Aqueous extracts of cigarette tar containing the tar free radical cause DNA nicks in mammalian cells.

The ability of aqueous extracts of cigarette tar to nick DNA was investigated using viable mammalian cells. Tar extracts contain a radical with a stable electron spin resonance (ESR) signal at g = 2.0036 characteristic of a semiquinone. The association of the tar component that carries the ESR signal with DNA was demonstrated using viable rat alveolar macrophages. The formation of single-strand DNA breaks caused by cigarette tar extracts in viable rat thymocytes follows saturation kinetics, indicating a tar component associates with DNA and then nicks it. These studies support our hypothesis that tar components that contain the cigarette tar radical can enter cells, associate with, and then nick DNA.


Introduction
Cigarette smoking is a major cause of human lung cancer (1), and the evidence suggests that radical mechanisms are involved in causing the DNA damage that could lead to cancer (2,3). Cigarette tar contains high concentrations (>1017 spins/gram) of stable radicals that can be observed directly by electron spin resonance (ESR) (4). At least four radicals have been identified based on their ESR spectral characteristics; the most interesting is a semiquinone in equilibrium with quinones and hydroquinones in a low molecular weight, tar matrix (2,4,5).
For each commercial cigarette smoked, up to 12 mg of particulate material (excluding nicotine and water) is deposited in the lungs of a smoker (6). This material comes in contact with pulmonary fluids that wash over it and extract the water-soluble components. Therefore, we believe aqueous cigarette tar extract (ACT) solutions are a realistic model of the mixture of chemicals that lung cells are continuously exposed to in the lungs of a smoker.
The tar fraction that contains the semiquinone radical associates with DNA (7). When calf thymus DNA is incubated This paper was presented at the Conference on with buffer solutions containing the tar radical, the DNA recovered by precipitation contains the tar radical ESR signal, and this signal is not removed by washing the DNA (7). Aqueous cigarette tar extracts nick plasmid DNA (8), producing nicks that are not easily repairable (9); thus, error-prone mechanisms may be involved in the repair of the tar radicalinduced nicks and these erroneous repairs could lead to mutations. When plasmid DNA is incubated with ACT, the resultant DNA nicking appears to follow saturation kinetics (10). Based on these preliminary DNA nicking studies, we have proposed a model in which the tar radical first associates with, and then nicks DNA (3).
It is our hypothesis (2) that superoxide and hydrogen peroxide (H202) result from the reduction of dioxygen by the hydroquinones and/or semiquinone radicals present in smoke extracts (Equations 1-3).
H202 + Fe+2 -HO + OH-+ Fe+3 [4] The iron in reaction 4 probably is complexed to the DNA via polyhydroxyaro-matic components in the tar (3,11). In fact, Ghio et al. have found that aqueous cigarette tar extracts complex iron in vitro and in vivo (personal communication).
We report here studies that show the tar radical component in ACT produces DNA nicks using viable mammalian cells. We also report, for the first time, ESR spectra of the tar radical in aqueous solutions and show that the number of DNA nicks produced is proportional to the amount of tar present. We also report the effects of reduced glutathione (GSH), deferoxamine, catalase, and superoxide dismutase (SOD) on the yield of DNA nicks caused by tar extracts.

Materials and Methods
Materials Chemicals and enzymes (unless noted otherwise) were purchased from Sigma Chemical Co. (St. Louis, MO) and used without further purification. Reduced glutathione, 800 mM, pH 8.1, catalase 1 mg/ml, and SOD 1 mg/ml were all prepared just before use. Catalase was inactivated by boiling for 2 min in 1 N NaOH and then the pH was adjusted to 7.4 with HCl; SOD was inactivated by boiling for 2 min. Inactivated enzyme solutions were stored at 40C.

