Int. J. Cancer: 116, 679–685 (2005) ' 2005 Wiley-Liss, Inc.
Microenvironmental influences on mutagenesis in mammary epithelial cells Erzsebet Papp-Szabo1, P. David Josephy2 and Brenda L. Coomber3* 1 Department of Chemistry and Biochemistry, University of Guelph, Guelph, Ontario, Canada 2 Department of Molecular and Cellular Biology, University of Guelph, Guelph, Ontario, Canada 3 Department of Biomedical Sciences, University of Guelph, Guelph, Ontario, Canada Tumor progression may be viewed as an evolutionary process at the cellular level. Because blood supply to solid tumors is inadequate, the cancer cells face a hostile microenvironment characterized by hypoxia or anoxia, acidic extracellular pH and nutrient deficiencies. It has been proposed that these factors result in increased levels of spontaneous mutagenesis and thereby contribute to tumor progression. We have examined spontaneous mutagenesis in vitro and in vivo, using previously characterized cell lines (mammary epithelial cells [ME] and mammary fibroblast cells [MFib]) from the mammary gland of the BigBlueTM rat, carrying a transgene construct suitable for the detection of mutations. Cells were exposed in vitro to control conditions, low pH, or to glucose deprivation, under normoxic or hypoxic culture conditions, and were also grown as xenografted tumors in immune-deficient mice. We examined cell survival and mutant frequency/spectrum at the cII locus. Significant increases in mutant frequency were observed in ME cells exposed to hypoxia alone or in combination with no glucose; the latter condition also resulted in reduced clonogenic survival. Cells grown as xenografts and then recovered and expanded in culture also had elevated frequencies of spontaneous mutations. We observed a shift in the spontaneous mutation spectrum between the ME cells and the MET cells (cultured in vitro or isolated from mouse xenograft tumors). These results support the concept that the tumor microenvironment contributes to tumor progression by enhancing spontaneous mutagenesis, that different cell types from the same organ can respond differently to these stresses and that differences in microenvironment may influence the types of mutations that arise. ' 2005 Wiley-Liss, Inc. Key words: mammary epithelial microenvironment; mutagenesis
cells;
cII
assay;
tumor;
Cancer cells from clinical tumors usually bear large numbers of mutations, at scales ranging from the single nucleotide and microsatellite to the chromosome. Many investigators believe that a ‘‘mutator phenotype’’1 is a hallmark of most cancers. Chromosomal instability leads to aneuploidy and microsatellite instability (MIN) leads to multiple errors in DNA sequence.2–5 Mutations in oncogenes and tumor suppressors have been implicated in these aspects of genetic instability.2,6 While some progress has been made in our understanding of the molecular defects involved in chromosomal stability (such as the importance of proteins involved in sister-chromatid segregation during mitosis)7,8 our understanding of the defective systems in MIN is much more advanced. Base mismatches arising during replication or DNA repair are normally corrected by the action of enzymatic ‘‘proofreading’’ machinery. Genes homologous to those encoding the E. coli mismatch repair proteins MutS and MutL are found in humans. MIN may arise due to mutations in the genes encoding these enzymes (MSH6, MLH1, etc.)4,5 or via other mechanisms, such as the persistent inflammation occurring during ulcerative colitis.9 Tumor progression is a form of cellular evolution, a consequence of both mutation and natural selection.3,10 Malignant progression of a cancer cell into a tumor is marked by increased cell proliferation, resistance to apoptotic cell death, angiogenic potential, invasiveness and metastatic tendency.11–19 Several previous studies using a variety of approaches have shown that the conditions present in the microenvironment of solid tumors (such as hypoxia, nutrient deprivation, acidosis, etc.) are mutagenic and/or enhance genetic instability.14,20–22 Thus, the microenvironment Publication of the International Union Against Cancer
may contribute to cancer progression by both promoting the formation of mutations and selecting for cells that are resistant to these physicochemical stresses.6 Human cancers most commonly originate from the parenchyma (epithelial cells) rather than from the stroma of organs,23 although stromal cells may contribute to the progression of carcinomas via cell-cell interactions.24 Multiple biological factors may be responsible for the predominance of carcinomas. Epithelium is directly exposed to environmental mutagens (inhaled and ingested chemicals, ultraviolet [UV] irradiation) and undergoes particularly rapid cell turnover (skin, intestinal mucosa, etc.). Stromal and parenchymal cells may also differ in carcinogen detoxification, DNA repair and other biochemical processes relevant to mutagenesis, but there is, as yet, rather little data bearing on these questions. Differences in the susceptibility of parenchymal and stroma cell types could also arise from differential responses to the physicochemic conditions in the tumor microenvironment. Several previous studies have directly measured mutagenesis in mammalian tumors or specific cancer cell types in vivo. For instance, Mirsalis25 reported that mutation frequencies in diethylnitrosamine-induced hepatocellular carcinomas in the Big Blue lacI transgenic mouse were elevated relative to normal liver tissue from diethylnitrosamine-treated animals. Baross-Francis et al.26 used the BC-1 transgenic lacI mutation detection system to study spontaneous thymic lymphomas arising in Msh2–/– mice defective in DNA mismatch repair and observed spontaneous mutation frequencies several-fold higher than in normal thymus tissue from these animals. Elevated spontaneous mutation frequencies in thymic lymphomas were also reported by Zhang et al.27 However, in a subsequent study,28 analysis was extended to other tumor types arising in these mice, e.g., squamous cell carcinoma, osteogenic sarcoma, and these tumors did not show elevated spontaneous mutation frequencies. We have generated epithelial and fibroblast cell lines derived from the mammary gland of a female transgenic BigBlue rat.29 The reporter transgenes present in these cells provide a convenient endpoint for detection and quantification of mutagenesis. In this study, we have investigated the possible effects of aspects of the tumor microenvironment (nutrient deprivation, hypoxia and reduced pH, alone and in combination) on spontaneous mutagenesis in an epithelial and a fibroblast cell line from the same organ. We have also measured mutant frequencies and mutational spectra in mammary epithelial cells grown as tumor xenografts in immune-deficient mice. Our findings confirm that spontaneous mutations occur in these cells when they are grown as xenografted tumors and that elements of the tumor microenvironment, such as reduced glucose combined with hypoxia, are mutagenic to these cells. Our results further demonstrate a significant difference in the microenvironmental mutagenic responses between stromal and
Grant sponsor: Canadian Institutes for Health Research; Grant number: MOP 49565. *Correspondence to: Department of Biomedical Sciences, University of Guelph, Guelph, Ontario, Canada N1G 2W1. Fax: 519-767-1450. E-mail:
[email protected] Received 23 July 2004; Accepted 7 January 2005 DOI 10.1002/ijc.21088 Published online 22 April 2005 in Wiley InterScience (www.interscience. wiley.com).
