Is α-lipoic acid a scavenger of reactive oxygen species in vivo? Evidence for its initiation of stress signaling pathways that promote endogenous antioxidant capacity



The chemical reduction and oxidation (redox) properties of α-lipoic acid (LA) suggest that it may have potent antioxidant potential. A significant number of studies now show that LA and its reduced form, dihydrolipoic acid (DHLA), directly scavenge reactive oxygen species (ROS) and reactive nitrogen species (RNS) species and protect cells against a host of insults where oxidative stress is part of the underlying etiology. However, owing to its limited and transient accumulation in tissues following oral intake, the efficacy of nonprotein-bound LA to function as a physiological antioxidant has been questioned. Herein, we review the evidence that the micronutrient functions of LA may be more as an effector of important cellular stress response pathways that ultimately influence endogenous cellular antioxidant levels and reduce proinflammatory mechanisms. This would promote a sustained improvement in cellular resistance to pathologies where oxidative stress is involved, which would not be forthcoming if LA solely acted as a transient ROS scavenger. © 2008 IUBMB IUBMB Life, 60(6): 362–367, 2008


ARE, antioxidant response element; DHLA, dihydrolipoic acid; GLUT4, glucose transport protein 4; GSH, glutathione; IGF-1, insulin-like growth factor-1; IR, insulin receptor; LA, lipoic acid; PTP1B, protein tyrosine phosphatase-1B; RNS, reactive nitrogen species; ROS, reactive oxygen species; TNFα, tumor necrosis factor-α.


The dithiol compound, α-lipoic acid (1,2-dithiolane-3-pentanoic acid; LA), is a necessary protein-bound cofactor for mitochondrial α-ketoacid dehydrogenases and thus serves a critical role in mitochondrial energy metabolism. However, there is a growing awareness that LA is readily absorbed from the diet, and orally derived LA has unique biochemical activities separate from this normal metabolic function. Based largely on in vitro studies, LA has been touted as a potent biological antioxidant, that is, LA or dihydrolipoic acid (DHLA) terminates a number of reactive oxygen species (ROS) and reactive nitrogen species (RNS) and also chelates redox active transition metals (e.g. iron, copper) (1, 2). This salutary antioxidant role has been implicated as part of the mechanism underlying the clinical benefits for orally supplied LA as a diabetes medicine and its means to improve age-associated cardiovascular, cognitive, and neuromuscular declines in animal models (3–6). As excellent reviews exist, which summarize the functions of LA as a mitochondrial enzyme cofactor (7), we will mainly confine the present discussion to the evidence suggesting that (DH)LA solely acts as a true oxidant scavenger in vivo or whether its physiological “antioxidant” actions may actually be provided through induction of stress-activated signaling mechanisms. The authors apologize in advance for the omission of many important studies on LA function because of space and citation limitations.


Halliwell and Gutteridge have defined a true antioxidant as “any substance that, when present at low concentrations compared to those of an oxidizable substrate, significantly delays or prevents oxidation of that substrate” (8). In vitro studies suggest that LA fits well into this standard definition. Because of its juxtaposed thiol groups and a high reduction potential of the DHLA/LA redox pair (−320 mV), LA and DHLA are capable of scavenging a variety of ROS and RNS (1, 2). However, most of the evidence that support (DH)LA as a direct scavenger of ROS/RNS mainly comes from in vitro studies and there is little direct support that LA acts as a physiological antioxidant. Moreover, a thorough understanding of its bioavailability and metabolic disposition must be considered to properly define the role of orally supplied LA as a biological antioxidant. The following section briefly reviews the evidence that LA is readily absorbed from the diet but also does not extensively accumulate in tissues and is rapidly metabolized.