Preparation ofACT
Research grade cigarettes (1R2) from the Kentucky Tobacco Research Council were smoked to a butt length of 30 mm using a standard puff profile (30 ml puff/30 sec).
The tar from four cigarettes was collected Environmental Health Perspectives on a Cambridge filter (a glass fiber filter that retains 99.9% of the particles larger than 0.1 micron), and the filter was extracted with 5 ml of pH 8 PBS (phosphate-buffered saline: 0.066 M phosphate, 0.85%, w/v, NaCl). After standing in the dark at 370C for 24 hr, the solution was filtered with a 0.2-pm filter (25 mm, polypropylene, Whatman International Ltd., Mainstone, England). This method produces ACT solutions that contain 8 to 10 mg of tar/ml of solution. An alternate method was employed to increase the tar yield. Tar deposits were washed off a Cambridge filter with approximately 15 ml of acetone until the filter was colorless, and the acetone was removed by evaporation with a stream of nitrogen. The dried residue was then resuspended in 5 ml of PBS, pH 8, in a vial and the solution was sonicated for 15 min using a Branson 2200 sonicator (Branson Ultrasonics). After standing at 37°C in the dark for 24 hr, this ACT solution was filtered with a 0.2-pm filter (25 mm, polypropylene, Whatman). This second method yields ACT solutions that contain 20 to 27 mg of tar/ml. We believe this second method of ACT preparation yields the same kind of tar only in a higher concentration. Therefore, for each experiment the volume of ACT solution used was adjusted to keep the amount of tar constant.

Separation ofACT by Membrane Filtration
Using Centricell 30 K and 10 K membrane filters (Polyscience, Inc., Warrington, PA), three molecular weight (mw) fractions of tar were obtained: fraction 1 >30,000 mw, fraction 2 between 10,000 and 30,000 mw, and fraction 3 < 10,000 mw. The amount of tar was determined to be 1.0 mg/ml, 1.6 mg/ml, and 24.4 mg/ml for fractions 1, 2, and 3, respectively. For each experiment, the volume of each fraction used was adjusted to keep the concentration of tar constant.

Isolation ofAlveolar Macrophages and Thymocytes
A male Sprague-Dawley specific pathogenfree rat 92 to 94 days old (360-400 g) (Harlan Sprague-Dawley, Houston, TX) was anesthetized with phenobarbital and then sacrificed by exsanguination. The thymus was removed, placed in a normal saline solution and then mashed with tweezers to release the thymocytes. The cells were pelleted by centrifugation at 400g and resuspended in 0.87% ammonium chloride/10 mM Tris-HCl, pH 7.2, 10 mM sodium bicarbonate solution to lyse any red blood cells. After 20 min, the cells were centrifuged at 400g and resuspended in a buffered isotonic solution, pH 7.2, containing 0.25 M myoinositol, 10 mM sodium phosphate, and 1 mM magnesium chloride. The thymocytes were counted, and their viability determined using the trypan blue exclusion assay. This procedure yields 1.5 to 1.8 x 109 cells with 90 to 95% viability.
To harvest macrophages from lung, 10 ml of a normal saline solution (37°C) was instilled intratracheally into the lungs, the solution was allowed to remain for one minute, and then recovered. This was repeated three times and the lavage solutions were pooled. Alveolar macrophages were separated from the lavage solutions by centrifugation at 400g. Typically, 5 to 6 million cells with 85 to 90% viability were harvested.