680
` ET AL. PAPP-SZABO
parenchymal cells from the same organ and provide molecular evidence of clonal evolution with carcinoma progression. Material and methods Cell culture Isolation and characterization of the mammary epithelial (ME) cell lines and the mammary fibroblast (MFib) cell lines has been reported previously.29 Cells were maintained in standard medium (DMEM: F12 5 1:1, 5% FBS) and passaged when confluent. Experimental exposures were performed under serum-free conditions chosen to reflect those existing within solid tumors.23,30–34 Conditions were: serum-free DMEM (control); serum-free, glucose-free DMEM (no-glucose condition); serum-free DMEM, pH 6.75 (low pH condition); serum-free, glucose-free DMEM, pH 6.75 (no glucose 1 low pH condition), as previously described.16,21,35 Cells were grown to 60–70% confluency and exposed to standard CO2 incubator conditions (room air; ~21% O2; normoxia) or continuous infusion of oxygen-free gas (5% CO2, 95% N2) in a modular incubator chamber (Billups-Rothenberg, Del Mar, CA) for 24 hr (hypoxia). After this treatment, the cultures were washed with PBS, allowed to recover in standard medium (normoxia) for 24 hr, trypsinized and viability was determined by trypan blue dye exclusion. For assessment of clonogenic survival, for each condition, 800 viable cells per plate were seeded, in triplicate in every experiment, as previously reported.29,36 For DNA isolation and mutation assay, 5 3 105 viable cells were seeded and cultured in standard medium until confluency was reached, followed by DNA isolation and measurement of cII mutant frequency, as described in our earlier publications.29,36,37 In assessments of the effects of specific stressors, we performed each experiment at least 3 times. Tumorigenicity ME cells (2 3 106 cells suspended in 0.1% BSA/PBS) were injected into the right flank of each of 3 immune-deficient RAG-1 null mice.38 The mice were monitored weekly, and after several months, 1 mouse developed a tumor at the injection site. Cells were recovered from the digested tumor tissue and expanded in vitro, and their BigBlue ME cell origin was confirmed by the presence of the transgene.39 This cell line, named MET, was used for the in vitro studies described below and was also reinjected into 4 additional RAG-1 null mice to generate xenografted tumors, for further analysis. DNA isolation and cII mutagenicity assay Details of the methods for the cII assay have been described previously.29,36 Briefly, genomic DNA was isolated from cultured cells or pieces of tumor, rehydrated in buffer (10 mM Tris-HCl, pH 7.6, 1 mM EDTA) and packaged with Transpack lambda bacteriophage packaging extract (Stratagene, LaJolla CA). The cII assay was performed following Stratagene protocols, using E. coli strain G1250. The packaged extract was used at 1:100 dilution for titer plates (3 plates per data point) and undiluted for test plates (9 plates per data point). Titer plates were grown overnight at 37°C and the plaques counted. Test plates were grown for 48 hr at 23°C, the temperature at which only cII mutant bacteriophage form plaques. Incubator temperature was controlled by a differential controller (Model CN76120, Omega Engineering, Laval, Quebec, Canada). Any dubious plaques were cored and replated. The mutant frequency was calculated as the ratio of the number of mutant plaques (23°C selective plates) to the total number of plaque-forming units, calculated from the average plaque numbers on the 37°C titer plates. cII gene sequencing Randomly chosen mutant plaques were cored, replated at low density, and incubated at the selective temperature (23°C) for 48 hr. From each plate, 1 plaque was transferred to sterile water
(100 ll) and boiled for 4 min; an aliquot (20 ll) of the sample was used for PCR amplification. The amplification mixture contained 2.5 U Platinum Taq DNA polymerase (Invitrogen, Burlington, Ontario, Canada), cII primers (each, 0.5 lM; cII.1 primer: ACCCCGCTCTTACACGTT; cII.2 primer: CTCTGCCGAAGTTGAGTATTT; Invitrogen) and dNTPs (each, 0.2 lM; Sigma, Oakville, Ontario, Canada) in MgCl2-containing PCR buffer. The final volume for each PCR reaction was 100 ll. PCR conditions were as follows: 95°C 3 5 min, 55°C 3 30 sec, 72°C 3 1 min; then 31 cycles of 95°C 3 30 sec, 55°C 3 45 sec, 72°C 3 80 sec; finally, 72°C 3 9 min. The PCR products were purified with MontageTM PCR centrifugal devices (Millipore, Billerica, MA). An aliquot (3 ll) from each sample was run on an agarose gel (1%) along with a quantifying marker, to determine the DNA