase activation [9]. This timing experiment suggests that ROS production is a downstream event in AMD cytotoxicity. The drug first impairs mitochondria (within hours), and the ensuing respiratory dysfunction leads to an accumulation of ROS a few hours later. Importantly, pretreatment of the cells with mitochondrial-specific antioxidants like ubiquinone (coenzyme Q₁₀) or membrane-targeted vitamin E provided significant protection against AMD cytotoxicity, more so than general free radical scavengers (Trolox or the spin-trap DMPO) [9]. This indicates that mitochondrial ROS are a critical mediator of cell injury in this model, and that shielding mitochondria from oxidative damage can rescue cell viability [9]. Similar observations were made by Szychowski et al., where, in the concentration range from 20 to 50 µM, the AMD-induced increase in cytotoxicity in HepG2 cells was preceded by an increase in ROS production [30]. These studies align well with later findings that enhancement of mitochondrial quality control such as via autophagy or exogenous antioxidants can mitigate AMD toxic effects in vitro [18].

Support for AMD-induced ROS in hepatic cells comes from studies on hepatocyte-derived lines. Lu et al. examined the effect of AMD on a mouse hepatoma Hepa1c1c7 cell especially under inflammatory conditions [25]. AMD alone caused a moderate increase in ROS generation and lipid peroxidation in these liver cells, accompanied by cell death. Moreover, flow cytometry with ROS-sensitive dyes showed higher fluorescence in AMD-treated cells compared to controls [25]. When low-dose TNF-α was added, ROS levels rose even further, and cell death was markedly potentiated, demonstrating an interaction between AMD and inflammatory stress [25]. Antioxidant interventions had differential effects. Specifically, water-soluble antioxidants (NAC, GSH, ascorbate) only minimally reduced cytotoxicity, whereas the fat-soluble antioxidant α-tocopherol nearly abolished AMD-induced cell death and significantly decreased ROS and MDA levels [25]. These data suggest that oxidative damage in hepatocytes occurs predominantly within lipid compartments (e.g., membranes, lipid droplets), consistent with AMD lipophilicity. Moreover, the membrane localization and protective effect of vitamin E further underscore the central role of lipid peroxidation in AMD toxicity in liver cells [25]. Indeed, by quenching peroxyl radicals in membranes, vitamin E likely interrupted the chain reaction of lipid peroxidation triggered by AMD. These in vitro liver results parallel those in lung cells in which ROS mediate much of the toxicity, and targeting ROS especially in membranes/mitochondria can significantly improve cell survival.

In the context of ocular cells, which are relevant to AMD-associated optic neuropathy, recent in vitro work has been illuminating. Liao et al. studied human retinal pigment epithelial (RPE) cells, which are analogous to cells potentially affected in AMD-induced optic nerve edema [32]. They found that AMD caused a dose-dependent decrease in RPE cell viability and increase in apoptosis, with an LC₅₀ around 50 µM [32]. Crucially, Liao et al. demonstrated that insulin-like growth factor-1 (IGF-1) can rescue RPE cells from AMD toxicity by activating the PI3K/Akt pathway—a pro-survival signaling route that upregulates antioxidant defenses [32]. IGF-1 pretreatment dramatically improved RPE survival, reducing ROS levels and MDA accumulation, and maintaining mitochondrial membrane potential in AMD-treated cultures [32]. Blocking the Akt pathway negated the IGF-1 protective effect, indicating that IGF-1 worked by enhancing cellular resistance to oxidative stress probably via induction of antioxidant enzymes or anti-apoptotic factors like Bcl-2 [33]. Similarly, IGF-1 was shown to protect retinal ganglion cells (RGCs) from AMD-induced apoptosis, requiring Akt signaling [33]. RGCs were even more sensitive to AMD measured by significant apoptosis at low micromolar levels, presumably because they have high metabolic rates and fewer intrinsic antioxidants than RPE [3]. The IGF-1 activation of endogenous antioxidant defenses (and inhibition of pro-apoptotic FoxO3a) significantly improved RGC survival [33]. These ocular cell studies underscore the role of oxidative stress in AMD neurotoxicity and highlight the potential of boosting cell survival pathways to counteract ROS-mediated damage.

Taken together, the in vitro evidence indicates that AMD increases ROS generation and perturbs antioxidant enzymes across multiple cell types. Key issues are as follows:

(1) Mitochondrial involvement. Often AMD first compromises mitochondria, then ROS rise subsequently.

(2) Antioxidant enzyme inhibition. AMD can diminish SOD, catalase, and GPx activities and deplete GSH in cells, contributing to oxidative stress.

(3) Lipid peroxidation. Many cell studies find elevated MDA or membrane damage, suggesting that ROS attack on lipids is central.

(4) Modulation by external factors. Inflammatory cytokines exacerbate ROS damage from AMD, whereas growth factors like IGF-1 or lipophilic antioxidants can mitigate it.

Table 1. Effects of AMD on ROS and antioxidant enzymes in various cell lines. Abbreviations: ↑—increase; ↓—decrease; CAT—catalase; HO-1—heme oxygenase-1; SOD—superoxide dismutase; ROS—reactive oxygen species; MDA—malondialdehyde; GSSG—oxidized form of glutathione. In this table, ROS denotes reactive oxygen species as operationally measured in the cited studies, including hydrogen peroxide (H₂O₂), superoxide (O₂•⁻), and hydroxyl radical (•OH). Because most assays used non-selective redox-sensitive probes (e.g., H₂DCFDA/DCF) that respond to multiple oxidants and report overall oxidative burden, the entries do not discriminate individual ROS species unless explicitly stated.

