An Oxidized Purine Nucleoside Triphosphatase, MTH1, Suppresses Cell Death Caused by Oxidative Stress*

MTH1 hydrolyzes oxidized purine nucleoside triphosphates such as 8-oxo-2′-deoxyguanosine 5′-triphosphate (8-oxo-dGTP) and 2-hydroxy-2′-deoxyadenosine 5′-triphosphate (2-OH-dATP) and thus protects cells from damage caused by their misincorporation into DNA. In the present study, we established MTH1-null mouse embryo fibroblasts that were highly susceptible to cell dysfunction and death caused by exposure to H2O2, with morphological features of pyknosis and electron-dense deposits accumulated in mitochondria. The cell death observed was independent of both poly(ADP-ribose) polymerase and caspases. A high performance liquid chromatography tandem mass spectrometry analysis and immunofluorescence microscopy revealed a continuous accumulation of 8-oxo-guanine both in nuclear and mitochondrial DNA after exposure to H2O2. All of the H2O2-induced alterations observed in MTH1-null mouse embryo fibroblasts were effectively suppressed by the expression of wild type human MTH1 (hMTH1), whereas they were only partially suppressed by the expression of mutant hMTH1 defective in either 8-oxo-dGTPase or 2-OH-dATPase activity. Human MTH1 thus protects cells from H2O2-induced cell dysfunction and death by hydrolyzing oxidized purine nucleotides including 8-oxo-dGTP and 2-OH-dATP, and these alterations may be partly attributed to a mitochondrial dysfunction.

Cellular DNA, RNA, and their precursor nucleotides are at a high risk of being oxidized by reactive oxygen species, which are inevitably generated by normal metabolic functions such as mitochondrial respiration or by environmental exposure to ionizing radiation and chemicals. The oxidation of nucleic acids appears to result in either spontaneous mutagenesis or cell death, thus being implicated in various age-related diseases such as cancer and neurodegeneration (1). Among the various types of oxidized damage in nucleic acids, 8-oxo-G, 1 an oxidized form of guanine, and 8-oxo-dGTP are both considered to be one of the major sources for spontaneous mutagenesis. Free guanine nucleotides have been shown to be more susceptible to oxidation than guanine in double-stranded DNA (2). 8-Oxo-dGTP can be incorporated into the nascent strand opposite adenine and cytosine in the template with almost equal efficiency, thus resulting in an A:T to C:G transversion mutation (3,4). A misincorporation of oxidized nucleotides also occurs during transcription, thus leading to the production of abnormal proteins (5).
To minimize the accumulation of 8-oxo-G in DNA, organisms come equipped with elaborate mechanisms, as well as DNA repair mechanisms. In Escherichia coli, MutT hydrolyzes 8-oxo-dGTP and 8-oxo-GTP to monophosphate forms, thereby avoiding the misincorporation of such mutagenic nucleotides into DNA or mRNA during DNA replication or transcription (4,5). The loss of the mutT gene increases the spontaneous occurrence of A:T to C:G transversion to a level 1,000-fold greater than the wild type level and also further increases the transcriptional errors (5). Homologs of MutT were identified in human and rodent cells, and they suppressed the elevated spontaneous mutation rates of mutT mutants to an almost normal level, thus designated as MTH1 (MutT homolog-1) (6 -9). In addition to the sanitizing mechanism of 8-oxo-dGTP, the base excision repair enzymes for 8-oxo-G in DNA, such as MutM (FPG)/OGG1 and MutY/MUTYH (MYH), are conserved among various organisms from E. coli to mammals (10 -12), thus indicating that the genotoxicity of 8-oxo-G in DNA is universal throughout all living organisms.
