Introduction

Abstract Introduction Materials and Methods Results Discussion References

Cigarette smoke is a complex mixture of more than 3,800 compounds, including both free radicals in high concentrations and chemical compounds that readily react to form other reactive substances (1). Mainstream and sidestream gas-phase cigarette smoke each contain about the same concentration of about 1 × 10¹⁶ free radicals per cigarette (or 5 × 10¹⁴ free radicals per puff) (3). These free radicals are reactive yet remarkably long-lived in gas-phase smoke, lasting for more than 5 min. They may randomly attack various cellular constituents and thereby impair some biochemical functions of the lung tissues of cigarette smokers. The free radicals cause peroxidation of membrane lipids, accumulation of oxidized dysfunctional proteins, and increased DNA damage. A large body of evidence suggests that the longitudinal decline in pulmonary function (4, 5) and increase in major pulmonary diseases in smokers, including emphysema (6) and cancer (9, 10), are associated with cigarette smoking.

Mitochondria are the intracellular organelles responsible for adenosine triphosphate (ATP) synthesis through the coupling of oxidative phosphorylation (OXPHOS) to mitochondrial respiration in human and animal cells. They contain their own genetic material in the form of mitochondrial DNA (mtDNA), which is the only extrachromosomal DNA in human cells. Human mtDNA is a circular, double-stranded, 16,569-bp DNA molecule that encodes 13 polypeptides which constitute the respiratory enzyme complexes (11). Human mtDNA is intronless, does not bind to histones or other specific DNA-binding proteins, and replicates faster than nuclear DNA, without proofreading or efficient DNA repair systems (12). It is located near the inner membrane of mitochondria and is continually exposed to high levels of reactive oxygen species (ROS) and free radicals generated from either the electron-transport chain of mitochondria or from extrinsic sources. These peculiar characteristics of mtDNA render it susceptible to both intrinsic and extrinsic oxidative damage. It has been estimated that the rate of mtDNA mutation is 17 times higher than that of nuclear DNA (15). In the past few years, a number of pathogenic mutations of mtDNA, such as large-scale deletions and point mutations, have been established as responsible for or associated with several distinct human diseases (16). Moreover, significant progress has been made in the study of age- related somatic mtDNA mutations in human tissues. The accumulation of somatic mtDNA mutations could contribute to the progression of mitochondrial diseases (16), the occurrence of various types of degenerative diseases (17), and the age-related decline in biochemical function of the OXPHOS system (18). More than 80 large-scale deletions have been identified in various tissues of human subjects with specific mitochondrial diseases, degenerative diseases, and aging (19). These large-scale deletions often cause the removal or truncation of multiple structural genes and transfer RNA (tRNA) genes of the mitochondrial genome, which can result in multiple respiratory-chain deficiencies. Among these deletions, a 4,977-bp deletion, termed the "5-kb common deletion," is most frequently seen in various tissues of aged individuals (22).

The lung is directly exposed to the oxidative stress elicited by cigarette smoke, environmental and industrial pollutants, infections and inflammatory reactions, and medical interventions such as hyperbaric oxygen therapy and chemotherapy with agents such as bleomycin and doxorubicin. A ³²P-postlabeling study of covalent DNA damage in different tissues of cigarette smokers revealed that the most extensive DNA damage occurred in the lung (25, 26). The extent of covalent modification of mtDNA by polycyclic aromatic hydrocarbons was found to be 40 times greater than that of nuclear DNA (27). Oral administration of N-acetylcysteine, a cancer chemopreventive agent endowed with nucleophilic and antioxidant properties, significantly decreased the formation of mtDNA adducts in cigarette smoke-exposed lung and liver of the rat (28). The mechanism by which cigarette smoke causes single-strand breaks in the DNA of cultured lung cells was proposed to involve endonuclease activation and hydroxyl-radical attack on DNA (29). Single-strand breaks in mtDNA may lead to slip-replication, which is one of the possible mechanisms of large-scale deletion of mtDNA (30). We hypothesized that large-scale deletion of mtDNA is associated with cigarette smoking. In a previous study, the frequency of occurrence and proportion of the 4,977-bp mtDNA deletion in human lung tissues were found to be unrelated to smoking habit or lifetime cigarette consumption (31). To the best of our knowledge, there is no published information about specific large-scale mtDNA deletion and lipid peroxidation in the lung tissues of smokers. In the present study, we found a novel 4,839-bp deletion of mtDNA that accumulates with time in the lung tissues of smokers. The frequency of occurrence and the proportion of the 4,839-bp mtDNA deletion were found to be associated with the smoking habit and the amount of cumulative cigarette consumption. We also examined the relationship between this 4,839-bp mtDNA deletion and lipid peroxidation in human lung tissues and the smoking index.

