The Role of Oxidative Stress in Chronic Obstructive Pulmonary Disease

Abstract

Tobacco smoke is the number one risk factor for chronic obstructive pulmonary disease (COPD) and contains a high concentration of oxidants. The lung has a high concentration of antioxidants and antioxidant enzymes; however, COPD patients show evidence of increased oxidative stress suggesting that endogenous antioxidants may be insufficient to prevent oxidative damage from cigarette smoke. The consequences of increased oxidative stress in the lung include increased transcription of inflammatory genes, increased protease activity, and increased mucus secretion. Oxidative stress is often associated with impaired skeletal muscle function and may be one of the causes of glucocorticoid resistance. While current pharmacologic approaches to the treatment of chronic obstructive pulmonary disease do not commonly include antioxidants, preclinical studies involving animal models suggest that antioxidant superoxide dismutase mimetics offer a potential new therapeutic approach to the prevention and treatment of chronic obstructive pulmonary disease.

• Reactive oxygen species
• Reactive nitrogen species
• Antioxidant

Introduction to COPD

Chronic obstructive pulmonary disease (COPD) is the fourth leading cause of death in the world and is the fourth leading cause of chronic morbidity and mortality in the United States [[1]]. COPD is a disease state that is characterized by airflow limitation that is not fully reversible and is associated with progressive decline in lung function as well as an abnormal inflammatory response. Clinically the disease is characterized by emphysema, which is the loss of lung parenchyma, and chronic bronchitis. The diagnosis of COPD is confirmed by spirometry. Typically the patients have an FEV₁ of < 80% in combination with an FEV₁/FVC ratio < 70%. The clinical signs of COPD include shortness of breath, as well as a cough with sputum production. The prevalence of COPD has been increasing over the past thirty years. Globally the prevalence rate in the mid 1990s in males was 9.3/1000 and in females was 7.3/1000 [[2]]. The morbidity of COPD increases with age and is greater in men than women. Furthermore, it is predicted that in the coming decades COPD mortality will rise from its current position as the fourth leading cause of death in the world. It is estimated that COPD will be the fifth leading cause of disability‐adjusted life years lost worldwide [[2]].

Overview of the Theories of Pathogenesis of COPD

The major risk factor for COPD is exposure to tobacco smoke. In the United States 80–90% of COPD patients are current or former tobacco smokers. The other well established risk factor for COPD is α1‐antitrypsin deficiency; however only an estimated 1–2% of COPD patients are α1‐antitrypsin deficient [[3]][[4]]. The incidence of α1‐antitrypsin disease associated alleles in white populations is only 3–4%, thus α1‐antitrypsin only accounts for a small minority of the pathogenesis of COPD. Other risk factors for developing COPD include environmental tobacco smoke and heavy exposure to occupational dust or other irritants, and air pollution.

Only 10–20% of heavy smokers develop COPD, indicating that there must be other factors that determine whether long‐term exposures to noxious agents cause the disease. Although the link between tobacco smoke and COPD is well established, there is an incomplete understanding as to the molecular mechanisms by which tobacco smoke leads to COPD. Research on α1‐antitrypsin deficiency has shown that an imbalance in proteases and anti‐proteases can play a key role in the development of COPD. Additional hypotheses are that tobacco smoke and other unknown factors lead to increased death or apoptosis of lung cells and impaired regeneration of healthy lung tissue, eventually leading to emphysema. It has also been proposed that there is a predisposition towards inflammation from tobacco smoke and infections may underlie the enhanced progression of loss of lung function. Finally, cigarette smoke has trillions of oxidants in each puff [[5]] and some believe that these oxidants are a major cause of COPD. The role of oxidative stress in the pathogenesis of COPD is the subject of this review.

Definitions of Oxidative and Nitrative Stress

Airways are unique in both their exposure to high levels of environmental oxidants and their unusually high concentration of extracellular antioxidants. In the resting state, the balance between antioxidants and oxidants is sufficient to prevent the disruption of normal physiologic functions; however, either increases in oxidants or decreases in antioxidants can disrupt this balance. The imbalance between oxidants and antioxidants is termed oxidative stress (Fig. 1). There is increasing evidence that oxidative stress plays a role in the pathogenesis of COPD.

Figure 1. Oxidative stress is an imbalance between reactive oxygen species (ROS) and antioxidants. (A) Under normal conditions there are sufficient antioxidants in the lung to overcome endogenous ROS. (B) During pathologic states there is an increase in ROS that may be balanced by increase in antioxidants. (C) When either ROS production is excessive or antioxidants are inadequate, the balance is tipped toward oxidative stress and injury from ROS may occur. (Full color version available online.)

Both reactive oxygen (ROS) and reactive nitrogen species (RNS) contribute to the imbalance seen in COPD. Common ROS include superoxide, hydrogen peroxide, and hydroxyl radical. RNS include nitric oxide (NO) and its derivatives such as nitrogen dioxide and peroxynitrite. Other molecules that contribute to oxidative stress include protein radicals and lipid peroxide radicals. These molecules can cause nonspecific damage to cells and extracellular matrix, yet production of ROS and RNS in highly localized domains is also recognized as essential for normal physiologic function; for instance superoxide is a crucial component of phagocytic cell killing [[6]] and NO mediates smooth muscle relaxation in both blood vessels [[7]] and airways [[8]].

The features that make oxygen ideal for aerobic energy production (i.e. atmospheric abundance and a high affinity for electrons) are also its Achilles' heel. For instance, one electron addition to oxygen produces superoxide (O₂⁻), a second electron yields hydrogen peroxide (H₂O₂), and a third electron leads to the formation of the hydroxyl radical (OH•). These three intermediates are called reactive oxygen species because they readily react with other molecules such as proteins, lipids, and DNA (Fig. 2). The hydroxyl radical (OH•) is the most reactive of all reactive oxygen species (extremely high rate constants > 10⁹ M^{− 1}s^{− 1} for reactions with most molecules) and typically reacts immediately with organic molecules at the site of production [[9]]. When O₂ gains a fourth electron, it has been fully reduced to water, and no longer readily reacts with other molecules. Other species of oxidants include: peroxyl (RO₂^•), alkoxyl (RO^•), and hydroperoxyl (HO₂^•) free radicals, hypochlorous acid (HOCl), ozone (O₃), singlet oxygen, and peroxynitrite (ONOO⁻). The term free radical denotes molecules with at least one unpaired electron. Oxygen (two unpaired electrons), superoxide, hydroxyl radical, and NO (each with one unpaired electron) are examples of free radicals. Hydrogen peroxide is not a free radical because all of its electrons are paired.

