Cancer Research Grants Advances in Breast Cancer Research
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
Cancer Research 66, 9845, October 15, 2006. doi: 10.1158/0008-5472.CAN-06-0732
© 2006 American Association for Cancer Research
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplementary Data
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Lee, D.-H.
Right arrow Articles by Chu, F.-F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Lee, D.-H.
Right arrow Articles by Chu, F.-F.

Molecular Biology, Pathobiology, and Genetics

Mutation Accumulation in the Intestine and Colon of Mice Deficient in Two Intracellular Glutathione Peroxidases

Dong-Hyun Lee1, R. Steven Esworthy2, Christy Chu2, Gerd P. Pfeifer1 and Fong-Fong Chu2

1 Department of Biology and 2 Department of Radiation Biology, City of Hope Cancer Center, Duarte, California

Requests for reprints: Gerd P. Pfeifer, Beckman Research Institute of the City of Hope, 1450 East Duarte Road, Duarte, CA 91010. Phone: 626-301-8853; Fax: 626-358-7704; E-mail: gpfeifer{at}coh.org.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Mice deficient in two glutathione peroxidases (GPX), Gpx1 and Gpx2, [Gpx1/2-double knockout (DKO) mice] are prone to ileocolitis on a mixed C57BL/6 and 129S1/SvJ (B6.129) genetic background. We reported previously that ~25% of B6.129 Gpx1/2-DKO mice develop ileocolonic tumors by 6 to 9 months of age, when their non-DKO littermates [having at least one wild-type (WT) Gpx1 or Gpx2 allele] rarely have inflammation and none have tumors. Because genetic background affects tumor susceptibility, we have generated a B6 Gpx1/2-DKO colony and discovered that these mice have fewer inflammatory cells, milder ileocolitis, and low mortality, and only 2.5% of B6 mice developed tumors. The mutant frequency of a cII reporter gene was about 2- to 3-fold higher in 28-day-old Gpx1/2-DKO and 4-fold higher in 8-month-old Gpx1/2-DKO ileal mucosa than in controls in both genetic backgrounds. In contrast, mutant frequencies in the unaffected B6 liver were not significantly different between WT and Gpx1/2-DKO mice. The mutant frequency of 8-month-old B6.129 Gpx1/2-DKO ileum was 38.94 ± 15.5–5, which was not significantly higher than the age-matched B6 ileum, 25.54 ± 10.33–5. The mutation spectra analysis has shown that B6 Gpx1/2-DKO ileum had a 3-fold increase in small nucleotide deletions at mononucleotide repeats over control B6, which are a signature mutation associated with oxidative stress. Unexpectedly, B6 Gpx1/2-DKO mice had fewer C to T transitions at CpG dinucleotides than the WT B6 (18.0% versus 40.1%; P < 0.001). Our results suggest that inflammation drives gene mutations, which leads to neoplastic transformation of intestinal epithelium in the B6.129 Gpx1/2-DKO mice but rarely in the B6 Gpx1/2-DKO mice. (Cancer Res 2006; 66(20): 9845-51)


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Oxidative stress occurs when the generation of reactive oxygen species (ROS) in a system exceeds the antioxidant capacity to neutralize and eliminate them (1–4). This imbalance can result from a lack of available antioxidants or from an overabundance of ROS from environmental or internal sources. An excess of ROS can damage cellular macromolecules, such as lipids, proteins, and nucleic acids. For example, ROS can directly interact with nucleic acids, resulting in a variety of modifications, including base damage, sugar damage, deletions, cross-linked lesions, and DNA strand breaks (1, 3, 5).

Inflammatory diseases are often associated with significant oxidative stress that can damage cells and tissues. A link between inflammation and cancer is well established (4, 6, 7). An increased incidence of cancer is seen in patients with chronic gastritis, chronic pancreatitis, and inflammatory bowel disease (IBD; refs. 4, 7–9). IBD includes Crohn's disease (CD) and ulcerative colitis (UC); each affects approximately one in a thousand people in Western countries. Although CD and UC may have different genetic origins and affect different areas of the gastrointestinal tract, both diseases result in an increased risk of cancer compared with the age-matched general population (10, 11). Clearly, chronic inflammation contributes to carcinogenesis in the gastrointestinal tract; however, the mechanisms are not understood.

Inflammation is caused by the recruitment of inflammatory cells, including neutrophils, monocytes, and eosinophils (7), which release ROS. Chronic inflammation can be a tumor promoter, which may affect signal transduction mechanisms, influence cell proliferation, and modulate apoptosis. However, it is also possible that inflammation acts as a tumor initiator by inducing DNA damage and gene mutations in the affected tissue. Such a proposal, in which inflammation, DNA damage, increased mutations, and tumorigenesis are directly linked in consecutive steps, has been put forward (12).

Selenium-dependent glutathione peroxidases (GPX) represent a major selenoprotein-containing gene family in mammals. The GPXs include four selenium-dependent hydroperoxide-reducing isozymes: (a) the ubiquitous GPX1, (b) the epithelium-specific GPX (GPX2), (c) the secreted plasma GPX (GPX-P), and (d) the monomeric phospholipid hydroperoxide GPX (PHGPX), which are encoded by the Gpx1, Gpx2, Gpx3, and Gpx4 genes, respectively (13). Among these GPXs, GPX1 and GPX2 are the major H2O2-reducing GPX activities in the gastrointestinal epithelium. The GPX1 and GPX2 isozymes have very similar properties, such as substrate specificity and cytosolic localization. They both reduce H2O2 and fatty acid hydroperoxides very efficiently but reduce lipid hydroperoxides poorly. Unlike the ubiquitous GPX1, GPX2 is expressed mainly in epithelium, most highly in the gastrointestinal epithelium. We found that, although homozygous mice deficient in either the wild-type (WT) Gpx1 or Gpx2 alleles appeared to be normal under standard housing conditions, homozygous Gpx1 and Gpx2 double knockout mice (Gpx1/2-DKO), with combined disruption of both Gpx1 and Gpx2 genes, are highly susceptible to ileocolitis beginning around weaning (14). Similar to other mouse IBD models, when these mice are derived into germ-free conditions, they are disease-free (15). With long-term follow-up, the tumor incidence in Gpx1/2-DKO mice raised conventionally is ~25% in mice harboring Helicobacter hepaticus (an enterohepatic Helicobacter species that can cause colitis in immunodeficient mice; ref. 16) on a mixed B6.129 genetic background (17). Oxidative stress associated with cellular damage is thought to play a key role in the pathogenesis of the colitis itself (8, 9). Lack of the antioxidant GPX system in the intestinal mucosa may set up a continuous cycle of ROS and inflammation; inflammation and production of ROS by inflammatory cells eventually leads to cancer.