Though de novo synthesis supplies all LA needed for its function in intermediary metabolism, it can also be absorbed from foods (leafy green vegetables and meats) and dietary supplements. Even though little information currently exists as to the general intake of LA from dietary sources, controlled studies in rodents suggest that 20–40% of LA given orally is absorbed into the plasma (9, 10). Plasma LA levels typically reach a maximum between 0.5 and 2 h after intake in a variety of animal species studied (1), and a single 600 mg dose given to rats was shown to produce a peak plasma concentration of 13.8 μM. Rapid gastrointestinal transport of LA into the blood plasma is followed by an equally rapid clearance, reflecting both uptake into tissues, (e.g. the liver, brain, heart, and skeletal muscle) as well as glomerular filtration and renal excretion (11). Radiolabeled LA, either fed or administered to rodents by intraperitoneal injection, is rapidly metabolized and excreted. Up to 98% of the radiolabel can be recovered mostly in the urine within 24 h after LA administration (12). This rapid clearance appears to be a universal mammalian trait as the urinary recovery of 14C-labelled carbon from the parent compound is 54.7%, 71.8%, and 63.5% in mouse, rat, and dog, respectively (12).

In keeping with its rapid plasma clearance and excretion, LA undergoes extensive catabolism following its transport into the cells and tissues. A study examining the metabolic fate of LA in rodents and dogs using liquid chromatography/tandem mass spectrometry revealed 12 major metabolites recovered from blood plasma and urine (12). Additionally, this study confirmed the previous ones that β-oxidation is the major metabolic outcome in vivo, accompanied by oxidation of the dithiolane ring.


These metabolic studies correlate with our recent observations that nonprotein-bound LA only reaches ∼60 μM in rat liver following its gavage (40 mg/kg, b.w.) (T.M.H., unpublished observations). This is in agreement with other studies that also reported the limited accumulation of LA following its oral supplementation (10). Thus, the extensive metabolism and clearance of LA results in only a transient and limited accumulation of the parent compound in tissues following an oral dose. The small cellular availability of LA is significantly lower than other endogenous low molecular weight antioxidants [e.g. glutathione (GSH), ascorbic acid], which are found in millimolar concentrations in most tissues. Thus, argued from a mass action standpoint, it is difficult to conclude that LA/DHLA solely acts as a physiological antioxidant.

A possible reason for the disparity between low tissue accumulation of LA and its potent oxidant scavenging properties may stem from the extensive use of cultured cell models that often employ high supraphysiological LA concentrations relative to that seen in the plasma. These conditions may drive results toward an antioxidant role for LA which is not reflective of its true nature in vivo. Moreover, cell-based studies certainly do not replicate the single-pass clearance that LA undergoes systemically via its glomerular filtration from the plasma (11). In fact, Packer and coworkers used cultured primary human fibroblasts and Jurkat cells and reported the rapid reduction of LA to DHLA followed by its efflux into the media (13). This suggests that cells in culture may chronically recycle (DH)LA prior to its catabolism, which could maintain effective cellular LA levels for much longer periods than would be achieved physiologically.

Despite these caveats, it is clear that dietary LA provides a remarkable range of positive therapeutic benefits, which includes limiting the neurological complications from diabetes, lowering the age-associated increases in oxidative damage, and reducing the toxicity from free radical generating redox-active transition metals (14–17). All these outcomes implicate an antioxidant role for LA in vivo. How, then, can both the considerable evidence of LA's potent ROS/RNS scavenging capabilities and its limited cellular accumulation and rapid metabolism be reconciled? It has been hypothesized that orally administered LA indirectly increases the overall cellular antioxidant capacity to achieve its therapeutic benefits rather than act directly to scavenge ROS/RNS on a long-term basis. The following section will now briefly describe the mounting evidence that orally supplied LA induces a variety of cellular signaling pathways and transcriptional arrays to achieve a more long-lasting antioxidant and antiinflammatory benefit to tissues than could be achieved if LA were only a transient nutritional antioxidant.


GSH is the most abundant low molecular weight cellular antioxidant, buffering the thiol redox state. It is synthesized in all mammalian cells and its levels are tightly controlled by transcriptional regulation of GSH synthesis genes, by substrate availability, and by allosteric feedback inhibition. GSH levels can thus be rapidly induced to meet both acute and chronic toxicological and oxidative challenges. There is mounting evidence from both in vitro and more physiological studies that LA increases or maintains cellular GSH levels by acting as a transcriptional inducer of genes governing GSH synthesis and, potentially, by increasing substrate availability.