Figure 2. Derivation of reactive oxygen species. Sequential electron (e⁻) addition to oxygen results in the formation of reactive oxygen species (red): superoxide (O₂‐), hydrogen peroxide (H₂0₂) and hydroxyl radical (OH). Superoxide can combine with nitric oxide (NO) to form peroxynitrite (ONOO‐). Both O₂⁻ and OH⁻ can initiate lipid peroxidation to form lipid peroxides (ROO) that can propagate free radical chain reactions. (Full color version available online.)

Exogenous Sources of Oxidants

Cigarette Smoke

Cigarette smoke contains > 4700 chemicals including hydrogen peroxide, superoxide, hydroxyl radical and nitric oxide [[5]][[9]][[10]] (Fig. 3). The constituents of cigarette smoke are often separated into a gas and particulate (tar) phase; both components contain abundant free radicals. The gas phase is less stable, but has been estimated to contain over 10¹⁵ free radicals per puff. The tar phase is more stable and has been estimated to contain over 10¹⁷ free radicals per gram [[9]]. Quinones, semiquinones, hyrdroquinones, and metals are prominent components of smoke that may propagate reactions that produce high levels of ROS and RNS. Iron may also catalyze the production of free radicals through Fenton or Haber–Weiss reactions. The oxidants in cigarette smoke are postulated to injure lungs via several mechanisms including: 1) direct damage to lipids, nucleic acids, and proteins (see subsequent section), 2) depletion of crucial antioxidants such as glutathione, 3) initiation of redox‐cycling reactions, 4) promotion of carcinogenesis, 5) enhancement of the respiratory burst in macrophages and neutrophils, 6) inactivation of protease inhibitors such as α1 antitrypsin inhibitor, 7) a decrease in the binding affinity and translocation of steroid receptors, and 8) impairment of skeletal muscle dysfunction [[9]].

Figure 3. Oxidants in cigarette smoke. Nitric oxide (NO), nitrogen dioxides (NO₂), carbon monoxide (CO), ammonia (NH₃), cyanide (HCN), hydrogen peroxide (H₂O₂), superoxide (O₂⁻), and peroxynitrite (ONOO‐). (Full color version available online.)

Air Pollution

In December of 1952 in London there was a weeklong episode of severe air pollution associated with excess mortality. The majority of the excess deaths were due to exacerbations of bronchitis and the relative risk of death during this period was 9.3 [[11]]. This event encouraged epidemiologists to document an association between the prevalence of COPD with geographic areas of high pollution. Air pollution contains a wide variety of chemicals that vary from one microenvironment to the next. Many of these particles contain reactive species such as nitrogen dioxide or ozone. The particulate matter in air pollution contains a large number of soluble metals including transition metals that are capable of redox cycling through the Fenton or Haber–Weiss reactions [[12]]. Other components in air pollution that cause oxidative stress include polycyclic aromatic hydrocarbons and endotoxin.

Air pollution causes up to 30% of increase in COPD symptoms and case control studies have suggested that ozone has one of the strongest correlations [[13]][[14]]. Each increase of 50 µg/m³ in air pollutants results in a 3–20% increase in the number of emergency room admissions due to chronic obstructive pulmonary disease [[11]].

Endogenous Sources of Oxidants in Lungs

Although the lung is unique in its exposure to high concentrations of environmental oxidants, a large fraction of oxidants found in the lung are the by‐products of normal metabolism. Mitochondria are the largest source of free radicals because electrons leak from the electron transport chain onto oxygen to form superoxide. It is estimated that 1–3% of O₂ reduced in mitochondria may form superoxide [[9]]. A significant cytosolic source of superoxide is the enzyme xanthine dehydrogenase (a.k.a. xanthine oxidase) [[15]]. Membrane oxidases that produce substantial quantities of superoxide include cytochrome P450 the b₅ families of enzymes in the endoplasmic reticulum [[16]] and the NADPH oxidases. Degenerate cloning of membrane oxidases has recently revealed a new family of membrane oxidases (Mox) that are found on advential smooth muscle cells of coronary arteries [[17]] and aorta [[18]]. Hydrogen peroxide is formed during the dismutation of superoxide, but also by glycolate oxidase in peroxisomes. Peroxisomes also contain hydrogen peroxide producing enzymes such as D‐amino acid oxidase, urate oxidase, L‐alpha‐hyroxyacid oxidase, and fatty acyl‐CoA oxidase [[19]]. Soluble enzymes such as xanthine oxidase, aldehyde oxidase, dihydrorotate dehydrogenase, flavoprotein dehydrogenase and tryptophan dioxygenase can generate ROS [[19]]. Hydroxyl radicals classically form in the presence of metals and hydrogen peroxide (Fenton reaction); however, decomposition from other molecules such as peroxynitrite may play a small role in hydroxyl radical formation [[20]].

Reactive nitrogen species are primarily derived from nitric oxide. The sources of NO in the lung are three NO synthases (NOS): 1) neuronal NOS, found in the nerve plexus of the trachea [[21]] (NOS1), 2) inducible NOS, found in respiratory epithelium and activated macrophages [[22]] (NOS2), and 3) endothelial NOS, found in respiratory epithelium and vascular endothelium (NOS3) [[23]]. In the resting state nitric oxide is considered a signaling molecule. When nitric oxide is produced in high concentrations, such as with inducible NOS, it can react with oxygen or superoxide to form the highly reactive compounds nitrogen dioxide and peroxynitrite. Only inducible NOS is highly upregulated by such cytokines as tumor necrosis factor–α (TNF‐α) and interleukin‐1ß [[24]]. Thus, location determines the primary functions of nitric oxide in the lung: in pulmonary vessels nitric oxide is a vasodilator; in airway muscle nitric oxide functions as a bronchodilator; and in airway epithelium nitric oxide modulates the immune response.

