Antioxidant-induced changes in oxidized DNA

Abstract

N-acetylcysteine (NAC), a strong antioxidant, has antigenotoxic and anticarcinogenic properties currently being investigated in clinical trials. NAC detoxifies free radicals (e.g., the hydroxyl radical, ·OH) that cause DNA changes implicated in disease (e.g., cancer). The ·OH reacts with purines to form mutagenic 8-hydroxypurine (8-OH) and putatively nonmutagenic formamidopyrimidine (Fapy) lesions. Fapy lesions inhibit DNA synthesis likely modulating the mutagenic potential of the 8-OH lesions, which would suggest that the ratio of these oxidized bases is biologically important. However, little is known about how NAC modifies oxidized DNA structure or how such modifications may affect cellular processes, such as replication and transcription. By using gas chromatography-mass spectrometry and Fourier transform-infrared spectroscopy, we found that dietary NAC (5% in the diet for 14 days) affected ·OH-induced structural changes in DNA of the hind leg of the BALB/c mouse. For example, mutagenic 8-hydroxyguanine (8-OH-Gua) was reduced ≈50% (P = 0.02) in mice fed NAC compared with controls. NAC reduced the log₁₀ (8-OH-Gua/FapyGua) ratio from 0.58 ± 0.15 to essentially zero, a virtually neutral redox status. DNA from control mice had a remarkably high variance compared with mice fed NAC. Moreover, the DNA from treated and control mice was distinct with respect to base structure and vertical base-stacking interactions. The findings showing that NAC lowered the concentration of 8-OH-Gua, the log ratio, and the variance (previously associated with neoplastic changes) suggest that NAC reduces the mutagenic potential of oxidized DNA. These benefits could be offset by the other structural changes found after NAC exposure, which may affect the fidelity of DNA synthesis.

The thiol-containing nucleophile, N-acetylcysteine (NAC), is capable of detoxifying electrophiles and free radicals that are implicated in cancer and other diseases (1). Initiation of clinical trials with NAC in the last several years in the United States (2) and Europe (3, 4) attests to a strong interest in this molecule as a promising antigenotoxic and anticarcinogenic agent having a very low toxicity (1). However, it also has been reported that NAC causes oxidative damage to cellular and isolated DNA, and it has been shown to both increase (5) and decrease (6, 7) cellular levels of 8-OH-guanine (8-OH-Gua), a hydroxyl radical (·OH)-induced mutagenic base lesion implicated in tumor formation (8–10). Moreover, results from the long-awaited EUROSCAN trial showed that patients with cancers of the lung or head and neck did not benefit from treatment with NAC (4). Understanding these varied findings is limited by the fact that very little is known about how NAC affects the oxidative status of DNA and how this might alter its biological properties.

Preliminary studies suggested the hind limb of the BALB/c mouse, which exhibited a relatively high degree of ·OH-induced base damage (unpublished data), could be an attractive model for studying how dietary NAC may alter radical-induced changes in DNA. The mouse leg is not known to be naturally prone to disease in response to elevated levels of reactive oxygen species. Moreover, the high degree of oxidation we observed may be exercise-related. For example, DNA damage reportedly occurs in human muscle tissue under exercise-induced oxidative stress (11). The mouse leg can probably tolerate higher levels of oxidative damage in DNA than cancer-prone tissues (e.g., breast and prostate) without the risk of malignancy, in part, because of the virtual lack of cell proliferation in the muscle.

In the present study, we used gas chromatography-mass spectrometry (GC-MS) to quantify ·OH-induced base lesions in the DNA (12, 13) of mice fed either a normal diet or an NAC-supplemented diet. We also used Fourier transform-infrared (FT-IR) spectroscopy (14, 15) to provide broader insight into NAC-induced changes (e.g., conformational) in the base structure. The complementary use of these two technologies provided a unique perspective on the potential of NAC to modify ·OH-induced DNA alterations known to be associated with adverse biological change.

Materials and Methods

Tissue Acquisition and DNA Isolation.