DNA Nlcking in Thymocytes
Thymocytes (17 million cells/tube, 1.2 ml total volume) were incubated on ice with aliquots of ACT for 90 min. For experiments involving inhibitors, the inhibitor was added to the cells prior to the incubation. The cells were then pelleted by centrifugation and resuspended in a buffered isotonic solution, pH 7.2, containing 0.25 M myoinositol, 10 mM sodium phosphate, and 1 mM magnesium chloride. The fluorescence analysis of DNA unwinding (FADU) assay has been described in detail by Birnboim (12,13). Briefly, the cells are divided into three types of tubes: T (total double-stranded DNA) tubes, in which care is taken to prevent DNA unwinding, indicate the total (maximum) amount of double-stranded DNA present. P (partially unwound DNA) tubes, which are treated with a viscous alkaline solution that allows for partial unwinding of the DNA, are used to assess DNA damage. B (blank, no double-stranded DNA) tubes, are treated with the viscous alkaline solution and then sonicated with a Branson sonifier 450, 15 pulses at 20% output (Branson Ultrasonics) to effect maximum unwinding of the DNA, and contain no double-stranded DNA. At this stage, the volume of each tube is 1.0 ml, then 1.5 ml of a 6.67 pg/ml solution of ethidium bromide is added to each tube. The amount of double-stranded DNA in each tube is determined by measuring the fluorescence of ethidium bromide using 520 nm excitation and 590 nm emission. Three of each type of tubes, T, P, and B, are used for each determination, and the results are averaged. The amount of double-stranded DNA remaining is calculated from Equation 5: The amount DNA damage is determined by calculation of the damage quotient Qd, as shown in Equation 6: Qd = (log DC-log DACT) X 100 [6] where DC = double-stranded DNA in control cells and DA,T = double-stranded DNA in ACT-treated cells (12). Using these calculations, Birnboim demonstrated a linear relationship between Qd and the amount of cobalt-60 gamma radiation that lymphocytes received (12). The amount of protection provided by the various inhibitors is calculated as shown in Equation 7: % Pro-= (1 Qdinhibitor/QdAC) x 100 tection [7] Binding of the Cigaette Tar Radical to DNA in Rat Alveolar Marophages RAM, isolated as described above, were resuspended in 2.0 ml of a solution containing 24 mM EDTA and 75 mM NaCl (the solution was adjusted to pH 7.5 with 1 N NaOH). The cells were then incubated with 1.0 ml of an ACT solution, or 1.0 ml of PBS, pH 8.0. In both incubations, the cells were allowed to sit on ice for 90 min to allow for maximum penetration of the tar into viable cells. The cell viability decreased markedly for incubations longer than 90 min. After the incubation, exclusion of trypan blue dye was used to determine cell viability to be 85 to 90%. Modification of the alkaline elution procedure (14)

Results racton ofthe Tar Radical into Aqueous Buffer Solutions
Previously, we have reported ESR spectra of the tar radical in organic solvents or on a filter (4). The tar radical signal is extracted into aqueous buffers at pH 8.0 and its ESR spectrum observed in aqueous buffers ( Figure 1). These ACT solutions contain a broad ESR signal with a g value of 2.0035, which we have assigned to the tar semiquinone radical (4). This g value is typical of organic semiquinones (16) and organic semiquinone radicals previously have been observed in aqueous solutions at pH 8.0 (17).

DNA-Tar Radical Complexes Formed with Caf-Thymus DNA
Tar extracts contain stable semiquinone radicals that become associated with double-stranded DNA (7). In preliminary experiments, calf thymus DNA was incubated with ACT for 18 hr at 37°C and the DNA precipitated with cold ethanol. The ESR spectra of the ethanol precipitates from incubation of ACT with thymus DNA are shown in Figure 2. The treatment of ACT alone with cold ethanol does not result in the precipitation of any material. The DNA that precipitated from ACT-DNA incubations had an ESR signal with a g value of 2.0044, in agreement with our previous report (7) and within the range for semiquinone radical signals (16).
Next, calf thymus DNA was incubated with ACT and the DNA was isolated on a polycarbonate filter that impedes the passage of long DNA strands (14). These incubations result in the semiquinone radical becoming retained by the filter; Figure  3 shows the ESR spectrum observed from the dried polycarbonate filter. Control incubations of ACT alone, or calf thymus DNA alone, showed no radical signal remaining on the filter. This indicates that the DNA-tar radical complex is retained on the polycarbonate filter, and there is no nonspecific retention of the tar radical on the filters in the absence of DNA.