Antioxidant Defenses in Lungs

Antioxidants can be classified as either enzymatic or non‐enzymatic and are the primary defense against ROS/RNS. Antioxidant enzymes include the families of superoxide dismutase (SOD), catalase, glutathione peroxidase, glutathione S‐transferase, and thioredoxin. Each family has isoenzymes that are distinguished primarily by their distribution. For instance, the three mammalian superoxide dismutases are cytosolic (SOD1) [[25]], mitochondrial (SOD2) [[26]], and extracellular (SOD3) [[27]]. There are also multiple thioredoxin (Trx) and thioredoxin‐like (Txl‐1) proteins in the lung [[28]][[29]][[30]].

The non‐enzymatic category of antioxidant defenses includes low molecular weight compounds such as glutathione, ascorbate, urate, alpha‐tocopherol, bilirubin, and lipoic acid. Concentrations of these antioxidants vary depending of both subcellular and anatomic location (Table 1). For instance, glutathione is one hundred fold more concentrated in airway epithelial lining fluid compared to plasma [[31]]. Other high molecular weight molecules that might be considered antioxidants include proteins that have oxidizable thiol groups such as albumin or proteins that bind free metals such as transferrin. Albumin and transferrin are found in high concentration in serum, but are in much lower concentration in airway lining fluid [[32]]. Thus, both the lung parenchyma and airways have several distinct antioxidant enzyme systems.

Table 1.  Low Molecular Weight Antioxidants in Normal Subjects

Antioxidant	Plasma	Epithelial lining fluid
Glutathione	1.0 ± 0.7 µM	109 ± 64 µM
Urate	387 ± 132 µM	207 ± 167 µM
Ascorbate (Vitamin C)	67 ± 25 µM	40 ± 18 µM
α‐tocopherol (Vitamin E)	16 ± 5 µM	0.7 ± 0.3 µM

*Adapted from Ref. [[31]].

Antioxidant Enzyme Families

SOD Family

One of the major enzymatic antioxidant families is the superoxide dismutase family, which catalyzes the dismutation of superoxide radicals into hydrogen peroxide and oxygen. The reaction is pseudo first‐order and almost diffusion limited (Michaelis‐Menten constant > 10⁹ M^{− 1}s^{− 1}) [[33]]. Three SOD isoenzymes have been identified in mammals. The major intracellular SOD is a 32 kDa copper and zinc (Cu/Zn SOD or SOD1) containing homodimer present throughout the cytoplasm and nucleus [[34]]. Cu/Zn mutants are important in the pathogenesis of amyotrophic lateral sclerosis [[35]]. The mitochondrial SOD (Mn‐SOD or SOD2) is a manganese‐containing 93‐kDa homotetramer that is synthesized in the cytoplasm and translocated to the inner matrix of mitochondria [[26]][[36]]. Lack of Mn‐SOD may cause cardiomyopathies and neurodegenerative diseases [[37]][[38]][[39]]. The last mammalian SOD to be discovered is primarily extracellular (EC‐SOD or SOD3) [[27]].

CuZn SOD

CuZnSOD is the most abundant superoxide dismutase in most tissues. In the lung, CuZnSOD is highly expressed in bronchial epithelium, alveolar epithelium, mesenchymal cells, fibroblasts, and arterioles in capillary and endothelial cells [[40]][[41]]. In the bronchial epithelium, CuZnSOD is highly expressed in ciliated epithelial cells. CuZnSOD is located primarily in the cytosol but is also present at substantial concentration in the nucleus. It is highly expressed in some organelles such as lysosomes [[40]]. Studies with mice overexpressing transgenic CuZnSOD show that these mice are resistant to allergen induced lung toxicity and pulmonary oxygen toxicity [[42]][[43]].

Manganese Superoxide Dismutase

MnSOD mRNA is prominently expressed in cells in airway walls, the septal tips of alveolar ducts and in arteriolar walls located adjacent to airways [[44]][[45]]. In the lung, MnSOD is exclusively localized to mitochondria [[40]]. Mice that do not express MnSOD die shortly after birth [[37]][[38]]; however, the lungs are only minimally affected. The organs most affected by MnSOD genetic deficiency are the central nervous system, liver, and cardiovascular system [[37]][[38]]. Mice that are heterozygous for the MnSOD deletion do not reveal sensitization to hyperoxia or pertussis toxin induced lung injury [[46]][[47]]; however, decreased lung glutathione concentrations in these animals suggest that there is increased oxidative damage in the airways [[48]]. Mice that overexpress transgenic MnSOD in the lungs are resistant to hyperoxia [[49]] yet are not protected from excessive irradiation [[50]]. Thus the role of MnSOD in modulating lung injury needs further investigation.

EC‐SOD

EC‐SOD is the primary extracellular SOD enzyme and is highly expressed in many organs. The uterus, pancreas, and lung have the highest EC‐SOD activity levels among solid organs; EC‐SOD activity in the human lung is approximately eight times that of the liver, six times that of the brain, two to three times that in heart, and one to two times that of the kidney [[51]]. Within the lung, EC‐SOD protein is particularly abundant in blood vessels and airways (Fig. 4). EC‐SOD mRNA is abundant in airway epithelium and vascular endothelium and is expressed at levels four times of that in heart and 15 times that in liver [[52]]. The protein is mainly associated with the connective tissue matrix around vessels and airways in the lung, but is also found in close proximity to airway and vascular smooth muscle cells [[53]][[54]]. Electron microscopic immunocytochemistry reveals that EC‐SOD is seen in areas rich in type I collagen and associated with extracellular matrix around airway and vascular smooth muscle cells. The prominent expression of EC‐SOD around airways and airway smooth muscle suggests that it may play a role in allergic airway diseases such as asthma. Unstressed mice that are deficient in EC‐SOD have a normal phenotype but their blood vessels have impaired vasodilation to nitric oxide [[55]].

Figure 4. Pulmonary distribution of extracellular superoxide dismutase. Histochemical localization of the superoxide dismutases in human lung using rabbit anti‐human extracellular superoxide dismutase extracellular superoxide dismutase (ECSOD) antibodies. ECSOD labeling of a small airway (A) and small pulmonary vessel (B). Strong labeling is seen in the airway and vessel walls (arrows). (A) Reprinted by permission from Ref. [[206]] and (B) reprinted by permission from Ref. [[207]]. (Full color version available online.)