Six- to eight-week old female BALB/c mice were obtained from Charles River Breeding Laboratories. The mice were kept on a 12 h light/12 h dark cycle and were allowed free access to food and water. The control group (n = 12) was fed a normal diet (Teklad chow, Dyets, Bethlehem, PA). The experimental group (n = 10) was fed the same diet supplemented with 5% NAC (>99% purity, Sigma) prepared by Dyets. Each group was maintained on the diet for 14 days, after which the mice were killed. The hind leg muscles were removed, immediately frozen in liquid nitrogen, and maintained at −80°C until the DNA was extracted. The experimental protocols were approved by the Institutional Animal Care and Use Committee of the Pacific Northwest Research Institute.

DNA (≈50 μg) was extracted from the hind leg muscle (≈250 mg) by using Qiagen 100/G Genomic-tips (Qiagen, Chatsworth, CA) following the prescribed Qiagen extraction protocol. After extraction, DNA was passed through a 5.0 micron Cameo 25N syringe filter (Osmonics, Minnetonka, MN). The DNA was washed three times with ice cold 70% ethanol. Then, it was dissolved in 10–40 μl of optima grade water (Fisher Scientific) in preparation for GC-MS (12, 13) and FT-IR spectral analysis (14, 15).

GC-MS: Procedure and Statistical Analysis.

The procedure for GC-MS has been described (12, 13). Briefly, DNA (≈ 20 μg) was hydrolyzed with 150 μl of 60% formic acid in Reacti-vials (Pierce). Lyophilized hydrolysates were then derivatized with 50 μl of bis(trimethylsilyl)trifluoroacetic acid containing 1% trimethylchlorosilane and acetonitrile (4:1, vol/vol). The samples were analyzed by GC-MS by using an HP model 6890 with an HP model 5973 mass spectrometer (Hewlett–Packard). The purine-base lesions quantified were 8-OH-Gua, 8-hydroxyadenine (8-OH-Ade), 2,6-diamino-4-hydroxy-5-formamidopyrimidine (FapyGua, or Fapyguanine), and 4,6-diamino-5-formamidopyrimidine (FapyAde, or Fapyadenine).

One outlier (>3 SDs from the mean) was removed from each group of base lesion (GC-MS) data and was not included in statistical analysis (16). An F test was used to find significant differences in variance for DNA concentrations of base lesions (lesions/10⁵ unmodified base) and log₁₀ base lesion ratios. (Variance measures variability around the mean; log values were used because graphical analysis has shown that the logarithm of base lesion ratios is more normally distributed than values on the natural scale; ref. 10). T tests were performed to identify significant differences in concentrations of base lesions and log ratio values. Equal variance was assumed, unless the F test revealed significant differences in variance.

FT-IR Spectroscopy: Procedure and Statistical Analysis.

The analysis of DNA was conducted as described (14, 15) on splits of the samples used for GC-MS. Briefly, ≈1.0 μg of dry film was analyzed with an FT-IR microscope spectrometer (System 2000, Perkin–Elmer; ref. 15). Baseline absorbance (the mean of 11 absorbances centered at the minimum of the flat region between 2,000 and 1,750 cm⁻¹) was subtracted from the absorbance at each wave number. To remove the effects of differing film thicknesses, baselined spectra were normalized by dividing the absorbance at each wave number by the mean absorbance between 1,750 and 1,550 cm⁻¹ to yield a mean of 1.0 (17).

Statistical analysis of FT-IR spectra was used to provide further insight into NAC-induced structural changes in the DNA of mice fed NAC. This technology is highly sensitive for detecting subtle alterations in DNA base and backbone structures that are often not visually apparent (14, 15, 18, 19). The total spectral area examined was 1,750 to 1,315 cm⁻¹. This region includes stretching and bending vibrations of the nucleotide bases, as well as conformational changes associated with vertical base-stacking interactions (without interference from backbone vibrations; refs. 18 and 19). These changes are revealed by statistical differences between spectral means based on P values. Principal components analysis (PCA) of the FT-IR spectra, which involves hundreds of thousands of correlations between peak heights, wave numbers and other spectral properties (10), was undertaken as described (15, 17). Principal component (PC) scores then were used to construct a two-dimensional scatter plot. The spatial location and distribution (e.g., clustering) of PC scores serves to identify structural differences between groups of DNA.