Cell Line (Species)	Measured Parameter	Reference
ARPE-19 (human retinal pigment epithelial cell line)	↑ ROS and ↑ MDA from 10 to 100 µM AMD	[33]
BEAS-2B (human bronchial epithelial cell line)	↑ ROS level at 3 µM AMD	[29]
D407 (human retinal pigment epithelial cell line)	↑ ROS level, ↑ HO-1 level at 5 µM AMD	[34]
EA.hy926 (human umbilical vein endothelial hybrid cell line)	↑ MDA level from 45 to 90 µM AMD	[35]
H9c2 (rat embryonic ventricular cardiomyoblast cell line)	↑ ROS level at 0.5–4 μM AMD	[36]
HeLa (human cervical adenocarcinoma)	↑ ROS level at 30 µM AMD	[37]
Hep3B (human hepatocellular carcinoma cell line)	↑ GSSG level at 7.5 µM AMD	[5]
Hepa1c1c7 (mouse hepatoma cell line)	↑ ROS and MDA levels at 35 µM AMD	[25]
HepG2 (human hepatocellular carcinoma cell line)	↑ MDA level from 52.5 to 105 µM AMD	[35]
HepG2 (human hepatocellular carcinoma cell line)	↑ ROS level from 20 to 50 µM AMD, ↑ SOD activity and ↓ CAT level at 10 µM AMD	[30]
HPL1A (human peripheral lung epithelial cell line)	↑ ROS level at 100 µM AMD	[9]
L132 (human embryonic lung epithelial cell line)	↓ SOD activity from 1 to 10 µM AMD	[5]
MRC-5 (human fetal lung fibroblast cell line)	↑ lipid peroxidation, ↑ protein carbonyls, ↓ SOD activity and ↓ CAT activity at 100 µ AMD	[38]
primary adult canine cardiac myocytes	0.1–100 µM AMD showed direct radical-scavenging ability	[31]
Vero (African green monkey kidney epithelial cell line)	↑ MDA level from 32.5 to 65 µM AMD	[35]

Table 1. Effects of AMD on ROS and antioxidant enzymes in various cell lines. Abbreviations: ↑—increase; ↓—decrease; CAT—catalase; HO-1—heme oxygenase-1; SOD—superoxide dismutase; ROS—reactive oxygen species; MDA—malondialdehyde; GSSG—oxidized form of glutathione. In this table, ROS denotes reactive oxygen species as operationally measured in the cited studies, including hydrogen peroxide (H₂O₂), superoxide (O₂•⁻), and hydroxyl radical (•OH). Because most assays used non-selective redox-sensitive probes (e.g., H₂DCFDA/DCF) that respond to multiple oxidants and report overall oxidative burden, the entries do not discriminate individual ROS species unless explicitly stated.

Cell Line (Species)	Measured Parameter	Reference
ARPE-19 (human retinal pigment epithelial cell line)	↑ ROS and ↑ MDA from 10 to 100 µM AMD	[33]
BEAS-2B (human bronchial epithelial cell line)	↑ ROS level at 3 µM AMD	[29]
D407 (human retinal pigment epithelial cell line)	↑ ROS level, ↑ HO-1 level at 5 µM AMD	[34]
EA.hy926 (human umbilical vein endothelial hybrid cell line)	↑ MDA level from 45 to 90 µM AMD	[35]
H9c2 (rat embryonic ventricular cardiomyoblast cell line)	↑ ROS level at 0.5–4 μM AMD	[36]
HeLa (human cervical adenocarcinoma)	↑ ROS level at 30 µM AMD	[37]
Hep3B (human hepatocellular carcinoma cell line)	↑ GSSG level at 7.5 µM AMD	[5]
Hepa1c1c7 (mouse hepatoma cell line)	↑ ROS and MDA levels at 35 µM AMD	[25]
HepG2 (human hepatocellular carcinoma cell line)	↑ MDA level from 52.5 to 105 µM AMD	[35]
HepG2 (human hepatocellular carcinoma cell line)	↑ ROS level from 20 to 50 µM AMD, ↑ SOD activity and ↓ CAT level at 10 µM AMD	[30]
HPL1A (human peripheral lung epithelial cell line)	↑ ROS level at 100 µM AMD	[9]
L132 (human embryonic lung epithelial cell line)	↓ SOD activity from 1 to 10 µM AMD	[5]
MRC-5 (human fetal lung fibroblast cell line)	↑ lipid peroxidation, ↑ protein carbonyls, ↓ SOD activity and ↓ CAT activity at 100 µ AMD	[38]
primary adult canine cardiac myocytes	0.1–100 µM AMD showed direct radical-scavenging ability	[31]
Vero (African green monkey kidney epithelial cell line)	↑ MDA level from 32.5 to 65 µM AMD	[35]

4. In Vivo Animal and Clinical Evidence of Oxidative Stress in AMD Toxicity

Findings from animal models and clinical observations largely corroborate the in vitro data, demonstrating that AMD triggers oxidative stress in living organisms. However, in vivo studies also highlight the complexity of whole-body responses, such as inflammation and tissue-specific effects, which can modulate the oxidative outcomes. Table 2 and Table 3 provide an overview of major in vivo animals and patients studies examining ROS production and antioxidant changes due to AMD treatment.

Pulmonary toxicity is the most feared adverse effect of chronic AMD therapy, and many animal studies have focused on the lungs. Early work, using rodent models of AMD-induced pulmonary fibrosis, gave mixed messages about ROS involvement [8,39]. For instance, Leeder et al. in the mid-1990s studied hamsters given intratracheal AMD and found that depleting lung GSH only modestly exacerbated lung injury, and they could not detect increased ROS production by H₂DCF fluorescence in alveolar macrophages acutely exposed to AMD [6]. The authors concluded that their results do not support a primary role for ROS in AMD lung toxicity [6]. Similarly, a chronic rat study by Sarma et al. did not find significant elevation in lung MDA which is a lipid peroxidation marker after 3 weeks of high-dose AMD (60 mg/kg/day), except in spleen tissue where AMD levels were highest [7]. Lung and liver MDA remained statistically unchanged despite clear signs of pneumonitis/fibrosis, leading the authors to suggest that oxidative stress was not the direct cause of lung injury [7]. However, it is important to note that these studies had limitations such as authors often measured end-point markers such as MDA, etc. at a single time, possibly missing transient ROS spikes, and their analytic sensitivity which especially in older studies might not detect all forms of ROS.