Although the role of EC‐SOD in airway diseases is not clear, two studies indirectly suggest that EC‐SOD activity might play a role in COPD. First, neutrophils secrete increased extracellular superoxide when stimulated with ultra‐fine carbon black particles [[56]]. Second, EC‐SOD mimetics reduce airway inflammation and epithelial cell hyperplasia in animal models [[57]]. Although the role of EC‐SOD has not been specifically investigated in COPD, EC‐SOD attenuates lung injury and inflammation in other pulmonary diseases. For instance, mice that overexpress EC‐SOD in the lung are partially protected from hyperoxia‐induced injury [[58]][[59]], hemorrhagic shock [[60]], influenza [[61]], bleomycin [[62]], and oil fly ash [[63]]. These studies suggest that under normal conditions the constitutive levels of EC‐SOD may be sufficient to attenuate the effects of superoxide in extracellular spaces, but not during pathologic states.

Another mechanism by which EC‐SOD might influence airway physiology is through modulation of nitric oxide. Nitric oxide is a potent signaling molecule that may cause non‐adrenergic mediated bronchodilation [[64]][[65]][[66]]. Superoxide inactivates nitric oxide by rapidly reacting with it to form peroxynitrite. A recent report implicates EC‐SOD as a modulator of nitric oxide mediated signaling in blood pressure control [[55]]. The relative abundance of EC‐SOD in both bronchi and blood vessels raises the possibility the EC‐SOD might also modulate pulmonary blood flow or neurally mediated bronchodilation; however, the significance of EC‐SOD in modulating nitric oxide signaling pathways in airways remains to be shown.

Thioredoxin System

The thioredoxin system consists of multiple species including thioredoxins (e.g. Trx‐1 and Trx‐2), thioredoxin like proteins (e.g. Tlx‐1, Tlx‐2, Sptrx‐1, and Sptrx‐2) and thioredoxin reductases (e.g. TrxR‐1 and TrxR‐2) [[29]][[67]][[68]]. These enzymes are important in protein disulfide reduction, but may have additional antioxidant functions. Thioredoxin is a 13‐kDa protein containing two redox‐active cysteines at its active site (Trp‐Cys‐Gly‐Pro‐Cys), which are highly conserved across species. Mammalian TrxR are selenoezymes that reduce oxidized thioredoxin and other protein disulfides. Thioredoxin reduces the viscosity of sputum in Cystic Fibrosis patients [[69]], but its role in patients with COPD is unknown. In humans there are multiple thioredoxin genes including Trx‐1, which encode for the classical thioredoxin and Trx‐2, an isoenzyme that is located in the mitochondria. These proteins have been shown to be present in bronchial epithelium and alveolar macrophages [[28]][[29]]. Spermatid‐specific like thioredoxins are not expressed in the lung [[70]]. There have been a few studies of thioredoxin in non‐malignant lung disease. These show increased staining for both Trx and TrxR in interstitial lung diseases, such as usual interstitial pneumonia (UIP) and desquamative interstitial pneumonia (DIP), particularly in areas of metaplastic epithelium and in inflammatory cells [[28]]. Exogenous thioredoxin has been shown to reduce ischemia reperfusion in the lung [[71]][[72]][[73]]. Transgenic mice that overexpress human thioredoxin were more resistant to influenza virus infection than control mice [[74]].

Glutathione System

Glutathione is the most abundant intracellular thiol‐based antioxidant. It has an important function in maintaining the redox status within cells in addition to detoxifying compounds via conjugation reactions through glutathione S transferases. The enzymes that are important in the metabolism of glutathione include glutathione peroxidases (GPx) and the glutathione regenerating enzyme, glutathione reductase (GR). The glutathione system has an important role intracellularly, however, glutathione is in very high concentrations in lung epithelial lining fluid as discussed above. Patients with COPD exacerbations have lower levels of glutathione peroxidase activity in red blood cells [[75]].

The mechanisms by which glutathione (GSH) could attenuate COPD exacerbations are unknown, but should be interpreted within the context of its distribution in the lung. For instance, bronchoalveolar lavage fluid (BALF) contains a 100‐fold concentration of glutathione compared to blood [[31]][[76]]. GSH is also highly concentrated in intracellular spaces [[9]]. Although GSH has been shown to be elevated in the BALF of chronic smokers [[77]], cigarette smoke acutely lowers intracellular glutathione [[78]][[79]].

Evidence of Oxidative Stress in COPD

There is now overwhelming evidence implicating an association between oxidative stress and COPD; however, most of this evidence is indirect. The two indirect methods that are most often used to detect ROS are spin trapping and footprinting (also called fingerprinting). Spin trapping is a technique by which a radical reacts with a more stable molecule that can be readily measured in biologic systems. Spin trapping has been used to demonstrate increased ROS in the bronchoalveolar lavage fluid of patients with COPD [[80]]. Footprinting involves the principle that ROS react with lipids, proteins, or nucleic acids to form specific markers of oxidative stress. Examples of footprint reactions include ROS attack on proteins to form carbonyls, peroxynitrite reacting with tyrosine to form nitrotyrosine, ROS reacting with lipids to liberate ethane and isoprostanes or DNA to form base pair adducts such as 8‐oxo‐2‐deoxyguanosine. Another commonly measured footprint is the oxidation of glutathione to oxidized glutathione. Footprinting has been used extensively to study local and systemic oxidative stress in COPD (Table 2).

Table 2.  Free Radicals and Their Footprints in COPD

Marker	Change
Carbon monoxide	↑ [[197]]
8‐isoprostanes	↑ [[84]][[86]]
Ethane	↑ [[85]]
Nitrotyrosine	↑ [[198]]
α1‐PI activity	↓ [[134]]
Hydrogen peroxide	↑ [[90]][[97]]
Nitric oxide	↑ [[99]]

*See text for details.

Many studies have demonstrated increased footprints of oxidative stress both locally (lung) and systemically (blood and urine) in COPD patients. Evidence for local oxidative stress has been found in exhaled breath condensates, sputum, and lavage fluid of both smokers and patients with COPD. Some, but not all of these studies have found a difference between smokers and patients with COPD. For instance, lipid peroxidation products such as aldehydes are increased in the exhaled breath condensates of smokers and patients with COPD, but only malondialdehyde is increased in patients with COPD compared to smokers without COPD [[81]]. The presence of systemic oxidative stress in COPD is suggested by the elevation of plasma lipid peroxidation products such as malondialdehyde during COPD exacerbations [[75]][[82]][[83]]. Urinary levels of 8‐isoprostanes are also increased in patients with COPD [[84]]. Exhaled ethane is increased in COPD patients [[85]]. Similarly, patients with COPD have increased exhaled breath condensate 8‐isoprostanes compared to healthy smokers and non‐smokers [[86]]. The exhaled breath condensate pH has also been shown to correlate with 8‐isoprostanes and hydrogen peroxide in patients with COPD [[87]] and the 8‐isoprostane levels increase with COPD exacerbations, but return to normal after 2–4 weeks [[88]]. Lung tissue of patients with COPD has been shown to have increased lipid peroxidation product 4‐Hydroxy‐2‐nonenal [[89]]. Long term N‐acetylcysteine does not appear to affect exhaled breath condensate lipid‐peroxidation as measured by TBARS [[90]].