Results and Discussion

Base Lesion Concentrations and Ratios.

Base lesion concentrations and ratios in the DNA of control mice fed the normal diet and mice fed the NAC-supplemented diet are given in Table 1. In both groups, the highest mean concentrations were found for 8-OH-Gua and FapyGua, with 8-OH-Gua having moderately higher values. The 8-OH-Gua values were two- to three-fold higher than those of 8-OH-Ade, and the values for FapyGua were three- to four-fold higher than those of FapyAde. The Fapy derivatives are putatively nongenotoxic (20), and recent evidence suggests that they modulate the mutagenicity of the 8-OH derivatives (8–10, 20).

Table 1.

DNA base lesions and base lesion ratios with P values from a t test for differences between hind leg tissues of mice on the normal diet and NAC-supplemented diet

Base lesions and log10 ratios	Controls (n = 11)		NAC group (n = 9)		P value
	Mean	Variance	Mean	Variance	Mean	Variance
8-OH-Gua/105 Gua	241	1.7  × 104	128	3.4  × 102	0.02	0.001
FapyGua/105 Gua	163	1.6  × 104	113	2.1  × 103	0.27	0.05
8-OH-Ade/105 Ade	80	8.7  × 102	57	4.2  × 102	0.07	0.23
FapyAde/105 Ade	46	8.6  × 102	32	1.4  × 101	0.16	0.02
Log10 (8-OH-Gua/FapyGua)	0.27	0.21	0.10	0.13	0.26	0.002
Log10 (8-OH-Ade/FapyAde)	0.30	0.05	0.27	0.22	0.79	0.61

The only significant difference between the mean base lesion values for DNA of the two mouse groups was found with 8-OH-Gua (P = 0.02). The relatively high concentration of 8-OH-Gua is consistent with the higher reactivity of the ·OH with guanine, compared with adenine, in aqueous solution (21). However, the proportion of guanine to adenine base lesions would be subject to modulation by cellular repair enzymes (22). The mean concentration of 8-OH-Gua was almost 50% lower in the group of mice on the NAC-supplemented diet, consistent with reports demonstrating the ability of this antioxidant to lower DNA concentrations of 8-OH-Gua (6). However, the NAC had little effect on FapyGua (Table 1), which is likely related to the propensity of this antioxidant to react with electrophilic species (1).

Control mice had 8-OH-Gua concentrations ranging from ≈150 to 500 base lesions/10⁵ Gua (Fig. 1A). This diverse group of base lesion values comprised two clusters: the higher (a) ranging from 280 to 500 base lesions/10⁵ Gua (mean = 381 ± 174), and the lower (b) ranging from 125 to 250/10⁵ Gua (mean = 161 ± 58). In sharp contrast, all values for the NAC group were tightly clustered, having concentrations from ≈ 80 to 150 base lesions/10⁵ Gua (mean = 147 ± 68). The mean values for the NAC group and cluster (b) were not significantly different (Fig. 1A). Thus, it seems that NAC is able to effectively reduce relatively high 8-OH-Gua concentrations to a restricted, low range of values, thereby decreasing mutagenic damage in the DNA.

Figure 1.

(A) Guanine lesion concentrations and (B) log₁₀ guanine lesion ratios are shown to illustrate structural differences in DNA of mice on the normal diet (n = 11; ●) and the NAC-supplemented diet (n = 9; ○); (a) and (b) identify distinct clusters of points.

Relationship Between 8-OH and Fapy Lesions.

It is useful to consider the relationship between 8-OH-Gua and the ring-opened (formamidopyrimidine) FapyGua structure (20). The ·OH arises from the metal (e.g., Fe⁺²)-catalyzed decomposition of H₂O₂, which readily crosses the nuclear membrane (8). The ·OH then reacts with purines to form the redox-sensitive 8-oxyl derivative (21), which is then converted to either the 8-OH or Fapy lesion. Oxidative conditions favor the formation of 8-OH lesions, whereas reductive conditions favor formation of Fapy lesions. The closely coupled redox-dependent reactions, together with the action of repair enzymes (e.g., glycosylases), are believed to be influential in establishing the proportion of 8-OH and Fapy lesions found in the DNA of different biological systems (10, 13). Fapy lesions block DNA synthesis (20, 23) and likely modulate the mutagenic potential of the 8-OH-Gua (13).