	Materials and Methods

Abstract Introduction Materials and Methods Results Discussion References

Subjects

We collected lung tissues from one stillborn infant and 151 patients at the Division of Thoracic Surgery, Veterans General Hospital-Taipei, who underwent surgical resection for various pulmonary disease entities, including spontaneous pneumothorax, metastatic lung tumors, and bronchogenic carcinoma. Patients with infectious diseases such as pneumonia, fungal infection, and pulmonary tuberculosis; those with centrally located tumors with obstructing pneumonitis; and those who had received chemotherapy or radiotherapy for primary or metastatic lung malignancies were excluded from the study. None of the patients recruited into the study had a known history of industrial or occupational exposure to asbestos or organic solvent. A small piece of the lung tissue was collected from the grossly normal region of the resected specimen and quickly frozen in liquid nitrogen until analysis. Patient data, including sex, age, and history of major systemic diseases and surgery were collected prospectively. Smoking habit was classified into three categories: (1) current smokers, consisting of subjects who had smoked more than one cigarette or part thereof daily for more than 1 yr; (2) ex-smokers, consisting of subjects who had been previously reported as smokers and had quit smoking for more than 1 yr; and (3) nonsmokers, consisting of subjects who had never smoked. The smoking index is expressed as pack-year (cigarettes or parts thereof smoked per day × years of smoking/20). All patients enrolled in the study had refrained from smoking for at least 7 d before their surgery. Central venous blood samples were withdrawn on the day of surgery.

Evaluation of Pulmonary Function

Spirometry was performed a minimum of three times on 151 patients in the standard (seated) position, using a CPI 5000 IV (Gould, Houston, TX), before exploratory thoracotomy was done. The best values of FEV₁ and FVC were selected for analysis according to the criteria set by the American Thoracic Society (32). The results of the pulmonary function tests were reviewed by a senior chest physician (S.-C.C).

DNA Isolation

Because only a small amount (50 to 100 mg) of lung tissue was available from each of the patients, total DNA was isolated by proteinase K digestion/sodium dodecyl sulfate (SDS) lysis, followed by phenol/chloroform extraction, as previously described (33). After ethanol precipitation, the DNA pellet was dissolved in 100 µl of sterilized, doubly distilled water and frozen at 30°C until used.

Polymerase Chain Reaction Analysis of mtDNA Deletions

Oligonucleotide primers were chemically synthesized by Bio-Synthesis, Inc. (Lewisville, TX). The nucleotide sequences of the primers and the sizes of the polymerase chain reaction (PCR) products amplified with the indicated primer pairs from normal mtDNA and the mtDNA with a specific deletion, respectively, are shown in Table 1. PCR was conducted for 31 cycles in a 100-µl reaction mixture containing 200 ng template DNA, 200 µM of each deoxyribonucleoside triphosphate (dNTP), 40 pmol of each primer, 1.0 unit of Taq DNA polymerase (Perkin-Elmer/Cetus, Norwalk, CT), 50 mM KCl, 1.5 mM MgCl₂, 10 mM Tris-HCl (pH 8.3, at 25°C), 0.1% Triton X-100, and 0.01% (wt/vol) gelatin. The first cycle consisted of denaturation at 94°C for 5 min, annealing at 55°C for 5 min, and primer extension at 72°C for 3 min. The thermal profile of the following 30 cycles was denaturation at 94°C for 40 s, annealing at 56°C for 40 s, and extension at 72°C for 40 s. The amplified PCR products were separated on a 1.5% agarose gel at 150 V for 1 h, and DNA bands were visualized by transillumination under UV light after staining with ethidium bromide (50 µg/ml).