A DNA-Tar Radical Complex in Rat Alveolar Macrophages
Since the DNA-radical complex can be trapped on a polycarbonate filter, incubations of ACT with viable RAM cells were performed. RAM were incubated with ACT solutions and the DNA was immobilized on polycarbonate filters using a modification of the alkaline elution method of Kohn et al. (14). The ESR spectra of the dried filters are shown in   The results of the FADU assay for ACTinduced DNA nicks in rat thymocytes are shown in Figure 5. The DNA damage quotient (Qd) represents the amount of DNA nicks in the rat thymocytes. (The calculation of Qd is explained in Materials and Methods.) When Qd is plotted versus the total concentration of tar in the aqueous extracts, the maximum Qd and the binding constant can be determined either from a curve-fitting program or by Lineweaver-Burk or Eadie-Hofstee plots of the data. These three analyses are summarized in Table 1, and the Lineweaver-Burk plot is shown in the insert of Figure 5. The maximum value for Qd is 98 ± 13 and the binding constant is 2144 ± 237 )Jg/ml for tar extracts. These data indicate saturation kinetics, with a maximum amount of DNA damage occurring at a tar concentration of 2144 pg/ml. This is the amount of tar produced by approximately 0.23 cigarette. Membrane filters were used to separate the ACT solution into three fractions, and these fractions were tested for DNA nicking activity. Only fraction number 3, the fraction containing material with MW less than 10,000 amu, caused DNA nicks, as shown in Figure 6. This fraction also contains the semiquinone radical that associates with the DNA in RAM, as shown in Figure 4.

Eflects ofInhibitors on DNA Damage Caused by ACI Solutions
Catalase, SOD, GSH, deferoxamine, and mannitol were tested to determine if they protect DNA against nicking caused by tar extracts; the results are summarized in Table 2. Catalase and GSH protected the cells from DNA damage caused by ACT solutions. As the GSH concentration was increased, the amount of protection against DNA nicking by ACT also increased, as shown in Figure 7. Neither deferoxamine nor SOD protected against DNA nicking caused by ACT solutions. These results are summarized in Table 2.