LA as a Transcriptional Modulator of GSH Synthesis Genes

Nrf2 is a key transcription factor that mediates the expression of antioxidant and detoxification genes regulated by the antioxidant response element (ARE) (18, 19). Numerous studies show that nuclear Nrf2 levels determine the expression of these genes, including those for GSH synthesis. This tight regulation is mediated through rapid degradation of Nrf2, which is effected by its binding to Keap1, an actin-associated protein that bridges Nrf2 to the Cul3 ubiquitin ligase (20). Keap1 contains several redox active cysteine residues, which upon their oxidation or electrophilic adduction disrupt the ability of Keap1 to facilitate Nrf2 degradation. This acutely increases Nrf2 nuclear availability and activation of ARE-containing genes. As a number of redox-active low molecular weight thiol compounds (e.g. sulforaphane, pyrrolidine dithiocarbamate, oltipraz) increase ARE-regulated gene expression through the transient modulation of critical Keap1 thiols (21), it is plausible that LA may act in a similar manner and at low concentrations to modulate the redox-sensitive cysteine groups on Keap1. Thus, a theoretical framework exists where LA may induce the endogenous antioxidant and detoxification genes through its potent redox nature.

We recently showed that LA treatment increases hepatic nuclear Nrf2 levels and induces Nrf2-mediated gene transcription in vivo (22). Additionally, supplementing the diets of old rats with (R)-α-lipoic acid (40 mg/kg, b.w.) for 2 weeks reversed the age-related decline in hepatic GSH levels (22). LA ultimately increased intracellular GSH by inducing the transcription of both the catalytic and regulatory subunits of γ-glutamylcysteine ligase, which is the rate-controlling enzyme for GSH synthesis. The improved GSH status afforded by LA also reversed the age-related increased susceptibility of hepatocytes to tert-butyl hydroperoxide, a model alkylhydroperoxide that is detoxified in a GSH-dependent manner (14). Thus, LA appears to be an inducer of Nrf2-mediated antioxidant gene expression, which significantly increases the cellular capacity to synthesize GSH.

LA as a Stimulator of Substrate Availability for GSH Synthesis

In addition to elevating the expression of enzymes involved in GSH synthesis, LA may also heighten GSH levels through its ability to elevate cysteine uptake (23, 24). Packer and coworkers showed that DHLA reduced intracellular cystine to cysteine when given to cultured cells, thereby increasing the rate of GSH synthesis, as cysteine is the rate-limiting substrate for the reaction (25). This is unlikely to be an artifact of tissue culture conditions as we also observed that LA reverses the age-related decline in myocardial GSH by raising cysteine availability to this organ in vivo (26). The LA-induced reduction of cystine to cysteine, while seemingly distinct from its Nrf2-inducing effect, may actually be one in the same mechanism. The expression of the xmath image transport protein responsible for cystine uptake into tissues is regulated partly in an ARE-dependent manner. Thus, diet-derived LA may enhance both GSH synthetic capacity and the intracellular levels of cysteine.

These results suggest that at least one plausible mechanism whereby diet-derived LA improves the cellular antioxidant capacity may be through induction of endogenous antioxidant levels (e.g. GSH), which would sustain an antioxidant effect far longer than if LA solely acted as a direct scavenger of ROS/RNS. In fact, there is mounting evidence that LA affects other important antioxidant and antiinflammatory pathways, which buttresses the concept of LA as a transient but potent mediator of stress response signaling. These additional mechanisms will now be described.


Thiol-modifying compounds, such as the Keap1 modifiers noted earlier, may affect proteins through the redox modulation of critical cysteine residues, which results in the activation (e.g. kinases) or inhibition (e.g. protein tyrosine phosphatases) of protein function. There is strong evidence that LA is a modifier of critical protein thiolates and thus may influence a host of pathways sensitive to thiol redox state (27) (see Fig. 1).