Primer-Shift PCR

A primer-shift PCR method (34) was used to ascertain the authenticity of the mtDNA deletions. By using the different primer sets listed in Table 1, we found that the length of the amplified DNA fragments changed in parallel with the change in the distance between each primer pair. Because no DNA fragment could be amplified from the wild-type mtDNA with the PCR thermal profile (short extension time) described previously, the only DNA fragment was that amplified from the mtDNA molecules with a specific (4,839-bp) deletion. The sizes of the PCR products amplified from this mtDNA with different primer pairs were 762 bp (L2/H8), 530 bp (L3/H4), 413 bp (L4/H7), and 281 bp (L5/H4), respectively (Figure 1). The sizes of the DNA fragments amplified from the mtDNA with a 4,977-bp deletion were 624 bp (L2/H8), 392 bp (L3/H4), and 275 bp (L4/H7), respectively. No PCR product was obtained by using the primer pair L5/H4, which indicates that the breakpoints of the putative (4,839-bp) mtDNA deletion are near L5 and H4, respectively. The mtDNA with the 7,436-bp deletion was amplified by using the primer pairs L2/H5, L4/H6, and L2/H6, as shown in Table 1.

View larger version (68K): [in this window] [in a new window]   Figure 1.   Electrophoretogram of PCR products amplified from mtDNA with the specific 4,839-bp deletion from human lung tissue. Lanes 2 to 5: 762-bp, 530-bp, 413-bp, and 281-bp PCR products amplified by primer pairs L2/H8, L3/H4, L4/H7, and L5/H4, respectively, from mtDNA with the specific 4,839-bp deletion in lung tissues of a 52-yr-old subject with a smoking index of 20 pack-yr. The 624-bp PCR product in lane 1 was amplified from mtDNA with the 4,977-bp deletion from this subject. Lane M: 100-bp DNA ladder size marker.

Semiquantitative PCR

The proportion of the mtDNA with the 4,839-bp deletion in each of the lung DNA samples was determined with a semiquantitative PCR method based on the same principle as that for the quantification of the mtDNA with the 4,977-bp deletion (23). The total DNA of lung tissues was serially diluted twofold with distilled water. The range of dilution was usually from 2³ and 2²⁰. The primer pair L1/ H1 was used for the amplification of a 610-bp DNA fragment from the total mtDNA, and the primer pair L2/H8 was used for the amplification of a 762-bp PCR product from the mtDNA molecules with the 4,839-bp deletion. Amplified DNA fragments were separated by electrophoresis on a 1.5% agarose gel at 150 V for 1 h, and were detected fluorographically after staining with ethidium bromide. The proportion of mtDNA with the 4,839-bp deletion was determined as the ratio of the highest-fold dilution that allowed the 762-bp PCR product to be visible on the stained gel to the dilution that allowed the 610-bp PCR product to be visibly amplified from the total mtDNA under identical conditions. A typical gel pattern is shown in Figure 2. This method was found to be reliable when the proportion of mtDNA with the 4,839-bp deletion was more than 0.005% of total mtDNA.

View larger version (27K): [in this window] [in a new window]   Figure 2.   Semiquantitative PCR analysis of mtDNA with the 4,839-bp deletion in human lung. (A) PCR amplification of total mtDNA by serial twofold dilution (lanes 1 to 8: dilutions of 213- to 220-fold, respectively) of total DNA from the lung tissues of a 52-yr-old subject using primers L1 and H1 under the conditions described in MATERIALS AND METHODS. (B) PCR amplification of mtDNA with the 4,839-bp deletion by serial twofold dilution (lanes 1 to 8: dilutions of 23- to 210-fold, respectively) of total DNA, using primers L2 and H8 under identical conditions. Lane M: 100-bp DNA ladder size marker.