concentration. Purified DNA was sequenced (dye terminator cycle sequencing, Laboratory Services, University of Guelph), using the cII.1 primer. Statistical analysis Within each clonogenic survival experiment, the numbers of colonies on the 3 control (DMEM) plates were averaged, and the colony numbers on other experimental plates were compared to this average. For each condition, the 9 relative CFU numbers (3 per experiment, 3 experiments) were pooled and the averages and standard errors were calculated.37 For analysis of the significance of mutant frequency and clonogenic survival data, we first performed ANOVA, and data with a significant F-value were further compared using the 2-tailed t-test at a 95% confidence level. The data on distribution of mutations among classes were analyzed by Fisher’s exact test. Results Influence of in vitro conditions on mutation and clonogenic survival We investigated the effect of environmental stress conditions reflecting the tumor microenvironment—namely, acidic pH; no glucose; combined no glucose 1 acidic pH—on the clonogenic survival and mutant frequency of mammary epithelial and mammary fibroblast cells. The cells were incubated in vitro in media adjusted to the desired stress conditions. Since oxygen depletion is another major stressor in solid tumors, the experiments were performed under both normoxic and hypoxic conditions. Results are shown in Figure 1. In general, the clonogenic survival of the ME and MFib cells was little affected by stress exposures under normoxic incubation (Fig. 1a and b). However, under hypoxia, there was a statistically significant reduction in clonogenic survival (to ~34 %) of the ME cells exposed to no glucose conditions (Fig. 1a). A less dramatic reduction in survival was seen in the same cells exposed to no glucose 1 acidic pH under hypoxia (Fig. 1a). The number of cells remaining viable after the acute exposure differed greatly between the cell lines and among the stress conditions (data not shown). The no-glucose treatment was acutely toxic to ME cells and, to a lesser extent, MFib cells. Nevertheless, in each case, the same number of viable cells (800) was plated for quantification of colony formation. The cII transgene assay was used to measure mutant frequencies in cells cultured after acute exposure to stress conditions (Fig. 1c and d). These data can be examined with respect to stressor effects (different bars within each panel), the effect of hypoxia (paired bars within each group), and differences between the 2 cell lines (Fig. 1c vs. d). In most cases, mutant frequencies were higher under hypoxia than under normoxia. Exposure of ME cells (Fig. 1c) to hypoxia under DMEM or noglucose conditions resulted in approximately 2-fold elevation of mutant frequency relative to the normoxic DMEM control, and these differences were statistically significant (p 5 0.0036 and p 5 0.0135, respectively; bars marked with asterisks on Fig. 1).
681
MUTATION SPECTRA IN THE TUMOR ENVIRONMENT
FIGURE 1 – Relative clonogenic survival (a,b) and relative mutant frequency values (c,d) of the ME mammary epithelial (a,c) and MFib mammary fibroblast (b,d) cell lines. Shaded bars represent normoxic exposures and open bars represent hypoxic exposures. DMEM: control conditions; A: acidic conditions; NG: no glucose conditions; A 1 NG: acidic plus no glucose conditions. For clonogenic survival, the values represent the pooled average (6 standard error of the mean [SEM]) of 3 independent experiments, each experiment executed in triplicate (n 5 9). The mutant frequency data represents the relative average 6 SEM. In each case, the mutant frequency values were compared to the DMEM (untreated) values in the same experiment. The error bars for the DMEM (control) values represent the interexperiment variations in mutant frequency values. The average mutant frequency values were 7.5 3 10–5 (epithelial cells) and 4.8 3 10–5 (fibroblast cells). Between 168,000 and 758,000 plaques were scored for each data point. *indicates statistically significant difference between this treatment and DMEM normoxia (p < 0.05).
TABLE I – SPONTANEOUS cII MUTANT FREQUENCIES FOR ME AND MET CELL LINES UNDER NON-STRESS CULTURE CONDITIONS ME cells
MET cells
Plaques scored
Mutants
Mutant frequency (3 1025)
Plaques scored
Mutants
Mutant frequency (3 1025)
100,167 151,200 71,834 79,667 46,333 123,667 185,167
6 17 5 5 3 7 18 Mean S.D. S.E.M.
5.99 11.2 6.96 6.28 6.47 5.66 9.72 7.5 2.1 0.81
49,667 105,467 33,367 32,167 48,500 30,000 24,000
7 11 5 3 8 8 5 Mean S.D. S.E.M.