Previously, we also showed that hMTH1, but not E. coli MutT, has an ability to efficiently hydrolyze oxidized dATP and ATP, such as 2-OH-dATP and 2-OH-ATP, as well as 8-oxo-dGTP (13,14), and thus, designated MTH1 as an oxidized purine nucleoside triphosphatase (15). 2-OH-dATP can be incorporated into the nascent strand of DNA opposite guanine as well as thymine, thus causing a G:C to T:A transversion mutation (2,16). Most parts of 2-hydroxyadenine (2-OH-A) in DNA are considered to be derived from a misincorporation of 2-OH-dATP but not from a direct oxidation of adenine in DNA (2).
Mice lacking the Mth1 gene exhibit an increased occurrence of spontaneous carcinogenesis especially in the liver, and to a lesser extent, in the lung and stomach (17), thus suggesting that the accumulation of oxidized purine nucleotides triggers such malignant transformation in vivo. Furthermore, the increased accumulation of 8-oxo-G in DNA along with the increased expression of hMTH1 is observed not only in human * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
To understand the biological significance of such oxidized purine nucleotides in mammals, it is essential to unveil the mechanism by which they cause cell dysfunction as well as mutagenesis that may thereby result in both degenerative diseases and cancers. In the present study, we demonstrated that hMTH1 protects cells from the cytotoxicity of H 2 O 2 by hydrolyzing oxidized purine nucleoside triphosphates.
Laser Scanning Fluorescence Microscopy-T5v and T5MTH1 cells were cultured on slides and incubated with 2 M MitoTracker Red CM-H 2 XRos (Molecular Probes) for 30 min prior to fixation. MTH1 protein was detected using anti-MTH1 in combination with Alexa Fluor 488-conjugated goat anti-rabbit IgG (Molecular Probes). Nuclei were counterstained with TOPRO-3 (Molecular Probes). Slides were observed under Eclipse TE300 (Nikon) equipped with the Radiance 2100 laser scanning fluorescence microscope system (Bio-Rad).
Morphological Examinations of Cells-MEFs cultured to ϳ70% confluence were exposed to medium containing 500 M H 2 O 2 for 24 h and then observed under a phase-contrast microscope. To detect the nuclear morphology, cells were additionally incubated with 10 M Hoechst 33342 (Molecular Probes) with or without 2 g/ml propidium iodide (PI) for 30 min and examined under a fluorescence microscope. At least 1200 cells were counted and classified as follows: (i) live cells with a normal nucleus and an organized structure stained only with Hoechst 33342, (ii) pyknotic live cells with a highly condensed nucleus stained only with Hoechst 33342, and (iii) pyknotic dead cells with a highly condensed nucleus stained both with Hoechst 33342 and PI. For electron microscopic observations, MEFs cultured to 80% confluence were exposed to medium containing 500 M H 2 O 2 for 4 or 24 h and processed as described previously (11,12).
Quantitation of 8-Oxo-2Ј-deoxyguanosine in Nuclear and Mitochondrial DNA-To quantify the nuclear content of 8-oxo-2Ј-deoxyguanosine (8-oxo-dG), MEFs cultured to ϳ80% confluence in a 14-cm dish were exposed to medium containing 500 M H 2 O 2 for 1, 4, 8, or 24 h and then harvested. The nuclear DNA was prepared and subjected to an HPLC-MS/MS analysis of 8-oxo-dG, according to the method described previously (28). The detection of 8-oxo-dG immunoreactivity in mitochondrial DNA was performed as follows. MEFs cultured on glass slides were exposed to medium containing 300 M H 2 O 2 for 1, 4, or 8 h at 37°C prior to fixation. The slides were incubated with 5 mg/ml RNase A for 2 h at 37°C and then with 50% ethanol containing 50 mM NaOH for 5 min at room temperature. 8-Oxo-dG in DNA was detected with an anti-8-oxo-dG mouse monoclonal antibody (MOG-020; Japan Institute for the Control of Aging) in combination with Alexa Fluor 488-conjugated goat anti-mouse IgG (Molecular Probes). The nuclei were counterstained with 2 g/ml PI. Digitized images for 8-oxo-dG immunoreactivity and nuclear PI staining were separately captured from identical fields using a laser scanning fluorescence microscope system, LSM-510 Meta (Carl Zeiss). To quantify the immunoreactivity of 8-oxo-dG in mitochondrial DNA, a previously described method (28) was used with some modifications. Briefly, images of 8-oxo-dG immunoreactivity were converted to 256-level gray-scale images, whereas those of nuclear PI staining were converted to binary monochromes by Adobe PhotoShop 5.0J (Adobe Systems). From the gray-scale images of 8-oxo-dG immunoreactivity, the signals residing in areas corresponding to nuclei represented by the monochrome PI images were deleted, and then the cytoplasmic signals for 8-oxo-dG were measured using NIH Image 1.61. The area for a single cell was manually defined, and an integrated pixel density for each cell was determined after setting a uniform threshold. At least 50 cells were analyzed, and an average value of the integrated pixel density per cell was calculated as the 8-oxo-dG index for each sample.