DNA Sequencing

The 281-bp DNA fragment amplified from mtDNA with the 4,839-bp deletion through use of the L5/H4 primer pair was purified with the Nucleotrap CR extraction kit (Macherey-Nagel GmbH, Düren, Germany). Direct sequencing of the purified PCR product was done with the ABI PRISM dye terminator cycle-sequencing ready reaction kit with the AmpliTaq DNA polymerase and a fluorescence detection system, according to the recommended protocol (P/N 402078, revision A, August 1995; The Perkin-Elmer Corp., Branchburg, NJ). For each reaction, the following reagents were added: 8.0 µl Terminator Ready Reaction Mix, 3 to 6 µl (10 to 30 ng/µl) DNA template, 3.2 pmol of each primer, and distilled water to a final volume of 20 µl. Cycle sequencing was done on a GeneAmp PCR System Model 9600 (Perkin-Elmer/Cetus) for 26 cycles, with the following thermal profile: rapid thermal ramping to 96°C for 10 s, rapid ramping to 50°C for 5 s, and rapid thermal ramping to 60°C for 4 min. The PCR product was then purified by ethanol precipitation. The DNA pellet was suspended in 6 to 9 µl of loading buffer (deionized formamide and 25 mM ethylenediamine tetraacetic acid [EDTA], pH 8.0, containing 50 mg/ml blue dextran at a ratio of 5:1 formamide to EDTA/blue dextran). The DNA sample was denatured by heating at 90°C for 2 min after vortexing and spinning, and was then placed on ice until it was ready for sequence analysis. An aliquot of 2 µl of the DNA sample was loaded onto an ABI PRISM autosequencer (Model 377) for nucleotide sequence determination.

Measurement of Lipid Peroxides by High-Pressure Liquid Chromatography

The concentration of lipid peroxides in blood plasma or lung tissues was measured as malondialdehyde (MDA) according to the method described by Wong and associates (35). The blood sample was centrifuged at 900 × g for 20 min. The plasma in the supernatant was collected with the platelet and the buffy coat, with great care taken to avoid contamination, and was stored in liquid nitrogen until use. An aliquot of 40 µl of the blood plasma was pipetted into Eppendorf vials containing 0.6 ml of 0.44 M phosphoric acid. After vortex mixing, 0.2 ml of 42 mM thiobarbituric acid (TBA) solution was added to each vial. The final volume was then adjusted to 1.2 ml with distilled water. The vials were capped and boiled at 95°C for 60 min, and then cooled in an ice-water bath. An aliquot of 0.25 ml of each boiled sample was pipetted into another vial and then neutralized with 0.25 ml of methanol-1 N NaOH (10:1, vol/vol) solution. The neutralized solution was centrifuged at 9,500 × g for 5 min to precipitate denatured proteins. An aliquot of 25 µl of the supernatant was then injected into a high-pressure liquid chromatography (HPLC) system (C₁₈ column, 4.6 × 250 mm; Waters Corp., Milford, MA) and eluted with a solvent system of methanol and 50 mM potassium phosphate buffer (pH 6.8) (4:6, vol/vol). The flow rate of the solvent was 1.0 ml/min, and the absorbance of each eluate was monitored with a fluorescence detector, using an excitation wavelength of 525 nm and emission wavelength of 550 nm. The area of the absorption peak of the MDA-TBA adduct was determined by an integrator, and the concentration of lipid peroxides of each sample was then calculated from a calibration curve constructed by using 1,1,3,3-tetraethoxypropane as the standard. The concentration of lipid peroxides was expressed as µmol/liter plasma.