14.1 10.4 15.0 9.33 16.5 26.7 20.8 16.1 6.0 2.3
Under the remaining 2 stress conditions, acidic pH and acidic pH 1 no glucose, mutant frequencies were not significantly different between hypoxic and normoxic conditions. With MFib cells, the mutant frequency was higher in hypoxic than in normoxic cells, under each of the 4 exposure conditions. However, in no case was the difference between hypoxic and normoxic conditions statistically significant (Fig. 1d). Tumorigenicity of mammary epithelial cells ME cells were weakly tumorigenic in immune-deficient mice, with 1 tumor obtained from 3 mice after a period of 5 months. This tumor was placed into culture and the MET cell line was thereby obtained. (We did not attempt to reclone MET cells from a single cell.) MET cells retained morphology and karyotype of
the ME cells but exhibited a lower plating efficiency and a greater tendency to detach from the surface of culture dishes. We compared the spontaneous mutant frequencies of ME and MET cells (under nonstressed culture conditions) and found that the frequency was more than twice as high in the latter cells (Table I). MET cells (2 3 106 cells) were injected into 4 immune-deficient mice. Within only 2–3 weeks, each animal developed a detectable tumor at the site of injection. Each of the 4 tumors (size, 300–500 mm3) was removed, cut into small pieces and snap-frozen in liquid nitrogen. Pieces were thawed and used for DNA isolation and mutant frequency analysis. As anticipated, the yield of bacteriophage plaques from this analysis was low, presumably due to the poor packaging efficiency of the tumor DNA (not all of which would be derived from the MET cells per se). Despite this drawback, we were able to measure mutant
` ET AL. PAPP-SZABO
682
TABLE II – cII MUTANT FREQUENCIES IN TUMORS THAT DEVELOPED IN FOUR RAG1-NULL MICE INJECTED WITH MET CELLS Animal 1
Expt. Plaques
1 2 3 Total Mutant frequency
2 Mutations
23,500 4 35,700 1 19,333 1 78,533 6 25 7.64 3 10
Plaques
3 Mutations
17,667 0 38,967 1 11,333 1 67,967 2 25 2.94 3 10
Plaques
4 Mutations
8,000 9 8,000 9 23 1.13 3 10
Plaques
Mutations
22,250 1 14,167 0 36,417 1 25 2.75 3 10
TABLE III – ANALYSIS OF TRANSITIONS, TRANSVERSIONS, AND INSERTIONS/ DELETIONS IN cII GENES FROM CULTURED ME AND MET CELLS, AND MET TUMORS1 Mutation
G:C to A:T At CpG sites2 A:T to G:C G:C to C:G G:C to T:A A:T to T:A A:T to C:G ins / del Total Transition Transversion Transition bias
DMEM
No glucose
Hypoxia
No glucose hypoxia
Tumor cells
Tumor tissue
Total
2 1 (50%) 1 3 2 0 1 1 10
3 1 (33%) 0 3 1 1 1 0 9
3 1 (33%) 1 3 1 1 1 1 11
5 1 (20%) 0 4 1 0 2 0 12
7 3 (43%) 0 0 2 2 1 2 14
4 2 (50%) 1 2 0 0 1 3 11
24 9 (38%) 3 15 7 4 7 7 67
3 6 0.50
3 6 0.50
4 6 0.67
5 7 0.71
7 5 1.40
5 3 1.67
27 33 0.82
1 Corrected for clonality.–2G:C to A:T transitions occurring at CpG sites. These mutations are counted among the total G:C to A:T transitions. Numbers for this row are excluded from the column totals, to avoid double-counting.
frequencies for each tumor, although only on the basis of a small number of mutant plaques. As shown in Table II, the cII mutant frequencies varied from 2.7 3 10–5 (lower than that of the original ME cells) to 1.1 3 10–3. Sequence analysis of cII gene mutations in vitro and in vivo We studied cII gene mutations in the ME cells, MET cells and tumors at the DNA sequence level. Sequence analysis was obtained for about 20 mutant plaques in each condition. In the case of the tumor tissue samples, only 18 mutant plaques were available (Table II), and all of these plaques were sequenced. We observed that each of the 6 spectra obtained was dominated by 1 or 2 particular mutations. Each of the ME spectra was dominated by transversion mutations at positions 113 (C to A) and 175 (G to C). Together, these 2 mutations accounted for approximately onehalf of all mutations observed in each in vitro condition. By contrast, neither of these mutations was detected in the spectra of the MET cells or the tumor tissue. Instead, we observed in the MET cells a transition at position 52, which accounted for 8 of the 22 mutations, and in the tumor tissue, a transition at position 64 which accounted for 9 of the 19 mutations (data not shown). We interpret these frequent mutations as representing founder effects in the starting cell populations, rather than ‘‘hot spots’’ (sites highly susceptible to de novo mutation). We analyzed the distribution of mutations among the mutational classes (2 transitions and 4 transversions) in 2 different ways. First, all observed mutations were treated as if they were independent (data not shown). However, since we interpret the common mutations as being clonal, rather than independent, we also performed an analysis of the distribution of mutations, corrected for clonality, i.e., each specific mutation is counted only once, and we calculated the ‘‘transition bias’’ (ratio of transitions to transversions) (Table III). Many positions of the cII gene were susceptible to mutation. Almost all of the mutations seen in our study were base substitutions, with only a few insertions or deletions, most of which were single-base insertions or deletions. For the ME cells exposed to microenvironmental conditions in vitro, the transition bias was close to unity (0.50–0.71) (Table III). Spontaneous mutations in