Assay of Cell Viability-After exposure to H 2 O 2 , the cell viability was measured by 2-(2-methoxy-4-nitrophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium, monosodium salt (WST-8) using Cell Counting Kit-8 (Wako Pure Chemical Industries). In brief, MEFs were cultured in 96-well plates to ϳ70% confluence and incubated with medium containing various concentrations of H 2 O 2 followed by incubation with WST-8. Absorbance of WST-8 formazan dye at 450 nm was measured, and relative cell viability was determined by comparing the ratio of absorbance for each experiment to that for the control experiments.
Colony Formation Assay-Exponentially growing MEFs were plated in 10-cm dishes at a density of 1000 or 5000 cells/dish. The cells attached to dishes by incubation for 6 h were incubated in a medium containing various concentrations of H 2 O 2 supplemented with 1 mM pyruvate for 24 h followed by incubation with fresh medium for 2 weeks to allow for colony formation. The colonies were fixed and stained in 25% ethanol containing 0.3% crystal violet. All experiments were performed in triplicate. The average value of plating efficiency was determined for each experiment, and the relative plating efficiency was determined as a ratio of the average value of the plating efficiency to that for the control experiment.

Establishment and Characterization of MTH1-null MEF Cell
Lines-MEFs were isolated from embryos obtained by the mating of Mth1ϩ/Ϫ mice, and spontaneously immortalized cell lines were established. The genotype of each cell line was determined by genomic PCR (Fig. 1, A and B). A Western blot analysis of mouse MTH1 (mMTH1) proteins in wild type and Mth1Ϫ/Ϫ MEFs demonstrated that there was no detectable mMTH1 protein in each Mth1Ϫ/Ϫ MEFs (Fig. 1C). Furthermore, none of the Mth1Ϫ/Ϫ MEFs were able to hydrolyze either 8-oxo-dGTP or 2-OH-dATP, thus demonstrating that the Mth1Ϫ/Ϫ MEFs were indeed MTH1-null mutants (data not shown). The expression levels of mMTH1 in wild type MEFs were considerably lower than those in the liver from wild type mice (Fig. 1C). In our experiments, no significant difference was observed between wild type and MTH1-null MEFs regarding their sensitivity to H 2 O 2 , sodium nitroprusside, and rotenone (data not shown).