The content of lipid peroxides in the lung tissues was determined in a similar fashion after pretreatment of the tissues according to the method of Draper and Hadley (36). A piece of the frozen lung tissue (0.05 g) was weighed and then homogenized in 0.5 ml of 0.44 M phosphoric acid plus 0.05 ml of 0.2% butylated hydroxytoluene (BHT) in an ice-water bath, using a microhomogenizer with a Teflon pestle at low speed for 1 min. The content of lipid peroxides of the lung homogenate in terms of MDA-TBA adduct was assayed as described previously. The lipid peroxide content of the lung tissue was expressed as µmol/g wet weight.

Statistical Analysis

The frequencies of occurrence of the mtDNA deletions examined in the study in the lung tissues of patients in different smoking-index cohorts were compared and analyzed statistically with the chi-square test and Fisher's exact test for frequencies of occurrence less than 5. The differences between two groups in continuous variables were compared through use of the independent Student's t test or, for more than two groups, through one-way analysis of variance (ANOVA). The difference between groups in the proportion of the mtDNA deletion relative to wild-type mtDNA was compared through one-way ANOVA with Bonferroni's multiple-comparison test. Stepwise multiple regression analysis was adopted to identify the variable with statistical significance. A value of P < 0.05 was considered statistically significant.

	Results

Abstract Introduction Materials and Methods Results Discussion References

We recruited 101 males and 51 females into this study. Their ages ranged from 34 wk of gestation to 81 yr, with a mean ± SD age of 57.4 ± 17.6 yr. The mtDNA in each of the lung samples was examined by agarose-gel electrophoresis of PCR products amplified with four specific pairs of primers (Table 1). The primer-shift PCR experiment clearly demonstrated a specific 4,839-bp deletion in the mtDNA of smokers' lung tissues, which is different from the common 4,977-bp deletion (Figure 1). No mtDNA with the 4,977-bp deletion could be amplified using primer pair L5 and H4, which encompass some nucleotide sequences inside the 4,977-bp deletion. The PCR products amplified with primer pairs L2/H8 and L5/H4 showed that the 4,839-bp deletion in mtDNA is different from the common 4,977-bp deletion in mtDNA. Eighty of the 152 lung tissue samples (52.6%) from subjects with a mean age of 63.3 ± 10.3 yr (mean ± SD) were found to harbor mtDNA with the 4,839-bp deletion. The remaining 72 subjects, who did not have PCR-detectable mtDNA with the 4,839-bp deletion, were younger and had smoking indices lower than those of the subjects with the 4,839-bp deletion (P < 0.001). The pulmonary function indices of the subjects in these two groups were similar, as shown in Table 2. The frequencies of occurrence of mtDNA with the 4,839-bp, 4,977-bp, and 7,436-bp deletions in the lung tissues of different smoking-index cohorts are presented in Table 3. The incidence of the 4,839-bp mtDNA deletion was found to increase significantly with an increase in smoking index.

Results of the determination of the relative proportion of mtDNA with the 4,839-bp deletion through PCR done on a serially diluted DNA sample from one of the 80 subjects with this mtDNA deletion is illustrated in Figure 2. The proportion of mtDNA with the 4,839-bp deletion in lung tissues was found to increase with the smoking index (Figure 3). The average proportion of mtDNA showing the 4,839-bp deletion in lung tissues of subjects with a smoking index above 50 pack-yr was significantly higher than in those with smoking index of less than 40 pack-yr as shown in Table 4 (P < 0.05, one-way ANOVA with Bonferroni's multiple-comparison test). Subjects with smoking indices of 41 to 50 pack-yr had a higher proportion of mtDNA with the 4,839-bp deletion in their lung tissues than did nonsmokers (P < 0.05). The other two types of large-scale mtDNA deletions, of 4,977 bp and 7,436 bp, showed no significant differences in incidence or abundance among subjects with different smoking indices. The general descriptive data for the nonsmokers, ex-smokers, and current smokers are shown in Table 5. Nonsmokers had a younger age distribution, better FEV₁/FVC ratio, and greater female dominance than did ex-smokers or current smokers. Moreover, the incidence and proportion of mtDNA with the 4,839-bp deletion in the lung tissues of current smokers were significantly higher than in the lung tissues of nonsmokers. Ex-smokers, with similar mean smoking indices, had a similar incidence (78.9% versus 78.6%) and proportion of mtDNA with the 4,839-bp deletion to those of current smokers. The incidences and proportions of mtDNA with the 7,436-bp deletion were similar for subjects with different smoking habits. Although the incidences of the 4,977-bp deletion among current smokers and ex-smokers were significantly higher than that in nonsmokers, the proportions of mtDNA with the 4,977-bp deletion were similar for subjects with different smoking habits. Thus, the intergroup difference in incidence of mtDNA with the 4,977-bp deletion between groups was attributed to age.