More recent and comprehensive animal studies provide stronger evidence that oxidative stress is indeed integral to AMD-induced lung damage, especially when assessed over time. A detailed time-course study in rats by Al-Shammari et al. followed daily AMD administration (80 mg/kg intraperitoneal) for up to 4–5 weeks, measuring multiple oxidative indices at weekly intervals [12]. In the first two weeks, the authors observed a somewhat paradoxical drop in lung MDA levels compared to controls [12]—which they attributed to an early antioxidant or radical-scavenging effect of AMD or reduced metabolic activity in the lung. However, as exposure continued, clear signs of oxidative injury emerged. Lung GSH content fell significantly (by 50% at week 1 and >50% at weeks 3–4), and by week 4 the activities of all major antioxidant enzymes (CAT, SOD, GPx) were markedly decreased ranging from ~30% to ~60% lower than control [12]. Notably, at week 3, the authors saw an aberrant increase in glutathione reductase activity and changes in SOD, which they interpreted as an influx of inflammatory cells with their own antioxidant enzymes and altered ROS handling during pneumonitis [12]. By week 4, an oxidative stress climax was evident. NO metabolites were elevated indicating nitrosative stress, and the imbalance between ROS and antioxidants was severe enough that some rats developed overt respiratory failure significant mortality by week 5 [12]. Histologically, lungs progressed from edema to interstitial inflammation to fibrosis, paralleling the oxidative stress timeline [12]. The authors concluded that AMD induces lung injury in a sequence such as edema, inflammation, oxidative stress, and fibrosis, with ROS playing a causative role in the later stages and possibly being involved in the earlier stages via signaling [12]. This comprehensive in vivo study underscores that while very early AMD exposure might not show ROS increase or might even show an antioxidant effect, prolonged exposure depletes antioxidant reserves and results in significant oxidative damage in the lungs.

The role of ROS in AMD pulmonary toxicity is further supported by studies testing antioxidant therapies in vivo. Vitamin E, a potent lipid-soluble antioxidant, was evaluated in a rat model of AMD lung toxicity by Gawad et al. [40]. In that study, rats receiving AMD for 6 weeks developed classic lung lesions with inflammation, and thickened alveolar septa along with elevated oxidative markers, whereas the group co-treated with vitamin E had reduced lung damage and improved antioxidant status [40]. Though detailed data from that abstract are limited, the authors reported that vitamin E co-administration led to “almost normal” histology and significantly lower markers of oxidative stress, highlighting that antioxidant supplementation can mitigate AMD lung injury [40]. Similarly, L-carnitine, an endogenous compound known to support mitochondrial metabolism and act as an ROS scavenger, has been studied in this context. Dawood et al. treated rats for 6 weeks with AMD causing lung fibrosis with or without L-carnitine [11]. AMD-only rats showed severe lung damage with disrupted mitochondria, fibrosis, and high inflammation, as well as significant oxidative stress evidenced by increased lung MDA levels and decreased CAT and SOD antioxidant enzyme activities, and GSH levels were all significantly depleted by AMD [11]. In contrast, rats receiving L-carnitine alongside AMD had much lower MDA levels and their CAT, SOD, and GSH levels were restored towards normal [11]. L-carnitine also reduced lung TNF-α levels and ameliorated the histopathological changes [11]. These results strongly indicate that oxidative injury is a major driver of AMD-induced lung damage. Scavenging ROS and bolstering antioxidants via L-carnitine or vitamin E largely protected the lung tissue [11,40]. Other natural antioxidants have shown similar protective effects, such as grape seed extract and Ginkgo biloba which is rich in flavonoids, decreased lipid peroxidation, and prevented fibrotic changes in AMD-treated rats [41]. Collectively, these antioxidant intervention studies provide in vivo functional evidence that oxidative stress is a key part of AMD toxicity.

Beyond the lungs, the hepatic toxicity of AMD also involves oxidative mechanisms. AMD can cause a chronic hepatitis-like picture in patients with steatosis and fibrosis. In a mouse model, Serviddio et al. compared AMD to dronedarone which is a structural analog developed to be less toxic in causing liver injury [10]. AMD, but not dronedarone, led to pronounced mitochondrial ROS generation and respiratory chain dysfunction in liver [10]. AMD-treated mice had increased markers of lipid peroxidation in the liver and activation of the unfolded protein response likely secondary to ROS and lipid accumulation, whereas dronedarone caused some mitochondrial uncoupling but did not induce ROS overproduction [10]. Moreover, co-treatment with NAC which is both an ROS scavenger and GSH precursor, prevented much of the oxidative damage from AMD, again highlighting the central role of ROS in the hepatic toxicity of AMD [10]. These findings are consistent with clinical observations that dronedarone, while not risk-free, tends to cause less severe pulmonary and hepatic side effects than AMD, presumably because it does not accumulate in tissues as extensively or generate as many reactive metabolites. Similarly, Takai et al. (2016) administered high-dose oral amiodarone (1000 mg/kg/day for 3 days) to mice and observed a significant fall in the hepatic GSH/GSSG ratio indicating oxidative stress along with elevated ALT [42]. In the nematode Caenorhabditis elegans, Balkrishna et al. (2024) found that amiodarone exposure induced ROS in whole worms, an effect reversed by an antioxidant treatment [43].