Because inflammatory responses produce DNA-damaging oxidative species and chronic inflammation in the gastrointestinal tract can cause cancer, we hypothesized that inflammation drives gene mutations, which can lead to carcinogenesis. To test this hypothesis in a mouse model of IBD, we compared mutation frequencies and mutation spectra of the cII reporter gene in the affected intestinal and unaffected liver tissues of Gpx1/2-DKO and control mice on both B6.129 and B6 genetic background. Because GPX1 is up-regulated by p53 activation (18) and GPX2 by p63 (19), a homologue of p53, to inhibit oxidative stress–induced apoptosis, studying the effect of GPXs in prevention of gene mutations may provide an insight to the mechanism of action of these antioxidant enzymes in IBD-associated tumorigenesis.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Animal breeding. We have established the B6.129 Gpx1/2-DKO mouse colony as described previously (14). Breeders were maintained on diet 5020 (Laboratory Rodent Diet, Purina Mills, Inc., Richmond, IN) with 9% fat. At weaning, pups were placed on diet 5001 with 5% fat until recruitment into breeding. To generate a B6 Gpx1/2-DKO colony, the B6.129 mice were back-crossed to B6 for seven generations. B6 Gpx1/2-DKO males were bred to homozygous B6 Big Blue females (Stratagene, La Jolla, CA) carrying a tandem array of {lambda} genomes ({lambda}LIZ) containing mutational reporter genes, and subsequent breeding of heterozygous female progeny to homozygous B6 Gpx1/2-DKO males was done to incorporate the {lambda}LIZ shuttle vector transgene into the B6 Gpx1/2-DKO colony. We have maintained a WT B6 {lambda}LIZ Big Blue line as controls. We then generated B6.129 {lambda}LIZ-transgenic (Tg) mice by crossing B6 Gpx1/2-DKO {lambda}LIZ-Tg males with 129S1/Sv females (The Jackson Laboratory; Bar Harbor, ME). Heterozygous B6.129 {lambda}LIZ-Tg progeny were back-crossed to the original B6.129 Gpx1/2-DKO colony to establish a B6.129 {lambda}LIZ-Tg colony. B6.129 {lambda}LIZ-Tg mice carrying at least one WT Gpx1 or Gpx2 allele, non-DKO {lambda}LIZ-Tg, were used as controls for B6.129 Gpx1/2-DKO mice. We observed similar pattern and severity of disease in Gpx1/2-DKO mice of the same genetic background regardless of whether the mice were carrying the {lambda}LIZ gene or not. All colonies harbor H. hepaticus, which can cause colitis but not hepatitis in B6 and B6.129 mice (20).3 Animal care and treatment was approved by the City of Hope Research Animal Care Committee (Duarte, CA).

Pathology and tumor studies. We assessed animal health by routine weighing and visual inspection for diarrhea (wet tail), perianal alopecia, and perianal ulceration (14). Terminal ileum and distal colon sections flanking the tissues for DNA isolation for cII analysis were processed for histology analysis. Tissues were fixed overnight in phosphate-buffered formalin and then processed for paraffin embedment and thin sectioning. Sections were routinely stained with H&E for general pathology surveys. Immunohistochemistry was done with rabbit polyclonal anti-myeloperoxidase antibodies (x100, Lab Vision, Fremont, CA) and biotinylated preabsorbed goat anti-rabbit IgG (Abcam, Inc., Cambridge, MA) to stain neutrophils and monocytes (17), rat monoclonal anti-macrophage antibody (20 µg/mL F4/80, CI:A3-1, Abcam) and biotinylated goat anti-rat IgG (Vector Laboratories, Inc., Burlingame, CA), and a bromodeoxyuridine (BrdUrd) immunohistochemistry system (Calbiochem, San Diego, CA) to detect proliferating cells (17) and counterstained with hematoxylin. Pathology and inflammation were scored by a 14-point system as described previously (21).

Tumors were generally sufficiently advanced at 5 to 9 months of age to be detected by naked eye after dissecting up the intestine longitudinally (17). All putative tumorous lesions were verified and graded by consultation with staff pathologists on H&E-stained sections. When harvesting DNA for cII mutation analysis, we excluded mice with tumors.

DNA isolation. Epithelial cells were isolated from the distal 10 to 12 cm of the ileum (lower 60% of small intestine, corresponding to the diseased segment in Gpx1/2-DKO mice at the height of pathology and distribution of tumors) and total colon (less ~0.5 cm flanking segments taken for histology) by the everted sac method as described previously (22) to recover the villus and crypt compartments. The cells were rinsed in PBS thrice and frozen at –80°C. Genomic DNA was isolated using a standard phenol and chloroform extraction and ethanol precipitation protocol (23). Liver DNA was isolated using a NaCl method. Briefly, the excised tissue was homogenized in 4 mL cold PBS in a 7 mL Wheaton Dounce tissue grinder (Millville, NJ), washed twice with PBS, and lysed with 4 mL of a solution containing 0.5 mol/L Tris-HCl (pH 8.0), 20 mmol/L EDTA, 10 mmol/L NaCl, 1% SDS, and 0.5 mg/mL proteinase K at 37°C overnight. Subsequently, 2 mL saturated NaCl (~6 mol/L) was added to each sample, and the samples were incubated at 56°C for 10 minutes. After centrifugation at 5,000 x g for 30 minutes, the supernatant containing the DNA was mixed with 2 volumes of cold 100% ethanol, and the DNA was spooled by gently inverting the mix. The DNA was washed thoroughly with 70% ethanol, air dried, and subsequently dissolved in TE buffer (pH 8.0), and kept at –80°C until further analysis.