The EBC concentration of nitrates in COPD patients is no different from controls [[91]]; however, the sputum of patients with COPD has increased nitrates and nitrites (NOx) [[92]]. Nitrotyrosine levels are increased in the sputum of patients with COPD compared to healthy subjects and those with asthma [[93]]. Four weeks of inhaled steroids reduce nitrotyrosine staining in sputum [[94]]. Since exhaled nitric oxide is not elevated in patients with stable COPD [[95]][[96]], further work is necessary to determine the specific role of NOx in COPD. It is likely that gaseous NO is consumed by formation of peroxynitrite in the presence of oxidative stress.

There is direct evidence that stable ROS are elevated in patients with COPD. For instance, patients with COPD have increased exhaled hydrogen peroxide [[90]][[97]][[98]] and nitric oxide [[66]][[85]][[99]][[100]]. Cigarette smoking has also been associated with increased superoxide production in airway leukocytes in both acute and chronic smokers [[101]]. A causal link between air pollution and COPD exacerbations is suggested by the finding that neutrophils from patients with COPD produce more superoxide when exposed to ultrafine carbon black [[56]]. Direct evidence implicating other ROS has been difficult to obtain because the evanescent nature of ROS make them difficult to measure in vivo. For instance, the hydroxyl radical reacts with molecules in its immediate vicinity, thus making it difficult to measure in vivo. Electron spin resonance is the only method that directly measures free radicals, but this method is not sensitive enough to detect free radicals in most in vivo conditions.

The Role of Oxidative Stress in the Pathogenesis of COPD

Gene Transcription

Transcription of genes requires multiple steps beginning at the level of chromatin. The basic unit of chromatin is the nucleosome, which consists of approximately 146 base pairs of DNA wrapped around multimeric subunits of nucleosomes. Genomic regions that are tightly wrapped around histones have suppressed gene transcription since the transcriptional complex does not have access to the DNA. One of the key events in unraveling the DNA from the histone complex is the acetylation of lysine residues on histone protein number 4 [[102]]. The amount of acetylation is controlled by histone acetyl transferases (HAT), which promote histone acetylation, and histone deacetylases (HDAC), which promote histone deacetylation. Although the specific role of oxidative stress in modifying HAT and HDAC activity is unknown, transcription factors such as NF‐κB AP‐1, cyclic APM response element binding protein (CREB) and other transcription factors are sensitive to the redox state and modulate HAT activity ([[19]][[103]]). Oxidative stress activates HAT activity in an epithelial line and leads to inflammatory gene transcription [[104]]. There is evidence that HDAC2 activity is reduced by cigarette smoking and oxidative stress [[105]] and that this may be due to nitration of HDAC2 by peroxynitrite formation [[106]]. This reduction in HDAC2 activity appears to account for corticosteroid resistance in COPD patients [[107]]. These pathways are also activated in the transcript of other pro‐inflammatory genes such as interleukin 1 and tumor necrosis factor α. The mechanism by which ROS modulate this activity is unknown, but the most likely mechanism is through the oxidation and reduction of disulfide bridges or thiol‐residues. For instance, when Rac‐GTPases are activated, they produce reactive oxygen species that down regulate Rho activity through oxidation of a key thiol residue on a low molecular weight tyrosine phosphatase that turns off Rho [[108]].

The mechanisms by which tobacco smoke modulates transcription of genes in patients with chronic obstructive pulmonary disease are unknown. Signaling pathways that have been shown to be sensitive to tobacco smoke include protein kinase C (PKC) whose activity increases 2–3 fold when stimulated with 5% cigarette smoke extract [[109]]. Cigarette smoke also modulates NF‐κB [[110]] and leads to a prolonged increase in c‐fos transcripts [[111]]. NF‐κB is increased in alveolar macrophages and airway epithelial cells of COPD patients [[112]], particularly during exacerbations [[113]]. A recent gene expression profiling study of rat respiratory tract epithelium which extensively characterized the effects of cigarette smoke on gene expression [[114]] revealed that cigarette smoke causes a rapid induction of oxidative stress response genes and drug metabolizing enzymes such as hemeoxygenase 1 and NADPH quinonone‐oxido‐reductase. However, after three weeks of exposure, the increased expression of these genes was markedly reduced.

Latent Adenoviral Infections

An intriguing hypothesis for the pathogenesis of COPD is that latent adenoviral infections contribute to the disease through the pro‐inflammatory effects of the adenoviral E1A gene [[115]][[116]]. The link between adenovirus infection and oxidative stress in the pathogenesis of COPD remains to be proven; however, adenovirus E1A has been shown to block the induction of ferritin, which sensitizes cells to cytotoxicity from hydrogen peroxide [[117]]. Additionally, human alveolar epithelial cells that have been treated with E1A show increased production of mRNA for the cytokine interleukin 8 after treatment with hydrogen peroxide [[118]].

Inflammation

Although asthma and COPD are both associated with increased airway inflammation, the type of inflammation is substantially different. The sputum of asthma patients contains a high percentage of eosinophils [[100]] whereas the sputum in COPD patients contains predominantly neutrophils [[119]]. Cigarette smokers have similar increases of neutrophils in their airways [[120]] and there is an inverse correlation between neutrophilic inflammation and FEV₁ [[121]]. One mechanism by which cigarette smoke might increase lung neutrophils is by decreasing neutrophil deformability. Since the average pulmonary capillary diameter is only 5 µM and the average neutrophil diameter is 7 µM, the average neutrophil must deform to pass through the lungs. Cigarette smoking has been shown to acutely increase neutrophil deformability in the lungs in vitro [[122]] and increase neutrophil lung‐transit time in vivo [[123]]. The loss of deformability can be reduced by the antioxidant glutathione [[122]]. Additional studies have shown that oxidants also play a role in recruiting neutrophils to the lung by causing the up regulation of neutrophil adhesion molecules [[124]]. Thus, oxidants in cigarette smoke may increase lung inflammation by trapping neutrophils in the lungs both through decreased deformability and through increased adhesion to epithelium.