View larger version (20K): [in this window] [in a new window]   Figure 3.   The proportions of mtDNA with three different deletions in subjects with different smoking indices. The average proportion of mtDNA with the 4,839-bp deletion in subjects with smoking indices of 41 to 50 pack-yr (n = 21) was significantly higher than that of nonsmokers (n = 74, *P < 0.05). Subjects with smoking indices greater than 50 pack-yr had a higher average proportion of mtDNA with the 4,839-bp deletion than did subjects with an index of less than 40 pack-yr (**P < 0.05, one-way ANOVA with Bonferroni's multiple-comparison test).

Direct sequencing of the 281-bp PCR product revealed that it was amplified from the mtDNA that harbored the novel 4,839-bp deletion, which is located between nucleotide position (np) 8711 and np 13,549, and is flanked by the 9-bp direct repeat sequence 5'-CATACACAA-3' (Figure 4). This deletion is 138 bp shorter than the 4,977-bp deletion. DNA sequencing was also performed on a cloned 275-bp DNA fragment from mtDNA with the 4,977-bp deletion. As expected, the results revealed a deletion of 4,977 bp, including the well-known 13-bp direct repeats associated with the common deletion (data not shown).

View larger version (23K): [in this window] [in a new window]   Figure 4.   DNA sequence of the 281-bp PCR product generated by using the L5/H4 primer pair, showing that the aberrant band was amplified from mtDNA with the novel 4,839-bp deletion. This deletion is located between np 8711 and np 13,549, and is flanked by a nine-nucleotide direct repeat of 5'-CATACACAA-3' between np 8703 and np 8711 or between np 13,541 and np 13,549, respectively.

On the other hand, we analyzed the content of lipid peroxides in the blood plasma and lung tissues. Thirty-six of 151 blood samples were excluded from analysis because of obvious hemolysis during sample preparation. The concentrations of lipid peroxides, measured as MDA, in the blood plasma of 115 subjects were similar in groups with different smoking indices (Table 4). The MDA levels in the lung tissues of current smokers were significantly higher than in those of nonsmokers (P < 0.05, one-way ANOVA with Bonferroni's multiple-comparison test). Moreover, the average MDA level in the lung tissues of ex-smokers fell between that of nonsmokers and current smokers. The MDA levels in lung tissues increased with the smoking index of the subjects examined (Table 4). As shown in Figure 5, patients with a smoking index above 50 pack-yr had significantly higher levels of MDA in their lung tissues than did nonsmokers (12.81 ± 4.99 µmol/g, n = 9 versus 5.39 ± 0.48 µmol/g, n = 64; P < 0.05).

View larger version (13K): [in this window] [in a new window]   Figure 5.   Lipid peroxide contents of lung tissues in subject cohorts with different smoking indices. The MDA content of lung tissue in subjects who smoked more than 50 pack-yr is higher than that of nonsmokers (*P < 0.05, one-way ANOVA with Bonferroni's multiple-comparison test). All data are presented as means ± SEM.