MET cells grown under standard in vitro culture conditions (5 independent DNA isolations) were analyzed. In this case, the transition bias was 1.4 (Table III). A total of 18 mutant plaques were obtained from DNA that was isolated directly from the 4 METderived mouse tumors (i.e., cells were not expanded in vitro). The transition bias for MET-derived tumors was 1.7. All of these transition biases were not significantly different among any of the 6 spectra (p > 0.20). The base substitutions occurred mainly at G:C sites (90 of 114 events). Among the microenvironment exposures, we observed 13 G:C to A:T transitions, 4 (31%) of which occurred at CpG sites. Discussion Tumors undergo an evolutionary process marked by mutation and natural selection processes, since stressor effects on the rate of spontaneous mutagenesis in cells could contribute to the development of ‘‘mutator’’ phenotypes in tumors. Our ME and MFib cells, carrying a bacteriophage lambda transgene insert, allow analysis of mutant frequencies and mutational specificities under a variety of conditions. In our study, hypoxia alone caused a significant elevation in the mutant frequency in the ME cells relative to control conditions, while having little effect on clonogenic survival. Under no-glucose plus hypoxia conditions, mutant frequency was also elevated and clonogenic survival was reduced. With MFib cells (representing the stromal compartment), mutant frequency was higher under hypoxia, but the difference did not attain statistical significance. In our previous studies with chemical mutagens, the ME cell line was much more sensitive to the toxicity of N-ethyl-N-nitrosourea,29 but the 2 cell lines showed very similar genotoxic responses to this agent. In studies with the polycyclic aromatic hydrocarbons benzo[a]pyrene and 7,12-dimethylbenz[a]anthracene, MFib showed much greater sensitivity than did ME cells, in terms of both genotoxic and cytotoxic responses.39 In the present work, ME cells were more sensitive to the in vitro stress exposures, as measured by acute toxicity and clonogenic survival, particularly following no-glucose 1 hypoxia stress. The relative sensitivities of the 2 cell types may be very different with respect to different chemical or microenvironmental stresses, since so many
683
MUTATION SPECTRA IN THE TUMOR ENVIRONMENT
biochemical and biological parameters influence cellular response. In support of this, when no glucose plus hypoxia was combined with acidic conditions, the reduced clonogenic survival of ME cells was partially ‘‘rescued,’’ suggesting that such conditions may induce a more robust or efficient stress response in these cells, which did not occur under neutral pH. This difference in cell type response when exposed to microenvironmental conditions may contribute to the great preponderance of carcinomas, compared to sarcomas and other solid tumors seen in human cancers.23 The large error bars associated with mutant frequency reflect the small absolute numbers of mutant plaques obtained in the experiments. This yield of mutants, in turn, is limited by the efficiency of recovery of the transgene and by the inherently low frequency of spontaneous mutation. Because statistically significant enhancement of mutant frequency by stress conditions was only seen with the epithelial cells, further experiments (in vivo tumor formation and mutational specificity analysis) were conducted with the epithelial cell line. Many studies have examined the effects of elements of the tumor microenvironment (hypoxia, low pH, glucose deprivation) on cultured cells, and in general find that parameters related to tumor progression, such as survival, invasion and growth factor production, are enhanced.16,40–43 Such stress conditions that exist in the microenvironment of solid tumors are believed to decrease genetic stability and to be mutagenic, hence mutagenic events are presumably responsible for many of these alterations observed in studies of tumor progression. In mouse 3340 fibroblasts, hypoxia and low pH suppressed DNA repair and enhanced mutagenesis.21 Mihaylova et al.35 reported decreased expression of the DNA mismatch repair enzymes MLH1 and the PMS2 under hypoxic conditions in 3340 fibroblasts and human HeLa cells. The downregulation of these enzymes paralleled the induction of mutagenesis in reporter genes; however, other members of the MMR system (MSH2, MSH3, MSH6) were not affected by in vitro exposure to hypoxia.35 Mouse LN12 cells containing the ksupF shuttle vector displayed enhanced mutagenesis when exposed to hypoxia in vitro,14 and in human colorectal carcinoma cells, hypoxia treatment greatly increased spontaneous mutation frequency in a lacZ shuttle vector.44 Hypoxia also enhanced cellular resistance to 6-thioguanine and cisplatin, presumably due to mutagenesis of key resistance genes.44 Thus, the hypoxic environment may influence mutagenesis by altering expression of stress response and/or DNA repair genes. Our findings with ME cells agree with these previous results. The transgenic mutation assay detects the presence of mutant transgenes within DNA samples pooled from large numbers of cells, and therefore measures mutant frequency, i.e., the prevalence of mutations, rather than mutation frequency, i.e., the frequency of induction of new mutations.45 Multiple occurrences of the same mutation in a population could represent hot-spots (specific mutational events which occur at