The recruitment of inflammatory cells to the lung in COPD patients may set up a positive‐feedback loop that perpetuates lung injury through other mechanisms, including the regulation of cytokine secretion. For instance, patients with COPD demonstrate increased levels of interleukin (IL)‐6, IL‐1 ß, TNF‐α and IL‐8 in sputum [[125]] and respiratory epithelial cells of smokers and patients with COPD have been shown to secrete increased TGF‐ß [[126]]. The link between cigarette smoke and cytokine secretion may be activation of NF‐κB, which is up regulated by oxidants [[127]]. ROS, particularly hydrogen peroxide, have also been shown to regulate signal transduction in inflammatory cells via oxidation of key cysteine residues in phosphatases [[108]].

Intracellular Adhesion Molecules

Cigarette smoke extract induces increased expression of intercellular adhesion molecule (ICAM)‐1, which is tightly controlled by NF‐κB, on airway epithelial monolayers compared with unstimulated cells in vitro [[128]]. This enhanced ICAM‐1 expression is associated with a greater capacity of airway epithelial cells to bind mononuclear cells. ROS such as superoxide have recently been shown to be important in the induction of intracellular adhesion molecules such as E‐and P‐selectins [[129]][[130]][[131]]. Recent evidence also implicates endothelial nitric oxide synthase (eNOS) in the induction of these adhesion molecules [[132]]. In COPD patients the plasma levels of L‐selectin, E‐selectin and P‐selectin are not increased above controls, [[133]]. Thus the relevance of oxidative stress in upregulating adhesion molecules remains to be proven in COPD patients.

Protease Activity

Neutrophil elastase and other lung proteases are capable of causing emphysema. Reduced activity of the α1‐protease inhibitors is strongly associated with COPD and is the only well‐established genetic risk factor. There are several links between protease activity and oxidative stress. First, ROS oxidize a key methionine in the α1‐antitrypsin leading to its inactivation [[134]]. The oxidation of plasma proteins is almost exclusively due to the oxidant and pro‐oxidant species found in the tar phase of cigarette smoke [[135]]. Second, induction of human neutrophil elastase is redox sensitive [[136]]. Third, activation of lung epithelial cell secreted proteases leads to increased release of ROS [[137]]. Additional investigations are needed to fully determine how ROS modulate protease activity and lead to degradation of lung tissue.

Mucus Secretion

Oxidants from cigarette smoke or inflammatory cells may be associated with chronic bronchitis and COPD exacerbations by impairing mucociliary clearance. Reactive oxygen species have been reported to stimulate mucin secretion [[138]]. The superoxide‐mediated increase in mucus from respiratory epithelial cells is mediated by prostaglandins and can be decreased by inhibitors of arachidonic acid metabolism [[139]]. Superoxide and hydrogen peroxide have also been shown to decrease beating of respiratory epithelium cilia, even as early as 15 minutes after exposure [[140]]. A large, randomized trial is underway to determine whether the antioxidant N‐acetylcysteine can reduce exacerbations of chronic bronchitis [[141]].

Apoptosis

Recently reports have suggested that apoptosis may also play a role in the pathogenesis of emphysema. In patients with COPD, increased numbers of apoptotic cells have been demonstrated in pulmonary endothelium [[142]] and in alveolar walls [[143]][[144]]. Airway lymphocytes [[145]] and stimulated peripheral blood lymphocytes [[146]] from patients with COPD have also shown increased apoptosis. The pathogenesis of increased apoptosis of lung tissue remains under active investigation, but is hypothesized to be due to decreased production of growth factors such as vascular endothelial growth factor (VEGF) [[147]][[148]]. This hypothesis is supported by the finding that emphysematous lungs have decreased VEGF ligand [[142]] and the observation that inhibition of the VEGF receptor will cause emphysema in rats [[149]]. A role for oxidative stress is suggested because a SOD mimetic is capable of not only blocking the oxidative stress associated with VEGF blockade, but also is able to prevent the resultant emphysema (Fig. 5) [[150]].

Figure 5. Treatment with an SOD mimetic (M40419) attenuates anti‐VEGF‐(SU5416) mediated emphysema. (a) SU5416‐treated lungs show increased size of alveolar structures. (b) SU5416 + M40419‐treatment protects against alveolar loss and these animals show normal alveolar size and architecture (bar = 100 µm). Adapted with permission from Ref. [[150]]. (Full color version available online.)

Oxidative Stress and Nitric Oxide Signaling in the Lung

Although there is strong evidence of a role for nitric oxide in asthma, data regarding the benefit of nitric oxide in COPD is less certain. Increases in exhaled nitric oxide correlate with sputum neutrophilia [[151]] but the effect of COPD on exhaled NO is not yet clear. Some investigators have found that exhaled nitric oxide is higher in COPD patients during exacerbations compared to healthy controls [[95]][[152]], but there is no difference when COPD patients are at baseline [[153]]. Other investigators have found no increases in exhaled nitric oxide in patients with COPD [[154]].

The role of nitric oxide in the lung is complicated because there are thee distinct sources of NOS, as previously discussed. Both nNOS and eNOS are constitutively active; however, it is primarily iNOS that is induced in asthma and in COPD exacerbations and is responsible for increased levels of exhaled nitric oxide during these diseases. Although inhaled nitric oxide reduces pulmonary vascular resistance in patients with COPD [[155]] and decreases airway resistance in asthmatics, there is no improvement in oxygen saturation or airway resistance in patients with COPD [[66]][[156]][[157]]. Thus a lack of significant bronchorelaxation associated with nitric oxide in COPD suggests that signaling pathways in bronchial smooth muscle are impaired. Concomitant increases in the production of reactive oxygen species during inflammation may be the source of this impairment. Additionally increases in exhaled NO may correlate with increases in tissue peroxynitrite rather than with increases in tissue NO if there is oxidative stress in the tissue with enhanced superoxide production.