We then examined the content of lipid peroxides in lung tissues, the magnitude of mtDNA deletions, and the patients' pulmonary function indices to search for their relationships with age, sex, smoking index, and smoking habit. Correlation analysis showed that smoking index in terms of pack-year was positively correlated with age, tissue MDA content, and proportions of mtDNAs with the 4,839-bp and 4,977-bp deletions, and was negatively correlated with pulmonary function indices, including FEV₁ and the FEV₁/FVC ratio (P < 0.01). Stepwise multiple regression analysis was applied to analyze the relationship between the proportions of the 4,977-bp and 7,436-bp deletions and tissue MDA content, as well as all possible confounding factors, including age, sex, disease category, smoking habit, smoking index, and pulmonary function indices. The results showed that the smoking index in terms of pack-year was a significant independent factor for tissue MDA content (P < 0.001). Age was found to be the only independent factor related for the proportion of mtDNA showing the 4,977-bp deletion in lung tissue. Smoking index was the only independent factor related to tissue MDA content. Moreover, smoking index, tissue MDA content, and FEV₁/FVC ratio were statistically independent factors associated with the proportion of mtDNA with the 4,839-bp deletion in lung tissue. mtDNA with the 4,839-bp deletion was found to have accumulated in the lung tissues of 80 subjects as a function of smoking index in a non-zero-value exponential manner (r = 0.57, P < 0.0001). Cigarette smokers with a smoking index of 41 pack-yr or more were found to have a higher rate of moderately to severely obstructive ventilatory impairment in pulmonary function assessment than smokers with a smoking index of less than 40 pack-yr (10 of 30 smokers with a smoking index of more than 41 pack-yr versus five of 37 smokers with a smoking index less than 40 pack-yr; P = 0.053). With similar age distribution, FEV₁ and the FEV₁/FVC ratio were significantly lower in smokers with a smoking index of 41 pack-yr or more than in those of smokers with a smoking index of less than 40 pack-yr (FEV₁: 76.4 ± 20.8%, n = 30, versus 88.2 ± 15.6%, n = 37; FEV₁/FVC ratio: 64.4 ± 12.5%, n = 30, versus 71.2 ± 11.2%, n = 37; P < 0.05). The MDA content in the lung tissues from this group of heavy smokers (smoking index > 40 pack-yr) was also significantly higher than in the smokers with a smoking index of 40 pack-yr or less (11.2 ± 2.1 µmol/g tissue versus 6.8 ± 0.8 µmol/g tissue, mean ± SEM; P < 0.05).