Accordingly, we determined log₁₀ (8-OH-Gua/FapyGua) values for both groups of mice. The mean log₁₀ (8-OH-Gua/FapyGua) value shown in Table 1 for mice fed NAC is almost three times lower than that for the control mice; however, this difference is not statistically significant. The wide range of values and limited number of samples in the control group potentially influenced this outcome. However, a different perspective was obtained when the values for individual mice were depicted in a scatter plot. The log₁₀ (8-OH-Gua/FapyGua) values (Fig. 1B) for the DNA of both groups of mice showed a distribution similar to those for 8-OH-Gua (Fig. 1A). In the controls, the cluster of higher DNA values (a) had a mean log ratio favoring 8-OH-Gua (0.58 ± 0.15; Fig. 1B). In contrast, the DNA of the NAC mice did not exhibit the high values of cluster (a). The mean log ratios for the lower cluster (b) and the NAC group were virtually identical (0.01 ± 0.21 and 0.01 ± 0.14). Thus, NAC seems to reduce the base lesion concentrations to a narrow range while also reducing log₁₀ (8-OH-Gua/FapyGua) to essentially zero (reflecting a neutral redox status for the DNA; ref. 21).

The ability of NAC to lower concentrations of 8-OH-Gua, together with the reduction of the mean log₁₀ (8-OH-Gua/FapyGua) value to a virtually neutral status is probably an important characteristic of this antioxidant in modulating mutagenicity. This property of NAC may be especially relevant in tumor-prone tissues where reduction in the mutagenicity of DNA and control of DNA synthesis are considered to be critical factors in limiting cancer risk (9).

Molecular Change Evinced by FT-IR Spectroscopy.

Mean FT-IR spectra of DNA from the control and NAC groups are shown in Fig. 2A. In such a comparison, subtle differences in spectral properties are sometimes barely visible. Nevertheless, t tests at each wave number identified areas where differences exist (Fig. 2B). Significant differences (P ≤ 0.05) between ≈1,711–1,699 cm⁻¹ (x) and ≈1,636–1,624 cm⁻¹ (y) both reflect CO-stretching and NH₂-bending vibrations (18, 19). Significant differences below ≈1500 cm⁻¹ are mainly related to weak NH vibrations, CH in-plane base deformations and vertical base-stacking interactions. The four base lesions (Table 1) undoubtedly contributed to these spectral differences; however, a number of additional ·OH-induced base lesions have been identified in biological systems (24). Some of these were almost certainly present in the mouse DNA and thus influenced the spectral findings.

Figure 2.

(A) Mean spectra of DNA from mice on the normal diet (n = 12; thick line) and NAC-supplemented diet (n = 10; thin line). (B) P values from a t test for differences in spectral means between DNA of mice on each diet (x and y represent spectral differences in CO-stretching and NH₂-bending vibrations of nucleotide bases).

The lower cluster of base lesion values and log₁₀ ratios for the controls (b) could not be distinguished from those for the mice fed NAC (Fig. 1). However, it was of interest to establish, by using FT-IR spectroscopy, whether the two groups of DNA from which these values were obtained were structurally different. This determination was accomplished by using PC scores 1 and 10 derived from the spectra of each group of DNA (representing almost 100% of the total spectral variance). Fig. 3A reveals a separation between the PC scores for the DNA of the control and NAC groups, thus indicating that each has distinct structural properties. It is likely that these distinct properties are largely attributable to differences in the unique assemblages of base lesions and the resulting conformational changes represented by vertical base-stacking interactions.

Figure 3.