Figure 1.

Modulation of protein function by thiol disulfide exchange between lipoic acid and critical redox-sensitive protein sulfhydryls.


Oxidative stress-associated inflammation is thought to provoke early vascular events in atherogenesis, including the upregulation of vascular adhesion molecules and matrix metalloproteinase activity. These events require the activation of NF-κB, a transcription factor that induces the expression of many genes involved in inflammation and endothelial cell migration. Given the prooxidant nature of inflammatory processes, therapeutic strategies aimed at mitigating oxidant production and oxidative damage have been investigated for decades. LA is widely known as an inhibitor of NF-κB (28). Interestingly, LA limits TNF-α-induced NF-κB activation and adhesion molecule expression in human aortic endothelial cells via a mechanism seemingly distinct from traditional endogenous antioxidants, such as ascorbate or reduced GSH (29, 30). Thus, LA/DHLA may reduce the proinflammatory conditions by its interaction with NF-κB.


Many of the genes involved in mediating the cellular stress resistance are linked to the insulin signaling pathway. The binding of a ligand such as insulin to its receptor creates its own localized burst of H2O2 that causes the autophosphorylation of the tyrosine kinase domain on the insulin receptor (IR), initiating a signaling cascade (31). LA is similar to insulin in its ability to activate signaling molecules in the insulin/insulin-like growth factor-1 (IGF-1) pathway, though it may not work as a ligand. In a similar manner with its interaction with Keap1, LA or DHLA may thus modulate the critical cysteine residues in the IR beta-subunit and other critical thiol-containing proteins of the insulin signaling pathway.

One key enzyme that is activated by the insulin/IGF-1 pathway is protein kinase B/Akt, a serine/threonine kinase that mediates cell survival through regulating the phosphorylation state of stress response enzymes and by limiting apoptosis. LA induces Akt phosphorylation in human umbilical vascular endothelial cells (32) and the THP-1 human monocyte cell line (33). Using isolated rat hepatocytes, Diesel et al. (34) also showed LA-induced PI3K activation, but suggested a mechanism different than as a thiol redox modulator. Using computer modeling and a kinase assay, this group showed that LA may bind to the tyrosine kinase domain of the IR β-subunit, suggesting that signaling is initiated through receptor-protein modification rather than a direct effect on Akt itself or adaptor substrate molecules.

Even though LA interaction with the insulin signaling pathway is now well recognized (35), it is not clear which proteins are targets of LA action. In a preadipocyte line, Cho et al. (36) showed that although Akt became phosphorylated in response to LA within 30 min, the IR and insulin receptor substrate-1 (IRS-1) protein did not. In 3T3-L1 adipocytes, however, Moini et al. (37) found that IR was indeed phosphorylated after LA treatment, and this was specific to LA and not DHLA. Klip and coworkers (38) found using the same cell type that both IRS-1 and PI3K were phosphorylated following LA treatment.


Work from the Rhee lab has long shown that protein phosphorylation is modulated by cysteine oxidation (39). Alternatively, or in addition to stimulating the IR, LA appears to inactivate cellular protein tyrosine phosphatases (PTP1B), thereby preventing the otherwise inhibitory dephosphorylation of the IR tyrosine kinase domain (40). The observed inhibition of PTP1B with H2O2 and LA coincided with decreased sulfhydryl content, again suggesting that LA either directly modified critical thiol groups on this phosphatase or induced oxidants that did the same. Other phosphatases may be inhibited by LA, which include protein phosphatase-2A, an important serine/threonine enzyme involved in insulin-mediated and other signaling processes (41).