cII mutant frequency analysis. The cII mutant frequency was examined by using the select-cII mutation detection system for Big Blue rodents as described previously (24). The assay is based on the ability of the phage to multiply either lytically or lysogenically in Escherichia coli host cells. Briefly, we recovered the LIZ shuttle vectors from mouse genomic DNA (5 µg) and packaged them into viable phage particles using the Transpack packaging extract (Stratagene). The phage particles were then preadsorbed to G1250 E. coli, and the bacteria were plated on TB1 agar plates. The plates were incubated for 48 hours at 24°C (selective conditions) or overnight at 37°C (nonselective conditions). The cII mutant frequency was expressed as the ratio of the number of plaques formed on plates incubated under selective conditions to the number of plaques formed under nonselective conditions. We screened a minimum of 3 x 105 rescued phages for each experimental condition.

cII mutation spectrum analysis. Plaques containing putative mutants of cII were verified after being replated and incubated under selective conditions. The verified plaques were amplified by PCR using select-cII sequencing primers (Stratagene) and purified with QIAquick PCR purification kits (Qiagen, Valencia, CA). The PCR products were sequenced by using a BigDye terminator cycle sequencing kit on an ABI automated DNA sequencer. At least 100 mutants were analyzed for each of five B6 Gpx1/2-DKO and five B6 WT 8-month-old ilea.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
We have reported previously that B6.129 Gpx1/2-DKO mice are prone to ileocolitis under conventional housing conditions, and their ileocolitis occurs around weaning (3-4 weeks of age). The inflammation is most severe in the distal ileum, and by 6 to 9 months of age, 25% of these mice have gross tumors mostly in the ileum and a few in the colon (14, 17). B6.129 non-DKO mice with at least one WT allele of Gpx1 or Gpx2 rarely had inflammation and none had tumors. Genetic background has a profound effect on both inflammation (16) and tumor susceptibility (25), with B6 being resistant to inflammation but sensitive to chemical-induced colon carcinogenesis. We back-crossed our original colony of B6.129 Gpx1/2-DKO to B6 mice and investigated whether B6 Gpx1/2-DKO mice are also more resistant to inflammation and inflammation-associated cancer than B6.129 Gpx1/2-DKO mice.

As expected, under the same housing conditions, which contain H. hepaticus, these B6 Gpx1/2-DKO mice are more resistant to inflammation than B6.129 Gpx1/2-DKO mice, which were evaluated, concurrently. Most importantly, no B6 Gpx1/2-DKO mice were too ill and required euthanasia before 50 days of age, when 15% to 20% of B6.129 Gpx1/2-DKO mice did. B6 Gpx1/2-DKO mice began to have ileocolitis later, with peak pathology at 6 to 8 weeks of age (2-4 weeks later than B6.129 mice), and the pathology was milder than that in B6.129 Gpx1/2-DKO mice. Figure 1 shows the typical ileal pathology of a 28-day-old B6.129 and B6 Gpx1/2-DKO mouse after staining with anti-myeloperoxidase antibodies. At this young age, there were few mature macrophages detectable by immunohistochemistry with anti-F4/80 antibodies in either genetic background (data not shown). Although the 28-day-old B6 Gpx1/2-DKO ileum still had a high level of apoptotic cells, crypt distortion, and disorganized gland architecture, it rarely had epithelium erosion or crypt abscesses. At this young age, B6.129 Gpx1/2-DKO ileum had a much higher number of inflammatory cells than age-matched B6 (Figs. 1 and 2 ). We also compared the levels of macrophages, monocytes, and neutrophils in 8-month-old Gpx1/2-DKO ileum and found that B6.129 Gpx1/2-DKO mice had significantly higher numbers of anti-F4/80-stained macrophage cells than B6, although they had similarly high levels of anti-myeloperoxidase-stained monocytes and a low number of neutrophils (Figs. 1 and 2). This lower level of inflammation also leads to lower pathology/inflammation scores in the B6 Gpx1/2-DKO ileum compared with B6.129 between 21 to 120 days of age as shown in Fig. 3A .


Figure 1
View larger version (108K):
[in this window]
[in a new window]
  		Figure 1. Immunohistochemistry of Gpx1/2-DKO mouse ileum on B6.129 and B6 genetic backgrounds. A, has B6.129 ileum. B, has B6 ileum. A and B, left, 28-day-old mice stained with anti-myeloperoxidase ({alpha}-MPO) antibodies. Although, both ilea have distorted gland architecture, only B6.129 mice have evident crypt abscess as shown by the exfoliated neutrophils (arrows). Middle, 8-month-old ileum stained with anti-myeloperoxidase antibodies. B6.129 mice have neutrophils located in the abscessed crypt and monocytes located in submucosa (arrows), and B6 mice had fewer monocytes and neutrophils. Right, 8-month-old ilea stained with anti-macrophage ({alpha}-M{phi}) antibodies. Only a few macrophage cells are detected in the submucosa of older ileum of both B6 and B6.129 mice. Original magnification, x100.

 

Figure 2
View larger version (14K):
[in this window]
[in a new window]
  		Figure 2. Inflammatory cells in the ilea of B6.129 Gpx1/2-DKO mice and B6 Gpx1/2-DKO mice. Columns, mean number of inflammatory cells in 28-day-old (28 d) and 8-month-old (8 mo) ileum of B6.129 control, B6.129, and B6 Gpx1/2-DKO mice; bars, SD. WT B6 mice were not included due to scarcity of inflammatory cells. The macrophages were detected with anti-F4/80 antibodies, when anti-myeloperoxidase antibodies detected monocytes in submucosal and neutrophils in the crypt abscess. Three to eight ilea were counted in each group. *, significant difference from others in the same type of cells, except that monocytes in 8-month-old B6 and B6.129 are not different from each other; P < 0.05, t test.