Some interesting nuances have emerged from these in vivo studies. In Sarma’s 1997 rat experiment, AMD-treated rats’ alveolar macrophages actually did not show an increased superoxide production upon zymosan stimulation [7]. In fact, phagocytic activity was reduced by ~20%, which could imply that AMD impairs immune cell function. One could speculate that the AMD presence in macrophages where it accumulates as foamy phospholipids might inhibit NADPH oxidase activity, explaining why no extra ROS was seen in an ex vivo burst test [7]. However, this does not contradict the overall oxidative stress involvement. It may simply indicate that at baseline AMD-laden macrophages produce less ROS, which perhaps contributes to immunosuppression in the lung. On the other hand, in damaged tissue, AMD generates ROS through non-phagocytic routes directly in mitochondria [7]. Moreover, Sarma et al. found that the spleen had the highest AMD levels and the only significant MDA increase among organs [7]. The spleen is rich in mononuclear phagocytes that can store the drug [44]. Apparently, by 3 weeks, those cells experienced oxidative lipid damage (MDA increased by ~30%) [7]. This suggests that organs such as spleen, lung, liver, and adipose with high AMD accumulation may suffer more oxidative injury.

Clinically, direct evidence of ROS in patients on AMD is limited. However, clinical scenarios and indirect markers strongly implicate oxidative stress. AMD-induced pulmonary toxicity (AIPT) in patients often presents as chronic interstitial pneumonitis or acute respiratory distress syndrome (ARDS) [45]. It is well documented that risk factors for AIPT include high FiO₂ oxygen therapy and surgery with general anesthesia which often involves high oxygen levels [45]. Oxygen itself can generate ROS especially in the hyperoxia context [46]. The fact that patients on AMD who undergo surgery or oxygen therapy have precipitous respiratory failure suggests a synergistic toxicity. The oxidative stress from high oxygen likely adds to the inherent toxicity of AMD, tipping the balance and causing acute lung injury [47,48]. Case reports detail patients developing fulminant ARDS soon after being started on AMD and then receiving oxygen for surgery [45]. The pathological findings often include diffuse alveolar damage with hyaline membranes (like classic ARDS), and sometimes eosinophils and foamy macrophages—consistent with both an acute oxidative injury and the chronic phospholipid accumulation from AMD. Steroids are usually effective if started early, presumably because they reduce inflammation and oxidative burst from immune cells [49].

Clinical biomarkers of oxidative lung injury in AIPT have been explored. Some authors measured serum surfactant protein levels and KL-6 which is a marker of lung fibrosis in AMD patients, though these are not specific to oxidative damage [50]. There is interest in whether serum MDA or antioxidant levels change in patients on long-term AMD, but systematic studies are scarce. A study from the 1990s in a Hungarian journal noted that patients on AMD had signs of lipid peroxidation and depressed vitamin E levels [51]. Also, it is known that 50% of patients on chronic AMD discontinue it due to side effects [52], suggesting that the cumulative oxidative burden in tissues becomes intolerable for many. Patients often exhibit corneal deposits (AMD keratopathy) which, while not harmful, indicate high drug exposure. Interestingly, these corneal microdeposits might result from oxidative aggregation of lipids/proteins in keratocytes loaded with AMD [53].

AMD neurotoxicity in the optic nerve (AMD-induced optic neuropathy, AAON) is thought to be related to ROS and impaired axonal transport. As reviewed by Mitchell and Chacko [3], AMD causes swollen optic disc and vision loss resembling non-arteritic ischemic optic neuropathy. The mechanism is not fully understood, but pathological studies have shown axonal degeneration with lipid-laden inclusions in optic nerve fibers [54]. It is hypothesized that these inclusions disrupt axoplasmic flow, and that local oxidative stress/inflammation contributes to optic nerve head ischemia. The fact that IGF-1/Akt was protective in retinal cell models has led to suggestions of testing neuroprotective or antioxidant therapies in patients who develop AAON [3,33]. Clinically, AAON is rare (incidence ~0.1–2% in AMD users), but it underscores that central nervous system (CNS) tissues are not immune to the oxidative effects of AMD.

In summary, animal models overwhelmingly indicate that AMD induces oxidative stress in vivo, manifesting as increased ROS markers and decreased antioxidant defenses in target organs. Antioxidant treatments in animals prevent or attenuate toxicity, reinforcing causation. Clinically, while direct ROS measurement is challenging, the circumstances and pathology of AMD toxicity strongly support that ROS-driven damage is a central component. The in vivo evidence bridges the gap between cellular mechanisms and patient outcomes, affirming that oxidative injury is a unifying theme in AMD organ toxicities.

Table 2. In vivo studies on AMD-induced oxidative stress in mammals. Abbreviations: ↑—increase; ↓—decrease; CAT—catalase; GSH—glutathione; GPx—glutathione peroxidase; i.p. injection—intraperitoneal injection; LPO—lipid peroxidation; MDA—malondialdehyde; ROS—reactive oxygen species; SOD—superoxide dismutase. In this table, ROS denotes reactive oxygen species as operationally measured in the cited studies, including hydrogen peroxide (H₂O₂), superoxide (O₂•⁻), and hydroxyl radical (•OH). Because most assays used non-selective redox-sensitive probes (e.g., H₂DCFDA/DCF) that respond to multiple oxidants and report overall oxidative burden, the entries do not discriminate individual ROS species unless explicitly stated.