The impairment of airway nitric oxide signaling in COPD patients is likely due to nitric oxide's reaction with reactive oxygen species. For instance, nitric oxide rapidly reacts with superoxide to form peroxynitrite [[158]]. Peroxynitrite formation increases during inflammation and is toxic to microbes; however, peroxynitrite can also cause airway hyperresponsiveness [[159]]. Presumably, peroxynitrite formation is also associated with a decrease in the amount of nitric oxide available for bronchodilation of smooth muscle.

One would predict peroxynitrite formation to be favored at sites of inflammation and infection, but one would also predict that in the normal lung peroxynitrite production would be inhibited at sites where nitric oxide mediated signaling is important. Two sites at which nitric oxide signaling is essential are respiratory smooth muscle and the vasculature. Indeed, EC‐SOD is abundant in both airway smooth muscle and pulmonary vasculature and may be a mechanism of preserving both vasoregulation and smooth muscle bronchodilation in response to nitric oxide. The specific role of EC‐SOD in COPD remains to be proven, but should be given consideration for a substantial role in both inhibiting airway inflammation and in preserving the bronchorelaxant effects of nitric oxide during oxidative stress.

Respiratory Muscle Dysfunction

There is significant skeletal muscle dysfunction in some patients with severe COPD [[160]]. This dysfunction is especially prominent in the diaphragm [[161]], but also in peripheral muscles [[162]]. Several studies have demonstrated that COPD patients have evidence of increased oxidative stress, both peripherally and locally, particularly after exercise [[10]][[162]][[163]]. Biopsies of skeletal muscles from patients with COPD have shown increased apoptosis [[164]]. The contribution of oxidative stress to apoptosis in skeletal muscles is unknown. Other possible mechanisms by which oxidative stress may lead to decreased muscle function are unclear, but may involve alterations in electrolyte movement, mitochondrial respiration, or contractile proteins.

The relationship between skeletal muscle strength and oxidative stress is complicated. Oxidative stress has also been shown to increase skeletal muscle transcript levels for the antioxidant enzymes glutathione peroxidase, catalase, CuZnSOD and MnSOD [[165]]. In healthy individuals, a training protocol leads to skeletal muscle adaptation to oxidative stress, presumably through the induction of these enzymes; however, patients with COPD have reduced capacity to adapt to oxidative stress with exercise training [[166]].

Glucocorticoid Resistance

The chronic inflammation of COPD is resistant to the anti‐inflammatory effects of corticosteroids and this may be linked to oxidative stress. Steroid resistance occurs at the level of specific cells including alveolar macrophages, which show reduced responsiveness to corticosteroids in normal smokers and complete resistance in COPD patients [[167]]. By contrast, patients with asthma are highly sensitive to the effects of low doses of corticosteroids. Some of the mechanisms by which corticosteroids switch off inflammatory genes in asthma have recently been elucidated [[168]]. A major mechanism for switching off inflammatory genes is the reversal of histone acetylation activated by transcription factors by histone deacetylases (Fig. 6). Corticosteroids induce the translocation of glucocorticoid receptors from the cytoplasm to the nucleus and these bind to activate inflammatory gene transcriptional complexes. This leads to specific recruitment of HDAC2, which reverses histone acetylation and switches off the activated inflammatory genes [[169]]. HDAC2 activity is markedly reduced in alveolar macrophages obtained from smokers and even further reduced in cells from COPD patients [[105]]. The reduction in HDAC activity is significantly related to the reduced responsiveness to corticosteroids in these cells. The mechanism for reduction in HDAC activity is likely oxidative stress. Both oxidative stress (hydrogen peroxide) and cigarette smoke extract reduce HDAC activity in vitro. The mechanism may involve formation of peroxynitrite, since peroxynitrite mimics these effects and nitrates HDAC2, presumably impairing its catalytic activity and then its destruction [[106]]. The same mechanisms may account for the resistance to corticosteroids seen in asthmatic patients who smoke [[107]].

Figure 6. Stimulation of normal alveolar macrophages activates nuclear factor‐κB (NF‐κB) and other transcription factors to switch on histone acetyltransferase leading to histone acetylation and transcription of genes encoding inflammatory proteins, such as tumor necrosis factor‐α (TNF‐α), interleukin‐8 (IL‐8) and matrix metalloproteinase‐9 (MMP‐9). Corticosteroids reverse this by binding to glucocorticoid receptors (GR) and recruiting histone deacetylase‐2 (HDAC2). This reverses the histone acetylation induced by NF‐κB and switches off the activated inflammatory genes. In COPD patients cigarette smoke activates macrophages, as in normal subjects, but oxidative stress (perhaps acting through the formation of peroxynitrite) impairs the activity of HDAC2. This both amplifies the inflammatory response to NF‐κB activation, and reduces the anti‐inflammatory effect of corticosteroids as HDAC2 is now unable to reverse histone acetylation. (Full color version available online.)

The Role of Antioxidants in COPD

Endogenous and Dietary Antioxidants

There is a well‐established epidemiologic relationship between dietary intake of antioxidants and lung function [[170]]. Diets rich in fresh fruit and vegetables that are rich in antioxidants are associated with improved lung function and have been proposed as treatment strategies for patients with COPD [[171]][[172]]. It is unclear which specific dietary antioxidants are responsible for the decrease in lung function, but epidemiologic studies have indirectly suggested a role for vitamin E [[173]][[174]], vitamin A [[175]], and Vitamin C (Fig. 7) [[176]]; however, not all studies have reported a correlation between these dietary antioxidants and lung function [[177]]. Direct evidence implicating specific dietary antioxidants is more complicated.

Figure 7. Low levels of plasma ascorbate increase the risk of COPD independently of smoking history. The relative risks of having COPD in light (white bar), medium (black bar) and heavy (cross hatched bar) smokers compared to never smokers are shown for each quintile of plasma ascorbate concentration. The risk of COPD was highest in the patients with the very low quintile of plasma concentration (3–37 µM), compared to low (28–49 µM), median (50–58 µM), high (59–68 µM), or very high quintiles (69–174 µM). Adapted from Ref. (176).