 

Figure 3
View larger version (16K):
[in this window]
[in a new window]
  		Figure 3. Ileal inflammation scores and number of proliferating epithelial cells in B6.129 and B6 mice. A, ileal pathology or inflammation scores analyzed at different ages in B6 Gpx1/2-DKO (bullet), B6.129 Gpx1/2-DKO ({lozenge}), and B6.129 non-DKO ({Delta}) mice. B6 non-DKO ileum has even lower scores than B6.129 non-DKO mice and thus is not included. The criteria for the scoring include inflammation foci, crypt distortion, abundance of apoptotic cells, neutrophils, monocytes, and lymphocytes, and degranulation of Paneth cells or mucin depletion in goblet cells as described previously (21). The differences at 21 and 41 days of age are statistically significant (P ≤ 0.007, t test). B, number of proliferating cells labeled with BrdUrd, detected by immunohistochemistry. Two hours before euthanasia, mice were injected with BrdUrd, which was detected with an anti-BrdUrd antibody immunohistochemically as we have described previously (17). Each group has 5 to 7 mice at 50 days of age. The B6 and B6.129 Gpx1/2-DKO ileum has 2.5- and 2.4-fold significantly higher number of proliferating cells than their controls. P < 0.05, t test.

 
Because the inflammation scores quantify the morphologic changes, including crypt distortion, which is partially attributed to hyperproliferation, we also analyzed cell proliferation in the crypts of ileal and colonic epithelium. As shown in Fig. 3B, B6 and B6.129 Gpx1/2-DKO ileal epithelia have 2.5- and 2.4-fold, respectively, higher numbers of proliferating cells than their age-matched non-DKO controls. Similarly, B6 and B6.129 Gpx1/2-DKO colonic epithelia have 1.9-fold higher number of proliferating cells than their non-DKO controls (data not shown). Thus, the higher pathology/inflammation scores in B6.129 over B6 Gpx1/2-KO mice are not contributed by subinflammatory responses, such as apoptosis (data not shown) and proliferation; rather, the higher scores in B6.129 Gpx1/2-DKO mice are contributed by a higher number of inflammatory cells and inflammation foci. This may be a major reason we only detected one tenth of the tumor frequency originally found in B6.129 Gpx1/2-DKO mice (25%) in the B6 Gpx1/2-DKO mice (2.5%) analyzed between 6-9 months of age.

Because inflammation can enhance tumor susceptibility, we hypothesized that inflammation causes gene mutations, which lead to cancer. To test this hypothesis, we bred B6 Gpx1/2-DKO mice with B6 {lambda}LIZ-Tg mice carrying a tandem array of {lambda} genomes ({lambda}LIZ) containing mutational reporter genes. These mice contain the transcriptionally silent cII reporter gene, which we used as a target for mutation analysis. Five to nine each of B6 WT and B6 Gpx1/2-DKO mice carrying {lambda}LIZ-transgenes were used to study cII mutation frequencies and mutation spectra. The mutation analysis was done on tissue samples free of macroscopic tumors. We have isolated mucosal epithelial cells from intestine for this analysis to minimize the contribution by inflammatory cells (Supplementary Fig. S1).

The 8-month-old mice from both non-DKO control and Gpx1/2-DKO groups of either B6 or B6.129 mice had elevated mutant frequencies over the 28-day-old mice (Table 1 ). These age-accumulated differences were about 3- to 4-fold for the colon and ileum samples (P < 0.05, t test) but only ~1.7-fold for the liver samples (P = 0.06). Comparing the 28-day-old mice, mutant frequencies in the ileum or the colon of B6 Gpx1/2-DKO mice were about 2- to 3-fold higher than those in B6 WT mice; but in the liver, B6 Gpx1/2-DKO mice had only a slight (1.5-fold) increase in mutant frequency. In the 8-month-old mice, the mutant frequencies in B6 Gpx1/2-DKO ileum were over 4-fold higher than B6 WT mice (P < 0.04, t test). In the mildly affected B6 Gpx1/2-DKO colon, we observed a ~2-fold increase of mutant frequencies than in B6 WT mice. As expected, no significant differences in mutant frequencies were observed in the uninvolved B6 Gpx1/2-DKO liver and in B6 WT mice.


View this table:
[in this window]
[in a new window]
  		Table 1. cII mutant frequencies in Gpx1/2-DKO and control mice

 
To test whether gene mutations are correlated with cancer incidence, we have also analyzed cII mutations in B6.129 Gpx1/2-DKO and non-DKO {lambda}LIZ-Tg mouse ileum. We used the age-matched B6.129 non-DKO mice as controls because they were close relatives of the B6.129 Gpx1/2-DKO mice and rarely had inflammation or pathology and did not have tumors. Three to four each of B6.129 non-DKO and Gpx1/2-DKO ilea were analyzed for the mutant frequencies. As shown in Table 1, at 8 months of age, the susceptible B6.129 Gpx1/2-DKO mice had similar, ~4-fold higher, mutant frequencies as the B6 Gpx1/2-DKO mice compared with their controls. Although the mutant frequency, 38.9 x 10–5, in the B6.129 Gpx1/2-DKO ileum was higher than the age-matched B6 ileum (25.54 x 10–5), the difference was not statistically significant (P = 0.09, t test).

To compare mutational spectra, we sequenced DNA isolated from the cII mutants obtained from 8-month-old B6 WT and B6 Gpx1/2-DKO ileum. From each of five mice per group, we sequenced >100 mutants and confirmed a mutated cII gene in over 96% of the plaques. Because the cII transgene in the Big Blue rodents is not transcribed (26), there is no "strand-dependent mutagenesis," a phenomenon caused by transcription-coupled DNA repair in mammalian endogenous genes (27), in this system. Therefore, we have combined the strand mirror counterparts of all transitions (e.g., G:C>A:T and C:G>T:A) and transversions (e.g., G:C>T:A and C:G>A:T) when comparing the mutational spectra. As shown in Fig. 4 , although the mutational spectrum of the B6 Gpx1/2-DKO mice was generally similar to that of the B6 WT mice, there were some important differences. Although the absolute frequency of all types of mutations were increased in the B6 Gpx1/2-DKO mice (Fig. 4A), their relative contribution was different (Fig. 4B). For example, transition mutations at CpG sites comprised 41% (median, 31-47%) of all mutations in the WT ileum but only 19% (median, 15-30%) in B6 Gpx1/2-DKO ileum (P < 0.001, t test). Surprisingly, the relative contribution of G:T transversions was decreased in the B6 Gpx1/2-DKO mice, although their absolute frequency was increased. The relative frequency of small deletions was significantly increased in the B6 Gpx1/2-DKO mice compared with that in the WT mice [~3-fold increase, from 6.6% (median, 3.3-10.3%) to 15% (median, 6.2-35.2%) of all mutations; P = 0.04, t test]. Additionally, mutations at A:T bp and G:C transversions were also more common in the B6 Gpx1/2-DKO mice.