Species (Strain)	Organ	AMD Dose	Duration	ROS/MDA/LPO	SOD	CAT	GPx	GSH	Reference
Rat (albino)	Kidneys and serum	30 mg/kg/day oral gavage	8 weeks (chronic)	↑ MDA level	↓ SOD activity	-	-	-	[55]
Rat (Fischer-344)	Lung, Liver, Spleen, Kidney, Heart	60 mg/kg/day i.p. injection	21 days (sub-chronic)	↑ MDA level	-	-	-	-	[7]
Rat (Sprague–Dawley)	Lung	80 mg/kg/day i.p. injection	1–4 weeks (chronic)	↑ MDA level	↓ SOD activity	↓ CAT activity	↓ GPx activity	↓ GSH level	[12]
Rat (Sprague–Dawley)	Heart	100 mg/kg/day oral gavage	7 days (acute)	↑ LPO level	↓ SOD activity	↓ CAT activity	-	↓ GSH level	[56]
Rat (Sprague–Dawley)	Lung	40 mg/kg/day oral gavage	4 weeks (chronic)	↑ MDA	↓ SOD activity	-	-	↓ GSH level	[57]
Rat (Sprague–Dawley)	Liver	40 mg/kg/day oral gavage	8 weeks (chronic)	↑ ROS and ↑ MDA levels	↓ SOD activity	↓ CAT activity	↓ GPx activity	-	[58]
Rat (Sprague–Dawley)	Lung	80 mg/kg/day i.p. injection	7 and 14 days (acute) 21 and 28 days (sub-chronic)	↑ MDA level	↑ SOD (7 and 14 days), ↓ SOD (21 and 28 days) activity	↓ CAT activity	↓ GPx activity	↓ GSH level	[12]
Rat (Swiss albino)	Lung	100 mg/kg/day i.p. injection	10 days (acute)	-	↓ SOD activity	↓ CAT activity	-	↓ GSH level	[59]
Rat (Wistar)	Liver, Kidney, Testis	0.17 mg/kg/day oral gavage	30 days (sub-chronic)	↑ LPO level	↑ SOD activity	↑ CAT activity	-	-	[60]
Rat (Wistar)	Lung	30 mg/kg/day oral gavage	60 days (chronic)	↑ MDA level	↓ SOD activity	↓ CAT activity	↓ GPx activity	↓ GSH level	[11]
Rat (Wistar)	Lung	30 mg/kg/day oral gavage	6 weeks (chronic)	↑ MDA level	↓ SOD activity	-	-	↓ GSH level	[40]
Rat (Wistar)	Liver	20 mg on day 1 and a maintenance dose of 8 mg from day 2 to day 6 oral gavage	6 days (acute)	↑ LPO level	-	-	-	↓ GSH level	[10]
Rat (Wistar)	Lung	40 mg/kg/day oral gavage	4 weeks (chronic)	↑ MDA level	-	-	-	↓ GSH level	[61]
Rat (Wistar)	Lung	30 mg/kg/day oral gavage	4 weeks (chronic)	↑ ROS and ↑ MDA levels	-	-	-	-	[62]
Rat (Wistar)	Optic nerve tissue	50 mg/kg/day and 100 mg/kg/day oral gavage	14 days (acute)	↑ MDA level	↓ SOD activity	↓ CAT activity	-	↓ GSH level	[63]
Mouse (C57BL/6)	Liver	1000 mg/kg p.o. daily	3 days (acute)	↑ LPO level	-	-	-	↓ GSH level	[42]
C. elegans (N2 and CF1553 strains)	Whole worm	10 µM in media	24 h	↑ ROS level	↑ SOD3 expression	-	-	-	[43]

Table 2. In vivo studies on AMD-induced oxidative stress in mammals. Abbreviations: ↑—increase; ↓—decrease; CAT—catalase; GSH—glutathione; GPx—glutathione peroxidase; i.p. injection—intraperitoneal injection; LPO—lipid peroxidation; MDA—malondialdehyde; ROS—reactive oxygen species; SOD—superoxide dismutase. In this table, ROS denotes reactive oxygen species as operationally measured in the cited studies, including hydrogen peroxide (H₂O₂), superoxide (O₂•⁻), and hydroxyl radical (•OH). Because most assays used non-selective redox-sensitive probes (e.g., H₂DCFDA/DCF) that respond to multiple oxidants and report overall oxidative burden, the entries do not discriminate individual ROS species unless explicitly stated.

Species (Strain)	Organ	AMD Dose	Duration	ROS/MDA/LPO	SOD	CAT	GPx	GSH	Reference
Rat (albino)	Kidneys and serum	30 mg/kg/day oral gavage	8 weeks (chronic)	↑ MDA level	↓ SOD activity	-	-	-	[55]
Rat (Fischer-344)	Lung, Liver, Spleen, Kidney, Heart	60 mg/kg/day i.p. injection	21 days (sub-chronic)	↑ MDA level	-	-	-	-	[7]
Rat (Sprague–Dawley)	Lung	80 mg/kg/day i.p. injection	1–4 weeks (chronic)	↑ MDA level	↓ SOD activity	↓ CAT activity	↓ GPx activity	↓ GSH level	[12]
Rat (Sprague–Dawley)	Heart	100 mg/kg/day oral gavage	7 days (acute)	↑ LPO level	↓ SOD activity	↓ CAT activity	-	↓ GSH level	[56]
Rat (Sprague–Dawley)	Lung	40 mg/kg/day oral gavage	4 weeks (chronic)	↑ MDA	↓ SOD activity	-	-	↓ GSH level	[57]
Rat (Sprague–Dawley)	Liver	40 mg/kg/day oral gavage	8 weeks (chronic)	↑ ROS and ↑ MDA levels	↓ SOD activity	↓ CAT activity	↓ GPx activity	-	[58]
Rat (Sprague–Dawley)	Lung	80 mg/kg/day i.p. injection	7 and 14 days (acute) 21 and 28 days (sub-chronic)	↑ MDA level	↑ SOD (7 and 14 days), ↓ SOD (21 and 28 days) activity	↓ CAT activity	↓ GPx activity	↓ GSH level	[12]
Rat (Swiss albino)	Lung	100 mg/kg/day i.p. injection	10 days (acute)	-	↓ SOD activity	↓ CAT activity	-	↓ GSH level	[59]
Rat (Wistar)	Liver, Kidney, Testis	0.17 mg/kg/day oral gavage	30 days (sub-chronic)	↑ LPO level	↑ SOD activity	↑ CAT activity	-	-	[60]
Rat (Wistar)	Lung	30 mg/kg/day oral gavage	60 days (chronic)	↑ MDA level	↓ SOD activity	↓ CAT activity	↓ GPx activity	↓ GSH level	[11]
Rat (Wistar)	Lung	30 mg/kg/day oral gavage	6 weeks (chronic)	↑ MDA level	↓ SOD activity	-	-	↓ GSH level	[40]
Rat (Wistar)	Liver	20 mg on day 1 and a maintenance dose of 8 mg from day 2 to day 6 oral gavage	6 days (acute)	↑ LPO level	-	-	-	↓ GSH level	[10]
Rat (Wistar)	Lung	40 mg/kg/day oral gavage	4 weeks (chronic)	↑ MDA level	-	-	-	↓ GSH level	[61]
Rat (Wistar)	Lung	30 mg/kg/day oral gavage	4 weeks (chronic)	↑ ROS and ↑ MDA levels	-	-	-	-	[62]
Rat (Wistar)	Optic nerve tissue	50 mg/kg/day and 100 mg/kg/day oral gavage	14 days (acute)	↑ MDA level	↓ SOD activity	↓ CAT activity	-	↓ GSH level	[63]
Mouse (C57BL/6)	Liver	1000 mg/kg p.o. daily	3 days (acute)	↑ LPO level	-	-	-	↓ GSH level	[42]
C. elegans (N2 and CF1553 strains)	Whole worm	10 µM in media	24 h	↑ ROS level	↑ SOD3 expression	-	-	-	[43]