(A) PC plot illustrating differences in the structure (distinct clustering) of DNA groups from mice on the normal diet [Fig. 1, cluster (b)] and mice on the NAC-supplemented diet (Fig. 1). (B) P values from an F test illustrating differences in variance for three spectral regions between 1,716 and 1,539 cm⁻¹. These regions represent CO-stretching and NH₂-bending vibrations of nucleotide bases.

Although the base lesion data per se showed no difference between the NAC group and cluster (b) of Fig. 1 A and B, the FT-IR spectral data did reveal significant differences associated with broader aspects of molecular structure that may alter important biological properties of DNA. In this regard, the ability of the ·OH to substantially modify FT-IR spectra (structure) of DNA has been demonstrated previously by comparing pure oligonucleotides with oligonucleotides containing an 8-oxo (OH) substitution. Notably, a single 8-oxoguanine lesion, centrally located on a 25-base strand, resulted in a number of changes compared with the parent structure, including those of vertical base-stacking interactions as well as conformational changes in the phosphodiester-deoxyribose structure (25). Backbone changes were not found in the present study, possibly because of the reduced conformational flexibility of the more rigid helical structure and the lower degree of radical-induced damage.

Variance.

Table 1 shows that the variance for concentrations of 8-OH-Gua (P = 0.001) and FapyGua (P = 0.05) is ≈10- to 50-fold lower, respectively, for the DNA of the NAC group compared with that of the control group. In addition, the variance of log₁₀ (8-OH-Gua/FapyGua) for the DNA of the NAC group is ≈60% lower than that for the control group (P = 0.002). The remarkably high variance values for the control mice also are reflected in the scatter plots of base lesion data (Fig. 1A and 1B).

An F test was conducted to compare the variance at each wave number of the FT-IR spectra for the control and NAC mice. Significant differences in variance were found in three distinct regions between 1,716 and 1,539 cm⁻¹ (Fig. 3B). These differences correspond to CO-stretching and NH₂-bending vibrations (18, 19). Thus, the differences in spectral variance for the DNA of the control and NAC mice are consistent with differences in variance for the base lesion data (Table 1). Significant differences in variance were not evident in the plot of PC values (Fig. 3A), in which each PC reflects a broad array of structural properties associated with nucleotide base structure. This result suggests that the variance differences found in the base structures are “obscured” by the complexity of broader aspects of molecular structure.

In previous studies, high variances in DNA spectra (26) and base lesion structure measured by GC-MS (27) were characteristic of primary tumor tissues prone to metastasis (e.g., of the breast and ovary) but not of metastases themselves. It has been suggested that the high variance of DNA structure in tissues prone to malignant transformation is obligatory for clonal selection (26). Thus, in some tissues, the capability of NAC to markedly reduce variance in DNA may well be an important biological property for inhibiting cellular transformations to the malignant state. The present findings show that NAC substantially reduces group variance in the oxidative status of DNA. Hence, the likelihood exists that NAC also would reduce periodic variations related to the oxidative status of DNA of individuals within the group.

Biological Implications.

The ability of dietary NAC to lower 8-OH-Gua levels in DNA, as we have shown, is perceived to be beneficial (8). Furthermore, the attendant reduction found in the log ratio, redox status, and variance is likely an important additional benefit. Given that the enzymatic machinery “reads” the coded information at the level of hydrogen bonds, the changes in hydrogen bonding associated with NAC-related modifications in vertical base-stacking interactions (Fig. 2B) may affect the fidelity of RNA and DNA synthesis (i.e., transcription and replication). Presently, it is not clear whether these NAC-induced modifications alter its anticarcinogenic properties (1) or modulate the beneficial effects of lower 8-OH-Gua levels and the other potentially beneficial changes found.

Abbreviations

NAC

N-acetylcysteine

FT-IR

Fourier transform-infrared

GC-MS

gas chromatography-mass spectrometry

8-OH-Gua

8-hydroxyguanine, Fapy, formamidopyrimidine

FapyGua

2,6-diamino-4-hydroxy-5-formamidopyrimidine, or Fapyguanine

FapyAde

4,6-diamino-5-formamidopyrimidine, or Fapyadenine

PC

principal component

PCA

principal components analysis