Figure 4
View larger version (21K):
[in this window]
[in a new window]
  		Figure 4. Mutational spectra of the cII transgene in B6 Gpx1/2-DKO and WT mice. A, absolute frequencies of each type of mutation. The cII mutants were sequenced to evaluate mutational spectra. The mutation spectra were compiled from at least 100 mutant sequences from each mouse ileum. Five each of 8-month-old B6 WT and Gpx1/2-DKO mice were analyzed. The strand mirror counterparts of all transitions (e.g., G:C to A:T and C:G to T:A) and transversions (e.g., G:C to T:A and C:G to A:T) were combined. The mutant frequency for each type of mutation was determined by multiplying the percentage of each type of mutation with the average mutant frequencies obtained from the WT (white columns) and Gpx1/2-DKO (black columns) group. Gray columns, mutant frequencies occurring at CpG sites. B, relative frequencies of each type of mutation. Combined data for each of the five WT and five Gpx1/2-DKO mice (white and black columns, respectively). Ins, insertion; Del, deletion.

 
The complete mutational spectra of individual ileum in five each of B6 WT and B6 Gpx1/2-DKO mice are shown in the Appendix (Supplementary Figs. S2 and S3). Mutational hotspots, defined as the occurrence of more than five mutations at one particular nucleotide position, occurred in one or more B6 WT and B6 Gpx1/2-DKO mouse. Many of the hotspots in B6 Gpx1/2-DKO (13 of 28) and B6 WT (5 of 31) mice are unique to a single mouse. The higher number of unique hotspots in B6 Gpx1/2-DKO mice may come from clonal expansion of cells carrying an initial mutation through rapid cell proliferation and thus better fixation. Nonunique hotspots are likely derived from an intrinsic mutability of a particular DNA sequence. In the WT mice, the hotspots were largely G:C>A:T transitions at CpG dinucleotide sequences. Strikingly, 23 of the 31 hotspots in WT mice but only 9 of the 28 hotspots in the Gpx1/2-DKO mice were of this type. The B6 Gpx1/2-DKO mice had other types of dominant mutational events, including hotspots of G:C>C:G transversions (at positions 31, 86, 211, and 224), G:C>T:A transversions (at positions 106, 167, and 190), and a hotspot of T:A>C:G transitions (at position 53). B6 Gpx1/2-DKO mice had frequent occurrence of small deletions at two very large mutational hotspots; one at a mononucleotide G repeat, between positions 179 to 184, and the other at a mononucleotide A repeat, between nucleotides 241 to 246, the latter occurring in every DKO mouse.

We did not analyze mutation spectra of B6.129 mice due to their lack of uniform genetic background. For example, 129 mice are defective in the error-prone DNA polymerase iota, which may be involved in replication, bypassing some of the DNA lesions produced under oxidative stress.3 Analyzing mice with mixed genetic background would have complicated data interpretation.


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
ROS play a significant role in the pathogenesis of diseases of the gastrointestinal tract, including IBD (4, 7, 9). B6.129 Gpx1/2-DKO mice, as an IBD model, are highly susceptible to ileocolitis beginning around weaning (13–15, 17). GPX enzymes in the gastrointestinal tract play a critical protective role in the detoxification of ROS produced during inflammation in response to bacterial colonization (13). We considered the possibility that inflammation and increased levels of ROS in B6.129 Gpx1/2-DKO mice may produce an elevated load of mutations, which can increase the risk for malignant transformation. Thus far, there are only few reports showing a direct relationship between inflammation and DNA mutation load in specific tissues. UC patients have an increased load of p53 mutations in their affected tissues within areas of active inflammation (28). Oxygen radical overload diseases, such as Haemochromatosis and Wilson disease, are characterized by accumulation of iron and copper, respectively, in the liver. These conditions produce enhanced oxidative stress, increase p53 mutations in premalignant tissue, and increase the risk for liver cancer (29).

In this study, using a mouse model for IBD, we have found significantly higher mutant frequencies in the affected ileal and colonic epithelium, but not in the unaffected liver, of B6 and B6.129 Gpx1/2-DKO mice compared with their control mice. It is unlikely that the increase in mutant frequencies is contributed by nonepithelial cells, such as infiltrating inflammatory cells. First, for DNA isolation, we separated intestinal epithelial cells from lamina propria and submucosa, where the inflammatory cells reside (Fig. 1; Supplementary Fig. S1). Second, the number of inflammatory cells present among epithelial cells is very small especially in the B6 mice. Third, B6 Gpx1/2-DKO ileum had significantly lower levels of inflammatory cells than B6.129 Gpx1/2-DKO mice (Fig. 3A) but still accumulated a 2-fold and a 4-fold higher mutation load, respectively, than control mice at 28 days and 8 months of age. Fourth, B6.129 Gpx1/2-DKO mice at 28 days of age had as many inflammatory cells as 8-month-old B6.129 Gpx1/2-DKO mice but yet had lower mutant frequencies. Lastly, intestinal epithelial cells have similar, if not higher, levels of mutant frequencies than bone marrow cells and spleen (30, 31), sources of inflammatory cells and lymphocytes. Thus, the increased mutant frequencies found in cells isolated from B6 and B6.129 Gpx1/2-DKO intestinal and colonic mucosa are most likely reflecting cumulative mutations present in the epithelium rather than in the inflammatory cells. A large variation of mutant frequency was detected in the ileum among both 8-month-old B6 and B6.129 Gpx1/2-DKO mice; this is likely due to differences in the disease history and severity.