Expression of Human MTH1 Protein in MTH1-null MEFs-The expression vectors for wild type and two mutant hMTH1 (W117Y and D119A) were introduced to the MTH1-null MEF cell lines (T2 and T5), and stable transfectants were thus established. By quantitative Western blotting, the expression level of hMTH1 protein was determined in each established cell line ( Fig. 2A). The hydrolyzing activities of 8-oxo-dGTP and 2-OH-dATP were also examined using whole cell extracts prepared from each cell line. MEFs expressing wild type hMTH1 exhibited substantial levels of an enzymatic capacity to hydrolyze both 8-oxo-dGTP and 2-OH-dATP, whereas those of W117Y or D119A mutant showed a selective loss of 8-oxo-dGTPase or 2-OH-dATPase activity, respectively (Fig. 2B). In MTH1-null MEFs that received pcDEB⌬ itself, no enzyme activity was detected in reactions with up to 20 g of whole cell extract (data not shown). The specific activities of hMTH1 protein in each cell line were determined based on the data obtained by quantitative Western blotting (Table I). The specific activities of wild type hMTH1 in whole cell extracts prepared from T2MTH1-1, -2, and T5MTH1 cells were close to those determined with purified hMTH1, 336.2 ϫ 10 2 units of 2-OH-dATPase/g of protein and 296.6 ϫ 10 2 units of 8-oxo-dGTPase/g of protein, respectively, and those of W117Y and D119A mutant hMTH1 also correlated with those described in our previous report (24).
The intracellular distribution of hMTH1 expressed in MTH1null MEFs examined by laser scanning fluorescence microscopy revealed that the signals for hMTH1 were detected mainly in the nuclei and cytoplasm, and to some extent, they merged with the signals of a mitochondrial fluorescent marker, Mito-Tracker (Fig. 3). The results thus indicate that hMTH1 also localizes in the mitochondria as we reported previously (23). From these results, it is concluded that hMTH1 maintains its function and subcellular localization in MTH1-null MEFs.
hMTH1 Suppresses H 2 O 2 -induced Cell Death and Ultrastructural Alteration of Mitochondria in MTH1-null MEFs-To investigate the cytoprotective function of MTH1 against oxidative stress, we examined the sensitivity of MTH1-null T5v cells to H 2 O 2 . After exposure to 500 M H 2 O 2 for 24 h, most cells died and thereafter became detached from the culture dish, and most of the attached cells showed a marked condensation of the cell bodies (Fig. 4A). An examination of the nuclear morphology using Hoechst 33342/PI staining by fluorescence microscopy showed most of the Hoechst 33342-stained nuclei to be highly condensed, and many of them were also stained with PI, thus representing dead cells (Fig. 4B). The dead cells were not accompanied with the classical morphological features of apoptosis, namely nuclear fragmentation, but their nuclei were much smaller than those of necrotic cells; we thus referred to this cell death as pyknosis (29). On the other hand, T5MTH1 cells expressing wild type hMTH1 exhibited little of these mor-phological features after exposure to H 2 O 2 . As summarized in Fig. 4C, about 50% of T5v cells died by pyknosis, and 25% of them were alive but pyknotic, whereas more than 95% of T5MTH1 cells exhibited a normal morphology after exposure to H 2 O 2 .
Next, the H 2 O 2 -induced ultrastructural alteration in each cell line was examined by transmission electron microscopy. In this experiment, the cells were prepared at a higher confluence than that shown in Fig. 4; thus, most cells remained on a culture dish even 24 h after exposure to H 2 O 2 . In T5v cells exposed to 500 M H 2 O 2 , nuclear apoptotic bodies were infrequently observed, whereas the mitochondrial cristae degenerated, and electron-dense deposits (EDDs) accumulated in such degenerating mitochondria as early as 4 h after exposure (Fig.  5A). Twenty-four h after exposure to H 2 O 2 , 20.1% of mitochondria were positive with EDDs in T5v cells, whereas those with EDDs were only 5.6% in T5MTH1 cells (Fig. 5B). No apparent enlargement of the mitochondria was observed in either T5v or T5MTH1 cells after H 2 O 2 exposure. As a result, we concluded that hMTH1 efficiently protects cells from the cytotoxicity of H 2 O 2 , together with the maintenance of the mitochondrial structure.