In summary, there is a growing body of evidence that (DH)LA, potentially through its redox nature, modulates important signal transduction pathways that ultimately increase the endogenous cellular antioxidants (e.g. GSH), lower inflammation (e.g. NF-κB), and improve the chances for cell survival (e.g. IR/Akt) in times of insult. Induction of kinases and transcription factors or the inhibition of phosphatases comprise classic examples of signal amplification that would have longer term effects than if LA acted solely as a transient oxidant scavenger. Moreover, very small concentrations of LA could affect cell-signaling systems, which is consistent with the limited accumulation of LA in tissues following its oral intake. More studies are warranted to further weigh the importance of dietary LA as an antioxidant and an inducer of endogenous antioxidant capacity in vivo.


While much of the work on LA has used animal models to define its function, there have also been several important trials that examine the potential health benefits of LA. Many of the outcomes for the therapeutic use of LA in humans lend support to its general role as an antioxidant. However, there is little current evidence that LA produces this effect by its antioxidant scavenging properties or by its modulatory action on cell signaling networks that maintain endogenous antioxidant capacity. The strongest evidence for diet-derived LA as a therapy comes from studies on type 2 diabetes.

LA has been used in Germany for over 30 years in the treatment of diabetic polyneuropathies (35). The ALADIN and SIDNEY clinical trials of LA showed that its oral or intravenous administration improves nerve conduction velocity and neuropathic symptoms, such as pain, burning, paresthesia, and numbness (42–45). To date, the therapeutic use of LA in the treatment of diabetic polyneuropathies remains the best documented and the most significant benefit of LA to human health. Furthermore, a substantial body of experimental evidence has accumulated, which not only supports a role for LA as a mitigator of oxidative stress in this disease but also for its means to affect glucose handling (46, 47). A number of reports now show that LA improves glucose disposal in patients with type 2 diabetes receiving LA (48, 49).

Animal studies, which augment these clinical trials, show that LA improves skeletal muscle glucose uptake, whole-body glucose tolerance, and is helpful against insulin resistance (50). This improved glycemic handling that LA affords may stem from its interaction with regulatory components of the insulin signaling cascade (see earlier). Mechanistically, LA stimulates the recruitment of glucose transport protein 4 (GLUT4) from its storage site in the Golgi to the sarcolemma. This eventually facilitates glucose uptake by increasing the number of cell surface GLUT transporters. Evidence from cell culture experiments further supports the role of insulin-mediated PI3K activity in LA-induced glucose uptake, notably the sensitivity of this effect on wortmannin, a PI3K inhibitor (38, 51). However, direct evidence of the role played by GLUT4 translocation in improving glucose disposal using a physiologically relevant LA concentration is still needed.


The potential of LA to work as a therapeutic agent appears to lie not only in its actions as a direct scavenger of ROS/RNS, but also in its ability to affect signaling cascades. Because of the high electron density of the dithiolane ring (7), (DH)LA could theoretically react with the redox-sensitive cysteine groups of Keap1, PTP1B (40), IRS-1 (38), and others. While LA is generally known as a potent antioxidant, it may also become a prooxidant under the right conditions (52). These redox properties actually strengthen the argument that LA may act as a weak and transient modifier of the cellular thiol redox environment. Thus a more proper view for diet-derived LA may not be as a simple antioxidant, but more as a weak hormetic agent that induces similar protective mechanisms that would be normally induced in times of stress. Stimulation by LA of signal transduction molecules containing critical cysteine residues may thus prove to be clinically useful and should be explored more thoroughly in the future.

Another important consideration for future research is the differential effects of LA in vitro and in vivo. Although there is still much to learn about the metabolic fate of LA and the effects of the various metabolites on cells, it is clear that LA is rapidly cleared from plasma and tissues, whereas (DH)LA may remain in cell culture media for as long as the researcher allows. There is some evidence that (DH)LA cycles in culture, and may exacerbate the observed antioxidant and prooxidant effects. This serves to underscore the importance of using in vivo models to confirm the data obtained on LA in vitro. Clinical work on LA should be conducted with these considerations in mind.


The authors gratefully acknowledge the helpful discussions and comments provided by Judy A. Butler and Brian M. Dixon. Work on LA was supported by grants from the National Institutes of Health RO1 ZAG17141 and PO1 AT002034-01.