unusually high frequencies). However, as stated earlier, we interpret the multiple occurrences that we observed as likely representing the clonal expansion of a single mutant cell. In support of this interpretation, the specific mutations observed repeatedly in our spectra have not previously been reported as hot-spots in other studies of the cII gene. Therefore, we corrected for clonality by counting only once per spectrum any mutation observed multiple times.46 Injection of our ME cells to 3 immune-deficient RAG1-null mice resulted in only 1 tumor, after 5 months of observation. Cells derived from this tumor (MET cells), were expanded in vitro and reinjected into 4 mice, all of which developed detectable tumors within 2–3 weeks. This difference in the latency period shows that MET cells have much higher tumorigenicity than ME cells. Such progression towards greater malignancy with in vivo ‘‘transfer’’ has been seen in many previous studies.47–50 Strikingly, in our study, the sites (113 and 175) that were frequently mutated in the ME microenvironment studies were never recovered from the MET cells or the tumor tissue. These 2 sites accounted for 47 of
81 mutations in the microenvironment studies, and 0 of 41 mutations in the MET cells and the tumor tissue. This difference is highly significant (p < 0.001). We interpret this as a consequence of the clonal selection51 that occurred during the development of the malignant tumors in vivo. During this process, the subpopulation of cells bearing the mutations at positions 113 and 175 was largely or completely eliminated. The transition bias of spontaneous mutations is typically reported to be greater than 1.0, even though there are twice as many possible transversions as transitions.52 Purine-purine or pyrimidine-pyrimidine mispairings (e.g., resulting from polymerasecatalyzed misincorporation events), which would lead to transversions, probably cause greater structural distortion of the double helix than do purine-pyrimidine mispairings. Consequently, DNA mispairs that would induce transversions are more easily recognized and repaired. However, the reported transition biases in the cII transgene are not large. A study of spontaneous cII mutations arising in the BigBlueTM Rat2 embryonic fibroblast cell line found approximately equal numbers of transversions and transitions.53 Spectra dominated by transversions at G:C base pairs are typically observed with oxidative mutagens, such as ozone or other reactive oxygen species54,55 and these mutations are usually attributed to formation of 8-oxo-7,8-dihydro-20 -deoxyguanosine. In our study, only about one-third of the G:C to A:T transitions occurred at CpG sites, which is lower than other reported values for the cII gene,53,56 and may possibly be reflective of the extensive hypomethylation known to occur in cancer development.57 The transition bias in the spectrum of spontaneous cII mutations in tissues derived from BigBlueTM transgenic rodents is reported to be between 1.6 and 2.8 and varies with tissue type56,58,59 Overall, we found a similar transition bias in tumor tissue and in cells recovered from tumors (1.4–1.7). Although there was likely considerable selection of tumorigenic cells during the initial formation, our results suggest that common factors may influence spontaneous mutation in vivo, independent of tissue types or animal species/strains. Reynolds et al.14 found a 5-fold increase in mutant frequency of the ksupF shuttle in LN12 cell xenografts compared to cultured LN12 cells. This increase in spontaneous mutagenesis between the same cells in vitro and in vivo was not seen in our study, and may be reflective of the difficulty we observed in packaging DNA from the xenografted tumors. It is also possible that since the LN12 cells used by Reynolds et al.14 showed greater propensity for large (>200 base pairs) deletions when grown in vivo, differences in the shuttle construct itself (cII vs. ksupF) may account for this discrepancy. We examined the possible mutagenic role of the tumor microenvironment in tumor progression by evaluating quantitative and qualitative changes in the cII transgene present in rat mammary epithelial and fibroblast cell lines. Hypoxia (in combination with control or no-glucose conditions) produced an approximate doubling in spontaneous mutant frequency compared to normoxic control conditions in the ME cells, but not the fibroblasts. Other studies report up to 5-fold increases in mutation frequency when comparing cells exposed to normoxia versus hypoxia.14 These results may be biased by the fact that only 1 mouse originally injected with ME cells was able to form a tumor. Had the ME cells been exposed to microenvironmental ‘‘stress’’ (such as hypoxia and hypoglycemia) just prior to injection, we may also have seen an enhancement of tumor take, as has been reported for cells exposed to X-ray irradiation prior to xenografting.20 However, the increased tumorigenicity, higher spontaneous mutant frequency, and altered mutational spectrum of the MET cells, relative to their parental cell line ME, indicate that the cell line underwent permanent changes in the process of tumor development. Acknowledgements We thank Dr Kathleen Hill, University of Western Ontario, for critical reading of the manuscript.
684
` ET AL. PAPP-SZABO
References 1. 2. 3. 4. 5. 6. 7.
8. 9.