Cytoprotective Function of hMTH1 Is Dependent on Both 8-Oxo-dGTPase and 2-OH-dATPase Activities-The viabilities of T5v and T5MTH1 cells after exposure to various concentrations of H 2 O 2 are shown in Fig. 6A. In the absence of pyruvate in the medium, the median lethal dose (D 50 ) of H 2 O 2 for T5v cells was 98 M, whereas it was 224 M for T5MTH1 cells, thus indicating that the expression of hMTH1 increased the D 50 of H 2 O 2 by more than 2-fold. Even in the presence of 1 mM pyruvate that decomposes H 2 O 2 , the D 50 for T5MTH1 cells was 200 M higher than that for T5v cells, whereas that for both cell lines uniformly increased (Fig. 6B).
To examine whether the cytoprotective potential of hMTH1 depends on the expression levels of hMTH1, the D 50 of H 2 O 2 for MTH1-null T2v cells and for T2MTH1-1, T2MTH1-2 cells, which exhibited different expression levels of hMTH1, were determined (Fig. 6C). As a result, a strong correlation was observed between the D 50 values and the total activities of 8-oxo-dGTPase and 2-OH-dATPase in whole cell extracts from each cell line (Fig. 6D).
Next, the sensitivity to H 2 O 2 was determined for T5W117Y and T5D119A cells, which express hMTH1 mutant proteins, in the same manner in order to evaluate the cytoprotective effect of 8-oxo-dGTPase or 2-OH-dATPase, respectively. Their D 50 values (119 M for T5W117Y cells and 153 M for T5D119A cells) were apparently higher than those for MTH1-null T5v cells (88 M) but lower than those for T5MTH1 cells expressing wild type hMTH1 (224 M, Fig. 6E). Again, there was a very strong correlation between the D 50 values and the total activities of 8-oxo-dGTPase and 2-OH-dATPase in the whole cell extracts (Fig. 6F). As a result, hMTH1 is thus considered to confer protection against the cell dysfunction and death caused   (Fig. 7). Accordingly, we concluded that hMTH1 has a long term protective effect for cells exposed to H 2 O 2 .
Accumulation of 8-Oxo-dG in DNA of MTH1-null MEFs after Exposure to H 2 O 2 Is Suppressed by hMTH1-After exposure to H 2 O 2 , MTH1-null MEFs died by pyknosis accompanied with the morphological alteration of mitochondria, and they were efficiently suppressed by hMTH1. Thus, the accumulated misincorporation of oxidized nucleotides such as 8-oxo-dGTP or 2-OH-dATP into DNA is thought to be a major cause for such cell dysfunction and death. To prove this hypothesis, we examined the 8-oxo-dG content in nuclear and mitochondrial DNA of T5v and T5MTH1 cells after exposure to H 2 O 2 for various periods. The 8-oxo-dG content in nuclear DNA was measured by an HPLC-MS/MS analysis. In a steady state, 2.5-3.0 molecules of 8-oxo-dG/10 6 molecules of dG were detected in nuclear DNA prepared from each cell line, and the nuclear 8-oxo-dG levels in each cell line equally increased by ϳ2-fold for the first 4 h after exposure to H 2 O 2 . In MTH1-null T5v cells, the 8-oxo-dG level remained high up to 8 h after the exposure and thereafter decreased slightly within 24 h, whereas that in T5MTH1 cells was restored to a steady state level within 24 h (Fig. 8A). In T5D119A cells with 8-oxo-dGTPase but not 2-OH-dATPase, the nuclear 8-oxo-dG level was reduced to nearly a steady state level 8 h after exposure to H 2 O 2 , whereas that was still more than twice the steady state level at the same time in T5W117Y cells with 2-OH-dATPase but not 8-oxo-dGTPase (Fig. 8B).