Table 3. AMD impact on oxidative stress markers in humans. Abbreviations: ↑—increase; ↓—decrease; AF—atrial fibrillation; CAT—catalase; CABG—coronary artery bypass grafting; GSH—glutathione; GPx—glutathione peroxidase; LPO—lipid peroxidation; MDA—malondialdehyde; NO—nitric oxide; ROS—reactive oxygen species; SOD—superoxide dismutase.

Population/Clinical Context	AMD Dose	Duration	Effect on Oxidative Stress	Source
Patients with arrhythmia and pulmonary toxicity	340–440 mg/day (oral)	12 months	↓ SOD	[64]
Healthy volunteers or stable patients	200 mg/day (oral)	4 weeks	↓ LPO; ↓ ROS	[65]
CABG patients (post-operative AF prevention)	4 × 400 mg day before surgery; 2 × 600 mg on surgery day; 2 × 400 mg/day for 4 days post-op (~6 g total)	5 days	↑ NO; no change in MDA, SOD, CAT, or GPx	[66]
Patients with autoimmune thyroid disease treated with AMD	Not specified, chronic oral therapy	6 months	↑ MDA, ↓ GSH and ↓ SOD; redox imbalance observed	[67]
Heart failure patients	Not specified, chronic oral therapy	6 months	↓ SOD, ↓ CAT, ↓ GSH, ↓ MDA	[68]

Table 3. AMD impact on oxidative stress markers in humans. Abbreviations: ↑—increase; ↓—decrease; AF—atrial fibrillation; CAT—catalase; CABG—coronary artery bypass grafting; GSH—glutathione; GPx—glutathione peroxidase; LPO—lipid peroxidation; MDA—malondialdehyde; NO—nitric oxide; ROS—reactive oxygen species; SOD—superoxide dismutase.

Population/Clinical Context	AMD Dose	Duration	Effect on Oxidative Stress	Source
Patients with arrhythmia and pulmonary toxicity	340–440 mg/day (oral)	12 months	↓ SOD	[64]
Healthy volunteers or stable patients	200 mg/day (oral)	4 weeks	↓ LPO; ↓ ROS	[65]
CABG patients (post-operative AF prevention)	4 × 400 mg day before surgery; 2 × 600 mg on surgery day; 2 × 400 mg/day for 4 days post-op (~6 g total)	5 days	↑ NO; no change in MDA, SOD, CAT, or GPx	[66]
Patients with autoimmune thyroid disease treated with AMD	Not specified, chronic oral therapy	6 months	↑ MDA, ↓ GSH and ↓ SOD; redox imbalance observed	[67]
Heart failure patients	Not specified, chronic oral therapy	6 months	↓ SOD, ↓ CAT, ↓ GSH, ↓ MDA	[68]

5. Potential AMD Impact on Nuclear Receptors and Its Links to Oxidative Stress

AMD has been shown to modulate multiple xenobiotic-sensing nuclear receptors, with downstream effects on detoxification enzymes and cellular redox status [30,69]. For example, in HepG2 cells, AMD (10 µM) markedly activated the aryl hydrocarbon receptor (AhR) pathway, increasing CYP1A1 and CYP1A2 activities [30]. AhR–ARNT heterodimers bind xenobiotic response elements (XREs) in genes like CYP1A1, driving their expression [70]. Co-treatment with an AhR antagonist (CAY10464) abolished AMD-induced CYP1A1/1A2 induction but did not mitigate AMD cytotoxicity [30]. Notably, AhR blockade raised CAT activity and GSH levels in AMD-treated cells [30], suggesting that AhR signaling modulates antioxidant defenses. In sum, these data indicate that AMD is an AhR agonist that triggers phase I enzyme induction, yet AhR activation appears secondary to toxicity. Instead, AhR may play a compensatory role in oxidative stress, e.g., via nuclear factor erythroid 2–related factor 2 (Nrf2) crosstalk, as its inhibition exacerbates oxidative markers [30,71].