Direct measurement of antioxidants in both airways and plasma has suggested an association between COPD and low antioxidant levels. Alveolar macrophages obtained from bronchoalveolar lavage of COPD patients show decreased levels of glutathione [[178]]; however, chronic supplementation with N‐acetylcysteine does not increase glutathione in bronchoalveolar lavage fluid or lung tissue in patients with COPD [[179]]. Smoking is associated with increased concentrations of airway total antioxidant activity and specific antioxidants such as glutathione, particularly in chronic smokers [[101]]. Airway glutathione concentrations are correlated with airflow obstruction in patients with COPD [[180]]. In plasma, both smokers [[181]] and patients with COPD exacerbations [[82]] have decreased plasma total airway antioxidants. Some studies have failed to find a correlation between plasma total antioxidant activity and % predicted FEV₁ [[182]], thus additional investigation is necessary to elucidate the role of dietary antioxidants in COPD.

Enzymatic antioxidants may also play a role in COPD, however there are fewer investigations of these. Homozygosity of microsomal epoxide hydrolase, an enzyme that detoxifies epoxides, has been associated with COPD [[183]]. Other studies have revealed a slightly increased risk of COPD with deletions in the glutathione‐S‐transferase allele [[184]]. In smokers there is a GT dinucleotide repeat length polymorphism in the 5′‐flanking region of human hemoxygenase‐1 gene that is of increased prevalence in patients with emphysema [[185]]. Decreased red blood cell catalase activity has been associated with COPD [[186]]. Red blood cell glutathione peroxidase has been shown to increase during COPD exacerbations [[75]]. To our knowledge, there are no reports of lung superoxide dismutase activity, thioredoxins, or other peroxidases in COPD.

Pharmacologic Antioxidants

N‐Acetylcysteine

N‐acetylcysteine (NAC) is hydrolyzed intracellularly to cysteine and stimulates glutathione synthesis. NAC reduces lung injury in many animal models of disease [[187]][[188]][[189]][[190]][[191]][[192]][[193]][[194]][[195]] and is used to treat acetaminophen overdosage in humans. Several randominized‐controlled studies have shown that 600 mg NAC given daily will reduce exacerbations of COPD (Table 3). The mechanism by which NAC protects the lung are unknown, but are presumed to be through upregulation of the glutathione system.

Table 3.  Randomized Controlled Trials Using N‐Acetylcysteine to Treat COPD Patients

Study	N	Odds ratio [95% CI] of at least 1 exacerbation
Boman, et al. [[199]]	203	0.36 [0.19–0.67]
British Thoracic Society [[200]]	148	0.65 [0.25–1.73]
Grassi, et al. [[201]]	69	0.45 [0.17–1.20]
Hansen, et al. [[202]]	129	0.60 [0.30–1.22]
Multicenter Study Group [[203]]	495	0.28 [0.19–0.42]
Pela, et al. [[204]]	163	0.39 [0.20–0.76]
Rasmussen, et al. [[205]]	91	0.59 [0.20–1.73]

Antioxidant Mimetics

Several classes of antioxidant mimetic have been created that share the properties of being relatively low in molecular weight and having high rate constants for enzymatic scavenging of free radical species (Fig. 8). A metal atom (commonly manganese) acts as the active center for electron transfer to or from free radical species. Mimetics of the macrocyclic class illustrated in Fig. 8 are restricted to one electron transfer and are therefore more specific for scavenging O₂⁻ while the salen and porphyrin classes carry out both one and two electron transfers and are potentially effective in scavenging a broader range of reactive oxygen species. One of the macrocyclic classes of antioxidant mimetics has been shown to prevent the formation of emphysema in response to VEGF‐receptor blockade (Fig. 5) [[150]].

Figure 8. Three classes of superoxide dismutase mimetics. Side chains (R‐) can be substituted.

A manganoporphyrin antioxidant mimetic has been shown to significantly reduce inflammation and squamous metaplasia induced by tobacco smoke (Fig. 9) [[57]]. In this study animals were treated with the mimetic AEOL 10113 by intratracheal aerosolization once per week and then exposed to high doses of tobacco smoke (80 mg particulate/m³) for 6 hours per day for three successive days and then repeated for 8 weekly cycles. At the end of 8 weeks, animals exposed to tobacco smoke alone had 12% of their trachea showing squamous metaplasia while those receiving tobacco smoke exposure plus the antioxidant mimetic had only 2% airway squamous metaplasia. Both the number of inflammatory cells and mediators of inflammation expressed in bronchoalveolar lavage fluids were decreased by treatment with the mimetic [[57]].

Figure 9. Effects of tobacco smoke on airways. Catalytic antioxidants attenuate squamous metaplasia from tobacco smoke. (A) Epithelial cell changes in airways from tobacco smoke‐exposed and filtered air (control) rats (epithelium is indicated between arrows). Lung sections were stained with hematoxylin and eosin for general morphology. Airway epithelium from filtered air control rats do not show the striking squamous cell metaplasia with prominent filamentous‐like cytoplasmic extensions (*) observed in the airways from rats exposed to tobacco smoke for 8 weeks. (B) Squamous metaplasia in rat lungs with or without a superoxide dismutase mimetic. The percent of airway surface expressing the type of squamous metaplasia illustrated in Fig. 9A is shown. Data are presented as mean ± SE (n = 5−6). *p < 0.05 for comparison to the tobacco smoke alone group. Adapted from Ref. [[57]]. (Full color version available online.)

Other strategies for reducing oxidant stress such as the use of spin traps as free radical scavengers [e.g. nitrones [[196]]] or iNOS inhibitors to prevent the formation of peroxynitrite could also be effective in preventing airway and lung inflammation. These alternative approaches have not yet been tested in rigorous models of emphysema.

Conclusions

Current efforts to ameliorate the symptoms of COPD focus on the use of bronchodilators to counteract the airway narrowing induced by cigarette smoking. There are no known medications that have been shown to substantially alter the natural history of COPD. Although there are substantial data to implicate oxidative stress in the pathogenesis of COPD, there have been few studies that have successfully used antioxidant strategies to modulate the progression of COPD. The need for strategies to supplement endogenous antioxidants or modulate oxidative stress is great. These approaches may represent one of the most promising avenues of research for disease modifying drugs for a disease that exacts a significant toll on quality of life and productivity.

Acknowledgments

This work was supported by National Institutes of Health HL‐04407 (R.P.B.), 5 P01 HL031992‐20 (J.D.C), and the Colorado Tobacco Research Program 3I‐005(R.P.B.).

Dr. Crapo is a founder and CEO of Aeolus Pharmaceuticals, which is developing antioxidant mimetics for human clinical trials.