Oxidative stress not only induces DNA damage, as will be discussed later, but also increases cell proliferation and apoptosis. We have found that both B6 and B6.129 Gpx1/2-DKO ileal epithelia have ~2.5-fold higher number of proliferating cells than the age-matched non-DKO controls (Fig. 3B). Because cell proliferation would increase the probability of mutation accumulation (32), the higher level of cell proliferation in Gpx1/2-DKO intestinal epithelium can also contribute to a higher mutation load with more efficient fixation.

It is interesting that genetic background plays a significant role in the severity of inflammation and cancer incidence in the Gpx1/2-DKO model, but B6 and B6.129 Gpx1/2-DKO ilea have similar levels of gene mutations and cell proliferation. Although we detected higher mutant frequencies in the ileum of B6.129 mice compared with the age-matched B6 mice in the 8-month-old control group (9.03 ± 1.68 x 10–5 versus 6.01 ± 1.31 x 10–5) as well as in the Gpx1/2-DKO group (38.94 ± 15.50 x 10–5 versus 25.54 ± 10.33 x 10–5), these increases were not statistically significant. This result suggests that, although mutations are driven by inflammation and may be required for tumor formation, differences in genetic background between B6 and B6.129 Gpx1/2-DKO mice are essential to produce malignancy in intestinal epithelium. There may be a maximal mutation load that cells can tolerate. The concept of maximal preneoplastic increase of mutant frequencies has been suggested by Busuttil et al. (33) when they observed a similarly high mutant frequency (20 x 10–5) in 6- and 12-month-old B6 Sod1-null mouse liver, which was 4-fold higher than that in the control B6 liver. The 4-fold increase of mutation load during 7 months may be reaching the maximal load in the ileum. Cells with higher mutation load may be subjected to apoptosis, which may be more prominent in the ileal and colonic crypt epithelium of B6.129 Gpx1/2-DKO mice (17). Alternatively, it is possible that B6.129 Gpx1/2-DKO ileum do have a slightly higher mutation load than B6 mice, and a 1.5-fold higher mutant frequency can result in 10-fold higher tumor incidence.

Analysis of mutation spectra was done to determine the likely cause of the mutations. The majority of G:C>A:T transition mutations in both B6 Gpx1/2-DKO and B6 WT mice occurred at CpG sites but this was much more pronounced for the WT mice (Fig. 4A). CpG transition mutations are the single most important endogenous mutations linked to cancer and inherited disease (34–36). The major mechanism for CpG mutagenesis is thought to be the deamination of 5-methylcytosine bases at CpG sequences. Other possibilities, including oxidation of 5-methylcytosine, have also been considered (37, 38). Our data suggest that increased production of ROS and inflammation associated with GPX deficiency does not substantially increase mutagenesis at methylated CpG sequences.

There are at least two possible reasons why the absolute frequencies of all types of mutations were increased in the B6 Gpx1/2-DKO mice. The first is that enhanced cell proliferation, relative to non-DKO mice, plays a role in enhanced mutant frequencies in general, and this would apply to all types of mutations. However, another possible explanation is that oxidative DNA damage can produce many different types of mutations. For example, damage of guanine or cytosine can produce all types of transitions and transversions observed at G:C bp, and damage at thymine (and less likely adenine) has the potential to produce any type of mutation at A:T bp. It is even possible that the mutations in the control mice are in part due to oxidative DNA damage.

One major mutation type induced by oxidative damage is the G:C>T:A transversion, presumably a consequence of the presence of 8-oxodeoxyguanosine (8-oxo-dG). B6 Gpx1/2-DKO mice not only had increased transversion mutation frequencies (Fig. 4A) but also had prominent G:C>T:A hotspots at positions 109 and 192 (Supplementary Fig. S3). However, the relative contribution of G:T transversions to the mutational spectrum was decreased in the B6 Gpx-DKO mice (Fig. 4B). Because 8-oxo-dG is induced by oxidative damage and mainly causes G:C>T:A transversions (39), we quantified 8-oxo-dG in ileal genomic DNA of B6 WT or B6 Gpx1/2-DKO mice using high-pressure liquid chromatography (HPLC)–mass spectrometry (MS)/MS (24). However, we observed no significant differences (data not shown).

The new hotspots occurring specifically in B6 Gpx1/2-DKO mice were characterized mainly by G:C>C:G transversions and single-base deletions in small mononucleotide repeats (Supplementary Figs. S2 and S3). Four unique hotspots of G:C>C:G transversions (at positions 31, 86, 211, and 224) were observed in B6 Gpx1/2-DKO mice and thus might represent clonal expansions. Even if we exclude the unique hotspots, which can contribute up to 40% of the total mutation load (Supplementary Fig. S3; 93 DKO), B6 Gpx1/2-DKO ileum would still have a ≥2.5-fold higher mutant frequency than its control ileum. The G:C>C:G transversions can be produced by replicative bypass of oxidation products of 8-oxo-dG, such as guanidinohydantoin and spiroiminodihydantoin (40), types of lesions, which currently cannot be quantified by HPLC/MS/MS analysis. This type of mutation also seems to be unique to B6 Gpx1/2-DKO mice and is not elevated in B6 mice deficient in SOD1 (33). However, increases in single-base deletions seem to be associated with oxidative stress and were also elevated in B6 Sod1-null liver and kidney (33). Deletions in runs of Gs may be induced directly by oxidative DNA damage because guanine is the base most susceptible to oxidation. Oxidative stress is known to increase frameshift mutations in mismatch repair proficient bacterial cells (41, 42) and in human cells (43). Oxidant-induced strand breaks at mononucleotide repeats may increase the probability of misalignment or strand slippage during DNA replication (2, 41). In addition, ROS associated with chronic inflammation have the capacity to damage the protein components of the mismatch repair system directly, leading to a failure to correct single-base mismatches and small insertion/deletion loops that occur during DNA replication (44).