The constitutive androstane receptor (CAR) and pregnane X receptor (PXR) are promiscuous xenobiotic receptors that heterodimerize with RXR and bind direct-repeat DNA motifs in drug-metabolizing genes [72]. Both CAR and PXR regulate CYP2B, CYP3A and other enzymes that can generate ROS as byproducts [73]. Recent studies suggest that CAR enhances AMD metabolism and alters redox balance. In HepaRG hepatocytes, stable CAR overexpression markedly increased energy and xenobiotic metabolism, e.g., CYP3A4 activity, while lowering basal mitochondrial superoxide levels [69]. However, when these CAR-overexpressing cells were exposed to AMD, they showed the highest accumulation of mitochondrial superoxide [69]. This means that CAR activation reduces resting ROS but amplifies AMD-induced mitochondrial ROS. This implies that CAR-driven upregulation of oxidative phosphorylation and drug metabolism sensitizes cells to the mitochondrial toxicity of AMD [69]. PXR activation which similarly induces CYP enzymes could also impact AMD metabolism and ROS. Indeed, pharmacological inhibition of PXR/CAR with allyl isothiocyanate (AITC) protected cells from AMD cytotoxicity [74]. AITC suppressed PXR/CAR transactivation and CYP3A4/CYP2B6 expression, reversing AMD-induced cell death [74]. These findings imply that AMD toxicity is potentiated by PXR/CAR-driven CYP induction. Inhibition of the above-mentioned receptors and thus CYP-mediated metabolism reduces AMD-induced oxidative injury [74]. In summary, CAR/PXR pathways enhance AMD biotransformation via DR4-bound target genes and thereby can increase ROS production during detoxification [69,74].

Peroxisome proliferator-activated receptors (PPARs) are lipid-sensing transcription factors forming PPAR–RXR heterodimers on peroxisome proliferator response elements (PPREs) that regulate energy metabolism and influence redox homeostasis [75]. Evidence from animal and cellular models indicates that PPARα and PPARγ modulate AMD toxicity. In mice, AMD induced hepatic steatosis and upregulated PPARα target genes, despite AMD itself not directly activating PPARα [20]. Crucially, PPARα-knockout mice suffered more severe AMD hepatotoxicity results in greater weight loss and AST elevation than wild-type mice, whereas the PPARα agonist fenofibrate mitigated these effects [20]. This implies that PPARα activation is protective against AMD-induced liver injury [20], likely by promoting fatty acid oxidation and antioxidant pathways. In contrast, PPARγ appears to be stimulated by AMD. In human bronchial epithelial cells, AMD significantly upregulated PPARG mRNA and PPARγ-dependent lipid metabolism genes such as fatty acid desaturase 2 (FADS2) and stearoyl-CoA desaturase (SCD) [21]. Inhibition of PPARγ with GW9662 attenuated AMD-induced expression of SCD and FADS2 and reduced phospholipid accumulation, indicating that AMD engages the PPARγ signaling axis in lung cells [21]. PPARγ is known to induce antioxidant enzymes, e.g., PPARγ agonists greatly increase CAT expression via PPARγ response elements in the CAT gene [22] and induce the extracellular glutathione peroxidase (GPx3) [23]. Thus, PPARγ activation raises cellular defenses (CAT, GPx) against oxidative stress. By upregulating PPARγ, AMD may provoke both lipid remodeling and an adaptive increase in antioxidant capacity.

Overall, nuclear receptors act as key switches in the oxidative-stress response. Clinically, AMD administration causes oxidative injury evidenced by depletion of CAT, SOD, GPx, and GSH and increased MDA in animal lungs [59,76]. PPARs drive antioxidant gene expression in which PPREs are present such as CAT and GPx3 [22,23], while AhR/Nrf2 crosstalk influences phase II defense enzymes. In sum, AMD alters AhR, CAR/PXR, and PPAR signaling. The above-mentioned changes affect CYP-mediated ROS production and the transcriptional regulation of antioxidant enzymes via XRE and PPRE elements. The net result is a complex interplay in which some receptor activations (e.g., PPARα) mitigate AMD-induced ROS, whereas others (e.g., CYP induction by PXR/CAR) could amplify it. These mechanisms have been documented in cellular and animal studies [20,21,30], underscoring the importance of nuclear–receptor-mediated transcriptional networks in AMD-associated oxidative stress (Figure 2).

The impact of AMD on molecular pathways regulating oxidative stress appears to be both dose- and tissue-dependent and includes interference with other transcriptional and signaling regulators of antioxidant enzymes such as Nrf2, nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB), and mitochondrial biogenesis pathways [57,77].

One of the most critical regulators of antioxidant gene expression is the Nrf2/Keap1 system [78]. Under basal conditions, Nrf2 is sequestered in the cytoplasm by Keap1 and targeted for degradation. In response to oxidative stress, Nrf2 translocates to the nucleus and upregulates genes encoding for SOD, CAT, GPx, and enzymes involved in GSH biosynthesis [79]. Evidence from experimental models suggests that AMD may inhibit Nrf2 activation in pulmonary and hepatic tissues [29]. In rat models, prolonged AMD exposure reduced Nrf2 nuclear translocation, resulting in lower SOD1 and GPx1 expression and increased lipid peroxidation [80,81]. Clinical observations in patients with AMD-induced pulmonary toxicity support this, showing reduced SOD activity and GSH levels in erythrocytes and alveolar macrophages [57,64]. However, some short-term, low-dose exposures have paradoxically increased Nrf2-dependent gene expression, suggesting that AMD may have hormetic effects on this pathway [65].

NF-κB is another transcription factor influenced by AMD. NF-κB activation can both induce and respond to ROS generation [82]. It has been reported that AMD can indirectly affect NF-κB activity, resulting in elevated levels of TNF-α and inducible nitric oxide synthase (iNOS), thereby contributing to oxidative damage [83,84,85]. Importantly, NF-κB may also negatively regulate Nrf2, compounding the suppression of antioxidant defenses during chronic AMD exposure [77]. This crosstalk creates a pro-inflammatory and pro-oxidative feedback loop, particularly in tissues like the lung and thyroid, where AMD accumulates and where its lipophilic metabolites interfere with redox-sensitive transcriptional networks [86,87].