Next, the intracellular distribution of 8-oxo-dG after expo-sure to H 2 O 2 was examined by immunofluorescence microscopy using an anti-8-oxo-dG monoclonal antibody (Fig. 8C). The signals for 8-oxo-dG as well as nuclear staining by PI were completely diminished by treatment with DNase and RNase but not with RNase alone (data not shown). We thus concluded that the immunoreactivity detected with RNase treatment represented 8-oxo-dG in DNA. Accordingly, the cytoplasmic 8-oxo-dG immunoreactivity, shown in Fig. 8D (Fig. 9A). Twenty-four h after exposure to H 2 O 2 , 3-AB only very slightly increased the D 50 levels for both cell lines to similar extents (Fig. 9B). As a result, 3-AB The results from one of three independent experiments are presented. A, the viabilities of T5v (circles) and T5MTH1 (squares) cells exposed to H 2 O 2 (0 -240 M) without pyruvate. B, the viabilities of T5v (circles) and T5MTH1 (squares) cells exposed to H 2 O 2 (0 -1200 M) in the medium supplemented with 1 mM pyruvate. C, the viabilities of T2v, T2MTH1 (numbers 1 and 2) cells exposed to H 2 O 2 (0 -270 M) in the absence of pyruvate. D, the correlation between D 50 of H 2 O 2 determined from the survival curves shown in C and total cellular activities for 2-OH-dATPase and 8-oxo-dGTPase (units/g of whole cell extract (WCE)) in T2v and T2MTH1 (numbers 1 and 2) cells. E, the viabilities of T5v, T5MTH1, T5W117Y, and T5D119A cells exposed to H 2 O 2 (0 -240 M). F, the correlation between D 50 of H 2 O 2 determined from the survival curves shown in E and the total cellular activities for 2-OH-dATPase and 8-oxo-dGTPase in each cell line.  showed little potential to suppress the delayed cell death caused by a lower concentration of H 2 O 2 , which was efficiently suppressed by hMTH1. Next, a general caspase inhibitor, Nbenzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone (Z-VADfmk, Wako Pure Chemical Industries), was applied to T5v and T5MTH1 cells 1 h prior to the exposure to H 2 O 2 . Again Z-VADfmk had no effect on their viabilities (Fig. 9C). The same concentration of Z-VAD-fmk (50 M) effectively suppressed the cell death of both cell lines induced by staurosporine (Fig. 9D), thus confirming that the caspases in these cell lines can be effectively inhibited by Z-VAD-fmk. Since there was no difference between the viabilities of the two cell lines exposed to staurosporine, hMTH1 was unable to suppress staurosporineinduced cell death (Fig. 9D). Moreover, even in the presence of both 3-AB and Z-VAD-fmk, we did not observe any protective effect for the cells after exposure to H 2 O 2 (data not shown). As a result, we concluded that the delayed cell death caused by H 2 O 2 , which was suppressed by hMTH1, was mediated by oxidized purine nucleoside triphosphates but not by either PARP or caspases. DISCUSSION The major conclusions of the present study are: (i) hMTH1 suppresses the accumulation of 8-oxo-G both in nuclear and mitochondrial DNA in cells exposed to oxidative stress; (ii) hMTH1 suppresses the delayed cell death, which is independent of PARP and caspases; (iii) and which may be partly attributed to the suppression of mitochondrial dysfunction; and (iv) hMTH1 suppresses the cell dysfunction and death induced not only by 8-oxo-dGTP but also by 2-OH-dATP.
MTH1 Suppresses Accumulation of 8-Oxo-G in the Cell DNA Exposed to Oxidative Stress-We herein demonstrated that an increased accumulation of 8-oxo-G in both nuclear and mitochondrial DNA occurred in MTH1-null MEFs exposed to H 2 O 2 , and the accumulation was efficiently suppressed by the expression of hMTH1. It is likely that the immediate increase of 8-oxo-dG is due to the direct oxidation of guanine in the cell DNA because it was equally observed in both MTH1-null MEFs and those expressing hMTH1. After 8 h of exposure to H 2 O 2 , the increased 8-oxo-dG levels in both nuclear and mitochondrial DNA of hMTH1-expressing MEFs but not MTH1-null MEFs significantly decreased, thus indicating that hydrolysis of 8-oxo-dGTP by hMTH1 resulted in the suppression of its incorporation into the cell DNA. A recent study showed that an increased expression of hMTH1 significantly reduced both the steady state level and the H 2 O 2 -induced level of 8-oxo-dG in nuclear DNA from Msh2Ϫ/Ϫ MEFs, thereby suppressing their increased spontaneous mutation rates (30). Our results together indicate that the oxidation of free nucleotides in the nucleotide pool is a major source of 8-oxo-dG accumulated in the cell DNA. This is the first demonstration that MTH1 suppresses the accumulation of oxidized bases not only in nuclear but also in mitochondrial DNA in vivo.