Discussion
Studies in our laboratory with cigarette tarinduced DNA nicking were done with isolated plasmid DNA (8) or calf thymus DNA (7). We have extended our studies to determine the effects of cigarette tar in viable cells. We have also shown that the tar radical can be extracted into aqueous buffer solutions and directly observed by ESR.
[We had previously reported ESR spin trap spectra of the hydroxyl radical that is produced by these ACT solutions (18).]   For DNA binding studie used RAM. These cells are targ rette smoke damage (19)(20)(21)(22)(23)(24)(25)(26)(27) ai isolated as a homogeneous cell (28) from the lavage of rat lung modified the alkaline elution Kohn et al. (14) so that it retains only double-stranded RAM on polycarbonate filters. method, we are able to isolate intact cells that have been expo and show that the tar radical l DNA in these cells.
For DNA nicking studi employed the FADU method    (12). This method allows the detection of DNA single-strand breaks in cells that have s, we have been exposed to ACT. For these studies we ;ets for ciga-have used rat thymocytes; while these are nd are easily not primary target cells for DNA damage population in smokers, large numbers are easily gs. We have obtained as a homogeneous cell populamethod of tion. We are able to routinely isolate 1.5 to selectively 1.8 billion cells, the number required to DNA from complete a dose response curve. Therefore, Using this thymocytes are a useful and experimentally DNA from accessible model system that we have used )sed to ACT in our initial probing of the effect of ACT binds to the in viable cells (13). Leanderson and Tagesson studied the es we have ability of aqueous cigarette tar extracts to of Birnboim promote human polymorphonuclear leucocyte and hydrogen peroxide-induced DNA single-strand breaks in cultured human + bronchiolar cells (29). The tar alone did not cause any single-strand DNA breaks. These workers used tar solutions that were 8-fold more dilute than the most dilute studied here, and apparently were too dilute to cause nicking. They also prepared their tar solutions differently than we, bub-150 200 bling whole smoke through chloroform, evaporating the chloroform and redissolvagainst ACT-ing the tar in 10 ml of buffer.
3tion of reduced Nakayama et al. reported that cigarette smoke produces single-strand breaks in cul-Environmental Health Perspectives f tured human A549 lung carcinoma cells and the number of breaks was reduced by radical scavengers (30). In their experiments, smoke extracts were prepared by bubbling the smoke from two commercial filtered cigarettes into a phosphate buffer. The authors noted the large variability (5fold) in their results when different solutions of smoke were employed, and they attribute these difference to fluctuating conditions in trapping the smoke and the variable quality of commercial filters and cigarettes. Fielding et al. sought to determine if the tar fraction of smoke extracts caused these single-strand breaks (31). These workers used the smoke from four types of commercial cigarettes: ultra low, low, and medium tar with filters and high tar without a filter. The smoke was bubbled through buffer and the extracts then tested for DNA nicking in cultured A549 cells. These workers find no correlation between the amount of tar from a particular cigarette and the extent of DNA nicking observed, and they conclude that tar is not involved in DNA nicking. Our results indicating that aqueous tar extracts do nick DNA conflict with this conclusion. Neither Nakayama et al. nor Fielding et al. separated gas-phase cigarette smoke and tar by the standard methods we used, and both used commercial cigarettes, whereas we used standard research cigarettes.
We find that increasing amounts of tar extracts result in higher amounts of DNA damage up to a maximum (saturation) point. Using DNA isolated from rat lung and a 32P-postlabeling assay for DNA adducts, Randerath et al. also found a correlation between the amount of DNA adducts and the amount of cigarette smoke extract used in the incubations (32). Their results, like ours, show that the yield of adducts reaches a saturation maximum (32). These workers found that glutathione inhibits adduct formation, and they proposed that radical mechanisms are involved in the formation of tar-DNA adducts. Model of the tar radical associating with and nicking DNA. 1) Aqueous extracts of cigarette tar contain a semiquinone/hydroquinone radical in a low molecular weight polymeric matrix that binds to DNA. 2) These semiquinones reduce oxygen to form superoxide. 3) Superoxide then dismutates to form hydrogen peroxide; or hydrogen peroxide may diffuse into the nucleus from other loci in the cell. 4) Metals chelated to the tar react with hydrogen peroxide to form hydroxyl radical (or ferryl) species that nick DNA. This figure has been modified from Pryor and Stone ( 1 1). Cigarette smoke solutions have been shown to produce hydrogen peroxide (33). The DNA nicking caused by ACT is inhibited by catalase; hydrogen peroxide is implicated in causing DNA nicks in mammalian cells exposed to ACT. Similarly, Evans et al. showed that catalase protected against cigarette tar damage to the protein alpha-1-proteinase inhibitor (34).
Glutathione protects against DNA nicking by ACT. Glutathione is known to form covalent adducts with quinones and hydroquinones (35)(36)(37). Thus, the GSH protection of DNA may be related to the ability of GSH to add to quinones and hydroquinones and perhaps prevent the addition to and nicking of DNA, or to the well known ability of GSH to quench radical signals (38).
Aqueous extracts of cigarette tar con-tain a tar semiquinone radical. These extracts cause DNA nicking in intact cells and this nicking follows saturation kinetics, suggesting that cigarette tar associates with DNA and then causes nicks (11). The maximum nicking occurs at concentrations of tar that are equivalent to 0.23 cigarette, using incubations of 17 x 106 thymocytes.
The tar radical also associates with DNA in viable rat lung macrophages. These results show that tar components are capable of entering a viable cell, penetrating the nucleus, and interacting with DNA, and support our model of the tar radical associating with, and then nicking, DNA (11) as shown in Scheme 1. This provides further evidence that the cigarette tar component containing the free radical is involved in the DNA damage and subsequent carcinogenicity of cigarette tar.