AMD directly affects mitochondrial bioenergetics by impairing complex I and II activity, which not only increases mitochondrial ROS (mtROS) but also downregulates mitochondrial biogenesis via PPARγ coactivator 1 alpha (PGC-1α) inhibition [68]. PGC-1α, a coactivator of NRF1 and mitochondrial transcription factor A (TFAM), also modulates antioxidant gene transcription. AMD-mediated suppression of this axis results in decreased mitochondrial SOD2 expression, as observed in optic neuropathy and myopathy models [77]. Furthermore, accumulation of AMD in mitochondrial membranes destabilizes redox signaling and impairs calcium buffering, further exacerbating oxidative imbalance [11].

There is emerging evidence that AMD can influence antioxidant defense not only transcriptionally, but also via epigenetic mechanisms such as, e.g., histone acetylation of antioxidant gene promoters and post-translational modifications of enzymes like SOD and GPx such as, e.g., nitration and oxidation [86,88]. While data in humans are limited, AMD-induced oxidative stress has been associated with decreased activity of antioxidant enzymes even when their gene expression remains unchanged, suggesting enzyme inactivation rather than transcriptional repression [66]. This explains the discrepancies in publications where the enzyme expression does not correlate with its activity [30].

6. Conclusions and Future Perspectives

The life-saving antiarrhythmic efficacy of AMD is counterbalanced by its complex, multi-organ toxicities, now firmly linked to excessive ROS production and oxidative injury. A consistent theme emerging from decades of research is that AMD-induced ROS generation and toxicity are highly time-, dose-, and model-dependent. Acute or low-dose exposures in certain models can even display paradoxical antioxidant effects—for example, short-term AMD scavenged hydroxyl radicals in cardiomyocytes—whereas prolonged or high-dose AMD exposure invariably shifts the balance toward net oxidative stress and damage (Figure 3). Such context-specific outcomes help explain early contradictory findings and underscore that experimental conditions critically dictate the redox effects of AMD.

Mechanistically, AMD instigates oxidative stress through multiple direct and indirect pathways. Directly, AMD accumulates within cells and disrupts mitochondrial respiration, causing electron transport leakage and superoxide formation. This mitochondrial dysfunction is a primary ROS source in both pulmonary and hepatic models. AMD metabolism can also yield reactive radical intermediates via redox cycling. These radicals trigger lipid peroxidation in cellular membranes, evidenced by increased malondialdehyde and membrane damage in AMD-exposed cells. Concurrently, AMD and its ROS byproducts impair antioxidant defenses by inactivating crucial enzymes (SOD, CAT, GPx) and depleting glutathione stores. The resulting loss of antioxidant protection creates a self-amplifying loop wherein ROS-induced enzyme damage begets further ROS accumulation. Indirectly, AMD perturbs cellular signaling networks that further promote oxidative stress. It can activate or modulate nuclear receptors—AhR, PXR, CAR, and PPARs—which regulate genes for drug metabolism and antioxidant responses. For instance, AMD-driven PXR/CAR activation induces CYP enzymes that enhance AMD metabolism to ROS-generating products, while PPARα and PPARγ pathways are altered in ways that can respectively mitigate or contribute to oxidative outcomes. In parallel, AMD stimulates pro-inflammatory signaling which elevates cytokines like TNF-α and can activate the angiotensin II pathway, leading to recruitment of ROS-producing immune cells and inflammatory oxidative cascades. These direct and indirect mechanisms act in concert to tip cells into a state of oxidative stress.

From a clinical standpoint, this knowledge encourages a multi-pronged approach to managing and possibly mitigating AMD side effects. Firstly, patient monitoring can be enhanced. Clinicians might watch for indirect signs of oxidative stress such as rising liver enzymes, inflammatory markers, or imaging changes in lungs even before overt toxicity, especially in patients on high doses or long duration of therapy. Secondly, judicious use of lowest effective doses and considering drug holidays might allow the body’s antioxidant systems to recover intermittently, although this must be balanced against arrhythmia control. Thirdly, the potential of co-therapy with antioxidants or mitochondria-protective agents deserves exploration in clinical trials. While evidence is not yet sufficient to recommend this routinely, the consistent success of antioxidants in experimental models suggests that patients at risk could benefit from such supportive therapy. Another angle is using anti-inflammatory strategies to reduce the baseline propensity for oxidative injury. Importantly, the insights from oxidative mechanisms open avenues for new drug development. Future antiarrhythmic compounds might be designed to retain the efficacy of AMD but without its toxic moieties, for example, analogues that do not accumulate in mitochondria or that have built-in antioxidant functionalities. The comparison of dronedarone to AMD has already shown some progress on this front: less accumulation and less ROS induction. Although the safety issues of AMD mean the problem is not yet fully solved. Research into novel mitochondria-targeted antioxidants could also find an application as adjuncts to AMD therapy, to specifically neutralize mitochondrial ROS.

Several questions remain and warrant further investigation. One area is the detailed mapping of the timeline of ROS vs. tissue changes. For instance, what molecular events mark the transition from mere oxidative stress to irreversible fibrosis in the lungs? Understanding this could yield early biomarkers that predict who will develop fibrosis. Another question is the role of genetic variability and differences in patients’ antioxidant enzyme genes or drug metabolism that could influence susceptibility to oxidative damage from AMD. Indeed, polymorphisms in manganese SOD or glutathione S-transferases might make some individuals more prone to AMD toxicity—this could be a research and precision medicine focus.

Future studies should also explore therapeutic interventions in models that more closely simulate the clinical scenario. For example, can starting an antioxidant therapy midway through an animal’s AMD exposure reverse or halt ongoing toxicity? This would mimic a patient who has been on AMD for some time and starts showing early toxicity signs—could intervention at that point salvage organ function? Some evidence from L-carnitine and corticosteroids in animals suggests it might. Additionally, combination therapies such as, e.g., a mitochondrial antioxidant and/or an anti-fibrotic agent, could be tested to see if they synergistically prevent end-stage damage like fibrosis, which is less reversible.