MTH1 Suppresses the Delayed Cell Death Induced Independently of PARP and Caspases by H 2 O 2 -Human MTH1 effectively suppressed H 2 O 2 -induced delayed cell death, which characteristically exhibited pyknosis, thus indicating that the accumulation of oxidized purine nucleotides including 8-oxo-dGTP and 2-OH-dATP resulted in pyknosis. Pyknosis is a common morphological feature during the process of cell death including apoptosis; however, both apoptosis and necrosis can be induced in various types of cells by exposure to H 2 O 2 (31)(32)(33). In general, a lower insult induces apoptosis, whereas a higher insult induces necrosis (29,33), as we observed in the present study.
The activation of PARP by several cytotoxic agents such as H 2 O 2 results in necrosis coupled with a depletion of NAD ϩ and ATP (34). In our experiments, 3-AB efficiently inhibited the cell death of both MTH1-null and hMTH1-expressing MEFs induced immediately after exposure to higher concentrations of H 2 O 2 but not the delayed cell death caused by lower concentrations of H 2 O 2 (Fig. 9, A and B).
Apoptosis is known to be dependent generally on various caspase-cascades, thus being inhibited by such caspase inhibitors as Z-VAD-fmk (35). However, Z-VAD-fmk did not suppress H 2 O 2 -induced delayed cell death of MTH1-null MEFs at all (Fig. 9C). Recent studies have revealed several pathways of caspase-independent cell death (35,36), and our study demonstrated for the first time that MTH1 has an inhibitory effect on one of such caspase-independent pathways of cell death. Further study is required to identify the pathway of cell death that can be suppressed by MTH1.
The inhibition of cell death by such inhibitors as 3-AB or Z-VAD-fmk is not permanent because they cannot remove the direct cause of cell death. Human MTH1-expressing cells acquired an improved colony-forming ability after exposure to H 2 O 2 (Fig. 7), thus indicating that oxidized purine nucleotides are the direct cause of the cell death, and oxidized purine nucleotides are efficiently sanitized by MTH1, thereby ensuring the long term survival of the cells exposed to oxidative stress.
MTH1 Suppresses Mitochondrial Dysfunction Caused by Oxidative Stress-An ultrastructural analysis by electron microscopy revealed that H 2 O 2 induced an accumulation of small round EDDs in the mitochondria of MTH1-null MEFs, which was suppressed by hMTH1. The morphological alterations of the mitochondria have been scrutinized in patients with mitochondrial encephalomyopathy (37)(38)(39). In muscle biopsy specimens from patients with such mitochondrial diseases, abnormalities in mitochondrial ultrastrucuture are reported as inclusion bodies or paracrystalline inclusions, and the abnormalities are associated with mitochondrial dysfunction such as a decline in cytochrome oxidase activity. Although the precise pathological basis for these abnormal structures is not clear, a similar morphological alteration has been observed in rat brain neurons exposed to severe ischemia/reperfusion injury (40). These observations strongly suggest that the accumulation of mitochondrial EDDs in MTH1-null MEFs caused by exposure to H 2 O 2 is associated with mitochondrial dysfunction.