10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.
28.
Loeb LA. Cancer cells exhibit a mutator phenotype. Adv Cancer Res 1998;72:25–56. Albertson DG, Collins C, McCormick F, Gray JW. Chromosome aberrations in solid tumors. Nat Genet 2003;34:369–76. Cahill DP, Kinzler KW, Vogelstein B, Lengauer C. Genetic instability and Darwinian selection in tumours. Trends Cell Biol 1999;9: M57–60. Heinen CD, Schmutte C, Fishel R. DNA repair and tumorigenesis: lessons from hereditary cancer syndromes. Cancer Biol Ther 2002;1:477–85. Sarasin A. An overview of the mechanisms of mutagenesis and carcinogenesis. Mutat Res 2003;544:99–106. Coomber BL, Yu JL, Fathers KE, Plumb C, Rak JW. Angiogenesis and the role of epigenetics in metastasis. Clin Exp Metastasis 2003;20:215–27. Bernal JA, Luna R, Espina A, Lazaro I, Ramos-Morales F, Romero F, Arias C, Silva A, Tortolero M, Pintor-Toro JA. Human securin interacts with p53 and modulates p53-mediated transcriptional activity and apoptosis. Nat Genet 2002;32:306–11. Zou H, McGarry TJ, Bernal T, Kirschner MW. Identification of a vertebrate sister-chromatid separation inhibitor involved in transformation and tumorigenesis. Science 1999;285:418–22. Hofseth LJ, Khan MA, Ambrose M, Nikolayeva O, Xu-Welliver M, Kartalou M, Hussain SP, Roth RB, Zhou X, Mechanic LE, Zurer I, Rotter V, et al. The adaptive imbalance in base excision-repair enzymes generates microsatellite instability in chronic inflammation. J Clin Invest 2003;112:1887–94. Kerbel RS. Growth dominance of the metastatic cancer cell: cellular and molecular aspects. Adv Cancer Res 1990;55:87–132. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990;61:759–67. Boudreau N, Myers C. Breast cancer-induced angiogenesis: multiple mechanisms and the role of the microenvironment. Breast Cancer Res 2003;5:140–6. Yu JL, Rak JW. Host microenvironment in breast cancer development: inflammatory and immune cells in tumour angiogenesis and arteriogenesis. Breast Cancer Res 2003;5:83–8. Reynolds TY, Rockwell S, Glazer PM. Genetic instability induced by the tumor microenvironment. Cancer Res 1996;56:5754–7. van Kempen LC, Ruiter DJ, van Muijen GN, Coussens LM. The tumor microenvironment: a critical determinant of neoplastic evolution. Eur J Cell Biol 2003;82:539–48. Jang A, Hill RP. An examination of the effects of hypoxia, acidosis, and glucose starvation on the expression of metastasis-associated genes in murine tumor cells. Clin Exp Metastasis 1997;15:469–83. Vaupel P, Hockel M. Blood supply, oxygenation status and metabolic micromilieu of breast cancers: characterization and therapeutic relevance. Int J Oncol 2000;17:869–79. Subarsky P, Hill RP. The hypoxic tumour microenvironment and metastatic progression. Clin Exp Metastasis 2003;20:237–50. Henning T, Kraus M, Brischwein M, Otto AM, Wolf B. Relevance of tumor microenvironment for progression, therapy and drug development. Anticancer Drugs 2004;15:7–14. Paquette B, Little JB. In vivo enhancement of genomic instability in minisatellite sequences of mouse C3H/10T1/2 cells transformed in vitro by X-rays. Cancer Res 1994;54:3173–8. Yuan J, Narayanan L, Rockwell S, Glazer PM. Diminished DNA repair and elevated mutagenesis in mammalian cells exposed to hypoxia and low pH. Cancer Res 2000;60:4372–6. Li CY, Little JB, Hu K, Zhang W, Zhang L, Dewhirst MW, Huang Q. Persistent genetic instability in cancer cells induced by non-DNAdamaging stress exposures. Cancer Res 2001;61:428–32. Oliveira AM, Nascimento AG. Soft tissues. In: Alison MR, editor. The cancer handbook. London: Nature Publishing Group, 2004. 733– 65. Almholt K, Johnsen M. Stromal cell involvement in cancer. Recent Results Cancer Res 2003;162:31–42. Mirsalis JC. Transgenic models for detection of mutations in tumors and normal tissues of rodents. Toxicol Lett 1995;82–83:131–4. Baross-Francis A, Andrew SE, Penney JE, Jirik FR. Tumors of DNA mismatch repair-deficient hosts exhibit dramatic increases in genomic instability. Proc Natl Acad Sci USA 1998;95:8739–43. Zhang S, Lloyd R, Bowden G, Glickman BW, de Boer JG. Thymic lymphomas arising in Msh2 deficient mice display a large increase in mutation frequency and an altered mutational spectrum. Mutat Res 2002;500:67–74. Baross-Francis A, Milhausen MK, Andrew SE, Jevon G, Jirik FR. Tumors arising in DNA mismatch repair-deficient mice show a wide variation in mutation frequency as assessed by a transgenic reporter gene. Carcinogenesis 2000;21:1259–62.
29. McDiarmid HM, Douglas GR, Coomber BL, Josephy PD. Epithelial and fibroblast cell lines cultured from the transgenic BigBlue rat: an in vitro mutagenesis assay. Mutat Res 2001;497:39–47. 30. Vaupel P, Schlenger K, Knoop C, Hockel M. Oxygenation of human tumors: evaluation of tissue oxygen distribution in breast cancers by computerized O2 tension measurements. Cancer Res 1991;51:3316– 22. 31. Parker C, Milosevic M, Toi A, Sweet J, Panzarella T, Bristow R, Catton C, Catton P, Crook J, Gospodarowicz M, McLean M, Warde P, et al. Polarographic electrode study of tumor oxygenation in clinically localized prostate cancer. Int J Radiat Oncol Biol Phys 2004;58: 750–7. 32. Kallinowski F, Vaupel P, Runkel S, Berg G, Fortmeyer HP, Baessler KH, Wagner K, Mueller-Klieser W, Walenta S. Glucose uptake, lactate release, ketone body turnover, metabolic micromilieu, and pH distributions in human breast cancer xenografts in nude rats. Cancer Res 1988;48:7264–72. 