In summary, we observed an increased mutation load in the ileum of both B6 and B6.129 Gpx1/2-DKO mice compared with their age-matched control mice. Because B6 Gpx1/2-DKO mice had milder ileocolitis and 10-fold lower tumor incidence than B6.129 Gpx1/2-DKO mice, these data suggest that, although inflammation is clearly associated with gene mutations, an increase in mutations rarely leads to neoplastic transformation in B6 Gpx1/2-DKO mice having a tumor-resistant genetic background.


   Acknowledgments
 
Grant support: NIH grant CA84469 (G.P. Pfeifer), Crohn's and Colitis Foundation of America Senior Research Award (F-F. Chu), NIH grants CA119272 and CA114569 (F-F. Chu), and Eli and Edythe L. Broad Foundation BMRP (R.S. Esworthy).

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.

We thank Dr. Paul Frankel for statistical analysis, Tina Montgomery and Sofia Loera for tissue processing, and Dr. Peiguo Chu for tumor grading.


   Footnotes
 
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

3 Unpublished data. Back

Received 3/ 2/06. Revised 8/ 3/06. Accepted 8/10/06.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Epe B. Role of endogenous oxidative DNA damage in carcinogenesis: what can we learn from repair-deficient mice? Biol Chem 2002;383:467–75.[CrossRef][Medline]
  2. Evans MD, Cooke MS. Factors contributing to the outcome of oxidative damage to nucleic acids. BioEssays 2004;26:533–42.[CrossRef][Medline]
  3. Termini J. Hydroperoxide-induced DNA damage and mutations. Mutat Res 2000;450:107–24.[Medline]
  4. Hussain SP, Hofseth LJ, Harris CC. Radical causes of cancer. Nat Rev Cancer 2003;3:276–85.[CrossRef][Medline]
  5. Cadet J, Douki T, Gasparutto D, Ravanat JL. Oxidative damage to DNA: formation, measurement, and biochemical features. Mutat Res 2003;531:5–23.[Medline]
  6. Jackson AL, Loeb LA. The contribution of endogenous sources of DNA damage to the multiple mutations in cancer. Mutat Res 2001;477:7–21.[Medline]
  7. Coussens LM, Werb Z. Inflammation and cancer. Nature 2002;420:860–7.[CrossRef][Medline]
  8. Seril DN, Liao J, Yang GY, Yang CS. Oxidative stress and ulcerative colitis-associated carcinogenesis: studies in humans and animal models. Carcinogenesis 2003;24:353–62.[Abstract/Free Full Text]
  9. Itzkowitz SH, Yio X. Inflammation and cancer IV. Colorectal cancer in inflammatory bowel disease: the role of inflammation. Am J Physiol Gastrointest Liver Physiol 2004;287:G7–17.[Abstract/Free Full Text]
  10. Kilgore SP, Sigel JE, Goldblum JR. Hyperplastic-like mucosal change in Crohn's disease: an unusual form of dysplasia? Mod Pathol 2000;13:797–801.[CrossRef][Medline]
  11. Rhodes JM, Campbell BJ. Inflammation and colorectal cancer: IBD-associated and sporadic cancer compared. Trends Mol Med 2002;8:10–6.[CrossRef][Medline]
  12. Touati E, Michel V, Thiberge JM, et al. Chronic Helicobacter pylori infections induce gastric mutations in mice. Gastroenterology 2003;124:1408–19.[CrossRef][Medline]
  13. Chu FF, Esworthy RS, Doroshow JH. Role of Se-dependent glutathione peroxidases in gastrointestinal inflammation and cancer. Free Radic Biol Med 2004;36:1481–95.[CrossRef][Medline]
  14. Esworthy RS, Aranda R, Martin MG, et al. Mice with combined disruption of Gpx1 and Gpx2 genes have colitis. Am J Physiol Gastrointest Liver Physiol 2001;281:G848–55.[Abstract/Free Full Text]
  15. Esworthy RS, Binder SW, Doroshow JH, Chu FF. Microflora trigger colitis in mice deficient in selenium-dependent glutathione peroxidase and induce Gpx2 gene expression. Biol Chem 2003;384:597–607.[CrossRef][Medline]
  16. Erdman SE, Poutahidis T, Tomczak M, et al. CD4+ CD25+ regulatory T lymphocytes inhibit microbially induced colon cancer in Rag2-deficient mice. Am J Pathol 2003;162:691–702.[Abstract/Free Full Text]
  17. Chu FF, Esworthy RS, Chu PG, et al. Bacteria-induced intestinal cancer in mice with disrupted Gpx1 and Gpx2 genes. Cancer Res 2004;64:962–8.[Abstract/Free Full Text]
  18. Tan M, Li S, Swaroop M, et al. Transcriptional activation of the human glutathione peroxidase promoter by p53. J Biol Chem 1999;274:12061–6.[Abstract/Free Full Text]
  19. Yan W, Chen X. GPX2, a direct target of p63, inhibits oxidative stress-induced apoptosis in a p53-dependent manner. J Biol Chem 2006;281:7856–62.[Abstract/Free Full Text]
  20. Ihrig M, Schrenzel MD, Fox JG. Differential susceptibility to hepatic inflammation and proliferation in AXB recombinant inbred mice chronically infected with Helicobacter hepaticus. Am J Pathol 1999;155:571–82.[Abstract/Free Full Text]
  21. Esworthy RS, Yang L, Frankel PH, Chu FF. Epithelium-specific glutathione peroxidase, Gpx2, is involved in the prevention of intestinal inflammation in selenium-deficient mice. J Nutr 2005;135:740–5.[Abstract/Free Full Text]
  22. Esworthy RS, Swiderek KM, Ho YS, Chu FF. Selenium-dependent glutathione peroxidase-GI is a major glutathione peroxidase activity in the mucosal epithelium of rodent intestine. Biochim Biophys Acta 1998;1381:213–26.[Medline]
  23. Pfeifer GP, Chen HH, Komura J, Riggs AD. Chromatin structure analysis by ligation-mediated and terminal transferase-mediated polymerase chain reaction. Methods Enzymol 1999;304:548–71.[Medline]