Biological Significance of 8-Oxo-G Accumulation in DNA-Our results indicate that an accumulation of 8-oxo-G in nuclear and mitochondrial DNA of MTH1-null MEFs after exposure to H 2 O 2 is followed by the morphological alteration of mitochondria and then the delayed cell death. In Ogg1 knockout mice, an increased level of 8-oxo-G was more evident in DNA from the mitochondria than in that from the nuclei (41). HeLa cells introduced with hOGG1 tagged with a mitochondrial targeting sequence showed an increased 8-oxo-G DNA glycosylase activity in mitochondrial extracts and also an increased survival under oxidative stress (42), thus supporting the idea that 8-oxo-G accumulation in mitochondrial DNA contributes, at least in part, to the induction of cell death in MTH1-null MEFs exposed to H 2 O 2 .
We recently reported that Ogg1 knockout mice exhibited a significant increase in lung carcinogenesis with an accumulation of 8-oxo-G in nuclear DNA, whereas no lung cancer was found in Ogg1/Mth1-double knockout mice, in which a slightly higher level of 8-oxo-G was found to be accumulated in nuclear DNA than in Ogg1 knockout mice (22). Based on these results together with the present findings, we propose that an OGG1 deficiency mainly causes mitochondrial dysfunction due to an accumulation of 8-oxo-G in mitochondrial DNA, and the concomitant loss of MTH1 further increases the accumulation of 8-oxo-G and other oxidized purines including 2-OH-A in nuclear and mitochondrial DNA, thus enhancing cell death during initiation or early progression of lung cancer, which may be exposed to a substantial level of oxidative stress due to an excessive cell metabolism as compared with normal tissue.
Biological Significance of Oxidized Purine Nucleotides Other than 8-Oxo-dGTP-There is little data indicating the biological implications of oxidized purine nucleotides other than 8-oxo-dGTP, including 2-OH-dATP, which is most effectively hydrolyzed by hMTH1 (24), in cancer or neurodegenerative diseases, because of the difficulty of their detection. Recently, inherited mutations in the MUTYH gene have been reported in a family affected with multiple colorectal adenomas without an inherited mutation of the adenomatous polyposis coli (APC) gene (43). Tumors from affected siblings frequently contained somatic inactivating G:C to T:A transversion mutations of the APC gene (44,45). The APC gene is thus considered to be a possible target of mutation due to oxidative damage. We reported previously that 2-OH-A paired with guanine in DNA is one of the substrates for human MUTYH protein (12). Furthermore, we found that Mutyh knockout mice are highly susceptible to both spontaneous adenocarcinoma and G:C to T:A transversion mutation of a reporter gene in the small intestine. 2 No increased spontaneous tumorigenesis was found in the small intestine of Ogg1 knockout mice (22). These observations suggest the biological significance of 2-OH-A in DNA that cannot be repaired by OGG1, and also of the precursor, 2-OH-dATP. We showed that mutant hMTH1(D119A) lacking 2-OH-dATPase activity only partially suppressed the delayed cell death of MTH1-null MEFs in comparison with the wild type, thus indicating that 2-OH-dATP is certainly cytotoxic and biologically significant.
Since it has been widely accepted that carcinogenesis and cell death, both of which are contrary to each other in terms of their biological consequences, are tightly associated with DNA damage (1), it is most likely that oxidized purine nucleotides are misincorporated into DNA, thus causing cell dysfunction including cell death. However, hMTH1 efficiently hydrolyzes 2-OH-ATP, 8-oxo-ATP, and to a lesser extent, 8-oxo-GTP. As a result, cell dysfunction may also be caused by their incorporation into RNA. Furthermore, an increasing concentration of 8-oxo-dGTP itself has been reported to reduce the amounts of newly synthesized DNA in a cell-free DNA replication system derived from Xenopus egg lysates (46), thus indicating the possibility that free forms of oxidized purine nucleotides themselves exert a certain degree of cytotoxicity. To further elucidate the biological significance of these oxidized nucleotides in vivo, we are now developing transgenic mice expressing wild type or mutant hMTH1 proteins with a selective loss of substrate specificity. Extended analyses on hMTH1 using these materials will hopefully shed some light on the molecular mechanism of carcinogenesis and neurodegeneration.