33. Hlatky L, Sachs RK, Alpen EL. Joint oxygen-glucose deprivation as the cause of necrosis in a tumor analog. J Cell Physiol 1988;134:167– 78. 34. Brown JM, Giaccia AJ. The unique physiology of solid tumors: opportunities (and problems) for cancer therapy. Cancer Res 1998;58:1408–16. 35. Mihaylova VT, Bindra RS, Yuan J, Campisi D, Narayanan L, Jensen R, Giordano F, Johnson RS, Rockwell S, Glazer PM. Decreased expression of the DNA mismatch repair gene Mlh1 under hypoxic stress in mammalian cells. Mol Cell Biol 2003;23:3265–73. 36. Papp-Szabo E, Douglas GR, Coomber BL, Josephy PD. Mutagenicity of the oral carcinogen 4-nitroquinoline-1-oxide in cultured BigBlue rat tongue epithelial cells and fibroblasts. Mutat Res 2003;522: 107–17. 37. McDiarmid HM, Douglas GR, Coomber BL, Josephy PD. 2-Amino1-methyl-6-phenylimidazo[4, 5-b]pyridine (PhIP)-induced mutagenesis in cultured BigBlueTM rat mammary epithelial and fibroblast cells. Environ Mol Mutagen 2002;39:245–53. 38. Mombaerts P, Iacomini J, Johnson RS, Herrup K, Tonegawa S, Papaioannou VE. RAG-1-deficient mice have no mature B and T lymphocytes. Cell 1992;68:869–77. 39. McDiarmid HM. An in vitro assay of chemical mutagenicity in rat mammary epithelial and fibroblast cells. PhD thesis. University of Guelph, Guelph, Ontario, Canada. 2003. 222 p. 40. Rofstad EK, Halsor EF. Hypoxia-associated spontaneous pulmonary metastasis in human melanoma xenografts: involvement of microvascular hot spots induced in hypoxic foci by interleukin 8. Br J Cancer 2002;86:301–8. 41. Peretz S, Kim C, Rockwell S, Baserga R, Glazer PM. IGF1 receptor expression protects against microenvironmental stress found in the solid tumor. Radiat Res 2002;158:174–80. 42. Postovit LM, Adams MA, Lash GE, Heaton JP, Graham CH. Oxygenmediated regulation of tumor cell invasiveness. Involvement of a nitric oxide signaling pathway. J Biol Chem 2002;277:35730–7. 43. Rofstad EK, Rasmussen H, Galappathi K, Mathiesen B, Nilsen K, Graff BA. Hypoxia promotes lymph node metastasis in human melanoma xenografts by up-regulating the urokinase-type plasminogen activator receptor. Cancer Res 2002;62:1847–53. 44. Kondo A, Safaei R, Mishima M, Niedner H, Lin X, Howell SB. Hypoxia-induced enrichment and mutagenesis of cells that have lost DNA mismatch repair. Cancer Res 2001;61:7603–7. 45. Heddle JA, Martus HJ, Douglas GR. Treatment and sampling protocols for transgenic mutation assays. Environ Mol Mutagen 2003;41:1–6. 46. Stuart GR, Oda Y, de Boer JG, Glickman BW. Mutation frequency and specificity with age in liver, bladder and brain of lacI transgenic mice. Genetics 2000;154:1291–300. 47. Wilkinson D, Sandhu JK, Breneman JW, Tucker JD, Birnboim HC. Hprt mutants in a transplantable murine tumour arise more frequently in vivo than in vitro. Br J Cancer 1995;72:1234–40. 48. Kobayashi H, Man S, MacDougall JR, Graham CH, Lu C, Kerbel RS. Variant sublines of early-stage human melanomas selected for tumorigenicity in nude mice express a multicytokine-resistant phenotype. Am J Pathol 1994;144:776–86. 49. Lev DC, Kiriakova G, Price JE. Selection of more aggressive variants of the gI101A human breast cancer cell line: a model for analyzing the metastatic phenotype of breast cancer. Clin Exp Metastasis 2003;20:515–23. 50. Rumpler G, Becker B, Hafner C, McClelland M, Stolz W, Landthaler M, Schmitt R, Bosserhoff A, Vogt T. Identification of differentially expressed genes in models of melanoma progression by cDNA array analysis: SPARC, MIF and a novel cathepsin protease characterize aggressive phenotypes. Exp Dermatol 2003;12:761–71.
MUTATION SPECTRA IN THE TUMOR ENVIRONMENT
51. Korczak B, Robson IB, Lamarche C, Bernstein A, Kerbel RS. Genetic tagging of tumor cells with retrovirus vectors: clonal analysis of tumor growth and metastasis in vivo. Mol Cell Biol 1988;8: 3143–9. 52. Rosenberg MS, Subramanian S, Kumar S. Patterns of transitional mutation biases within and among mammalian genomes. Mol Biol Evol 2003;20:988–93. 53. Watson DE, Cunningham ML, Tindall KR. Spontaneous and ENUTM induced mutation spectra at the cII locus in Big Blue Rat2 embryonic fibroblasts. Mutagenesis 1998;13:487–97. 54. Jorge SA, Menck CF, Sies H, Osborne MR, Phillips DH, Sarasin A, Stary A. Mutagenic fingerprint of ozone in human cells. DNA Repair (Amst) 2002;1:369–78. 55. Will O, Gocke E, Eckert I, Schulz I, Pflaum M, Mahler HC, Epe B. Oxidative DNA damage and mutations induced by a polar photosensitizer, Ro19-8022. Mutat Res 1999;435:89–101.
685
56. Jakubczak JL, Merlino G, French JE, Muller WJ, Paul B, Adhya S, Garges S. Analysis of genetic instability during mammary tumor progression using a novel selection-based assay for in vivo mutations in a bacteriophage lambda transgene target. Proc Natl Acad Sci USA 1996;93:9073–8. 57. Ehrlich M. DNA methylation in cancer: too much, but also too little. Oncogene 2002;21:5400–13. 58. Chen T, Gamboa dC, Marques MM, Shelton SD, Beland FA, Manjanatha MG. Mutations induced by alpha-hydroxytamoxifen in the lacI and cII genes of Big Blue transgenic rats. Carcinogenesis 2002;23: 1751–7. 59. Boyiri T, Guttenplan J, Khmelnitsky M, Kosinska W, Lin JM, Desai D, Amin S, Pittman B, El Bayoumy K. Mammary carcinogenesis and molecular analysis of in vivo cII gene mutations in the mammary tissue of female transgenic rats treated with the environmental pollutant 6-nitrochrysene. Carcinogenesis 2004;25:637–43.