	Discussion

Abstract Introduction Materials and Methods Results Discussion References

Lung tissues appear to be bathed in a continuous flux of ROS generated by endogenous sources through electron leak from the respiratory chain and exogenous sources such as cigarette smoking. It has been suggested that the free radicals in cigarette tar have five properties that enable them to damage DNA (9). First, the semiquinone radicals in cigarette tar are long-lived. Second, they are practically water soluble. Third, the free radicals from cigarette tar tend to associate with DNA. Fourth, the tar semiquinone radicals functionally reduce dioxygen to form superoxide anions, which are then metabolized to generate hydrogen peroxide. Fifth, the free radicals from tar (perhaps through phenolic functionalities) can chelate divalent iron or copper ions (2). Damaging hydroxyl radicals are then produced in the presence of these metal ions through the Fenton reaction. Free radical-mediated damage to DNA is a continually ongoing biochemical event with which the lung must deal throughout life. Human mtDNA, with its unique structural characteristics, is especially vulnerable to oxidative damage caused by ROS and free radicals (14, 18). Thus, it is conceivable that mtDNA in lung tissues is quite susceptible to oxidative damage compared with the DNA in other internal organs. Somatic mutations of mtDNA during the human lifespan have been proposed as one of the factors contributing to aging and degenerative diseases (17, 18). The accumulation of mtDNA with the 4,977-bp deletion is well recognized to be associated with aging in various human tissues (20). In one of our previous studies (31), we found that the incidence and amount of mtDNA with the 4,977-bp deletion in human lung tissues was no less than in other energy-demanding tissues such as brain (20), muscle (23), and heart (24). Exogenous oxidative stress, as exerted by relatively high oxygen tension in the alveoli, noxious environmental agents, cigarette smoking, and the defense systems (e.g., alveolar macrophages) of the lung may enhance the generation of free radicals that are damaging to mtDNA and other biomolecules in the lung tissues. mtDNA mutations, and especially large-scale deletions, may serve as genetic markers (scars) of molecular changes in human lung tissues during the aging process, and of past exogenous adventitious injuries. A vicious cycle of electron leak, oxidative stress, mtDNA mutation, and enhanced lipid peroxidation keeps operating in lung tissue throughout the human lifespan (37). Enhanced mtDNA damage has been shown to correlate with the degree of lipid peroxidation both in vitro (38) and in various human tissues (39, 40). It was previously demonstrated that mtDNA deletions and the lipid peroxide content of submitochondrial particles (SMP), as well as the activity of manganese-superoxide dismutase (MnSOD) in mitochondria, increased simultaneously in human muscle, liver, and testis tissues with aging (39, 40). A positive correlation was found between the proportion of mtDNA with the 4,977-bp deletion and the lipid peroxide content of mitochondria in human tissues (40). Ballinger and colleagues (41) recently reported that cigarette smokers have increased genetic damage in both the nuclear and mitochondrial genomes of their pulmonary macrophages. By using extra-long PCR and serial dilution assay for mtDNA with the 4,977-bp deletion, they demonstrated that the pulmonary macrophages of the smokers harbored 5.6 times more mtDNA damage and 2.6 times more damage at a nuclear locus than did the pulmonary macrophages of nonsmokers. Although the proportion of mtDNA with the 4,977-bp deletion in the smokers was almost seven times greater than in the nonsmokers, the difference was not significant because of great interindividual variation. This is consistent with our previous finding that the frequency of occurrence and proportion of mtDNA with the 4,977-bp deletion in the lung are only related to aging (31). It may also indicate that mtDNA with the 4,977-bp deletion is not a good marker for evaluation of the oxidative damage caused by cigarette smoking.

In the present study, we demonstrated a novel 4,839-bp deletion of mtDNA in human lung tissues. This novel 4,839-bp deletion generates a fusion protein of adenosine triphosphatase 6 (ATPase 6) and ND5 peptides and causes a frameshift mutation downstream from the deletion. In the case of the ND5 polypeptide, translation of the fusion protein is terminated at amino acid residue 29 downstream of the deletion (11). The frequency of occurrence and the proportion of mtDNA with the 4,839-bp deletion increase with an increasing smoking index in terms of pack-year. We found the proportions of mtDNA with the 4,839-bp deletion in the lung tissues of current smokers to be significantly higher than in those of nonsmokers. Because the difference might be small and the interindividual variation was greater, we used multiple regression analysis to dissect out the significance of the association between various possible confounding factors. Further analysis of the relationship between the proportion of mtDNA showing deletions and smoking index (pack-year), with consideration of other possible confounding factors, revealed that smoking index, tissue lipid peroxide content, and a pulmonary function index (the FEV₁/FVC ratio) were variables independently related to the proportion of mtDNA with the 4,839-bp deletion. Furthermore, mtDNA showing this deletion was found to accumulate in the lung tissues of smokers in an exponential manner, with a correlation coefficient of 0.57 (P < 0.001). Cigarette smokers with a high smoking index (41 pack-yr or more) showed a significant decrease in pulmonary function indices in terms of FEV₁ and the FEV₁/ FVC ratio, an increase in the proportion of mtDNA with the 4,839-bp deletion, and an increased lipid peroxide content in their lung tissues compared with smokers who had a smoking index less than 40 pack-yr. These findings provide support for the concept that cigarette smoke plays a significant role in the increase in mtDNA mutation and lipid peroxidation in the lung tissues of smokers. We suggest that mtDNA with the novel 4,839-bp deletion described here may serve as a good and reliable molecular marker for evaluating the damaging effect of cigarette smoking on the human lung.