  24. Besaratinia A, Synold TW, Xi B, Pfeifer GP. G-to-T transversions and small tandem base deletions are the hallmark of mutations induced by ultraviolet a radiation in mammalian cells. Biochemistry 2004;43:8169–77.[CrossRef][Medline]
  25. Jacoby RF, Hohman C, Marshall DJ, et al. Genetic analysis of colon cancer susceptibility in mice. Genomics 1994;22:381–7.[CrossRef][Medline]
  26. Swiger RR, Cosentino L, Shima N, et al. The cII locus in the MutaMouse system. Environ Mol Mutagen 1999;34:201–7.[CrossRef][Medline]
  27. Hanawalt PC. Transcription-coupled repair and human disease. Science 1994;266:1957–8.[Free Full Text]
  28. Hussain SP, Amstad P, Raja K, et al. Increased p53 mutation load in noncancerous colon tissue from ulcerative colitis: a cancer-prone chronic inflammatory disease. Cancer Res 2000;60:3333–7.[Abstract/Free Full Text]
  29. Hussain SP, Raja K, Amstad PA, et al. Increased p53 mutation load in nontumorous human liver of wilson disease and hemochromatosis: oxyradical overload diseases. Proc Natl Acad Sci U S A 2000;97:12770–5.[Abstract/Free Full Text]
  30. Hernandez LG, Heddle JA. A carcinogenic western diet does not induce somatic mutations in various target tissues of transgenic C56BL/6 mice. Mutat Res 2005;570:185–96.[Medline]
  31. Ono T, Ikehata H, Nakamura S, et al. Age-associated increase of spontaneous mutant frequency and molecular nature of mutation in newborn and old lacZ-transgenic mouse. Mutat Res 2000;447:165–77.[Medline]
  32. Busuttil RA, Rubio M, Dolle ME, et al. Mutant frequencies and spectra depend on growth state and passage number in cells cultured from transgenic lacZ-plasmid reporter mice. DNA Repair (Amst) 2006;5:52–60.[Medline]
  33. Busuttil RA, Garcia AM, Cabrera C, et al. Organ-specific increase in mutation accumulation and apoptosis rate in CuZn-superoxide dismutase deficient mice. Cancer Res 2005;65:11271–5.[Abstract/Free Full Text]
  34. Jones PA, Baylin SB. The fundamental role of epigenetic events in cancer. Nat Rev Genet 2002;3:415–28.[Medline]
  35. Gonzalgo ML, Jones PA. Mutagenic and epigenetic effects of DNA methylation. Mutat Res 1997;386:107–18.[CrossRef][Medline]
  36. Pfeifer GP. p53 mutational spectra and the role of methylated CpG sequences. Mutat Res 2000;450:155–66.[Medline]
  37. Lee DH, O'Connor TR, Pfeifer GP. Oxidative DNA damage induced by copper and hydrogen peroxide promotes CG->TT tandem mutations at methylated CpG dinucleotides in nucleotide excision repair-deficient cells. Nucleic Acids Res 2002;30:3566–73.[Abstract/Free Full Text]
  38. Yoon JH, Iwai S, O'Connor TR, Pfeifer GP. Human thymine DNA glycosylase (TDG) and methyl-CpG-binding protein 4 (MBD4) excise thymine glycol (Tg) from a Tg:G mispair. Nucleic Acids Res 2003;31:5399–404.[Abstract/Free Full Text]
  39. Wood ML, Dizdaroglu M, Gajewski E, Essigmann JM. Mechanistic studies of ionizing radiation and oxidative mutagenesis: genetic effects of a single 8-hydroxyguanine (7-hydro-8-oxoguanine) residue inserted at a unique site in a viral genome. Biochemistry 1990;29:7024–32.[CrossRef][Medline]
  40. Henderson PT, Delaney JC, Muller JG, et al. The hydantoin lesions formed from oxidation of 7,8-dihydro-8-oxoguanine are potent sources of replication errors in vivo. Biochemistry 2003;42:9257–62.[CrossRef][Medline]
  41. Jackson AL, Chen R, Loeb LA. Induction of microsatellite instability by oxidative DNA damage. Proc Natl Acad Sci U S A 1998;95:12468–73.[Abstract/Free Full Text]
  42. Jackson AL, Loeb LA. Microsatellite instability induced by hydrogen peroxide in Escherichia coli. Mutat Res 2000;447:187–98.[Medline]
  43. Gasche C, Chang CL, Rhees J, et al. Oxidative stress increases frameshift mutations in human colorectal cancer cells. Cancer Res 2001;61:7444–8.[Abstract/Free Full Text]
  44. Chang CL, Marra G, Chauhan DP, et al. Oxidative stress inactivates the human DNA mismatch repair system. Am J Physiol Cell Physiol 2002;283:C148–54.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 Mol Biol EvolHome page
N. Phillips, M. Salomon, A. Custer, D. Ostrow, and C. F. Baer
Spontaneous Mutational and Standing Genetic (Co)variation at Dinucleotide Microsatellites in Caenorhabditis briggsae and Caenorhabditis elegans
Mol. Biol. Evol., March 1, 2009; 26(3): 659 - 669.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Res.Home page
M. A. Hahn, T. Hahn, D.-H. Lee, R. S. Esworthy, B.-w. Kim, A. D. Riggs, F.-F. Chu, and G. P. Pfeifer
Methylation of Polycomb Target Genes in Intestinal Cancer Is Mediated by Inflammation
Cancer Res., December 15, 2008; 68(24): 10280 - 10289.
[Abstract] [Full Text] [PDF]

Home page
 J. Clin. Endocrinol. Metab.Home page
Y. Song, N. Driessens, M. Costa, X. De Deken, V. Detours, B. Corvilain, C. Maenhaut, F. Miot, J. Van Sande, M.-C. Many, et al.
Roles of Hydrogen Peroxide in Thyroid Physiology and Disease
J. Clin. Endocrinol. Metab., October 1, 2007; 92(10): 3764 - 3773.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplementary Data
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Lee, D.-H.
Right arrow Articles by Chu, F.-F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Lee, D.-H.
Right arrow Articles by Chu, F.-F.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online