November 2001, Volume 1, Issue 2

Developing Methods of Genetic Analysis to Improve Cancer Risk Assessment

Barbara L. Parsons and Page B. McKinzie

Division of Genetic and Reproductive Toxicology
Food and Drug Administration, National Center for Toxicological Research, Jefferson, Arkansas 72079

Abstract:

Cancer risk assessment is currently based mainly on the two year rodent cancer bioassay. Advances in the scientific understanding of the genetic basis of cancer are now making it possible to improve risk assessment by obtaining and applying information about a chemical's mode of action. However, routine methods of monitoring the genetic damage that leads to cancer are not yet available. At the FDA National Center for Toxicological Research (NCTR), genotypic selection methods are being developed as an approach for obtaining such information. Genotypic selection refers to the DNA-based detection of rare mutation. Using genotypic selection, the oncogene and tumor suppressor gene mutations that are the earliest persistent changes in tumor development can be measured and used as biomarkers for cancer risk. Three different genotypic selection methods have been developed. The first assay (MutEx enrichment) uses the mismatch binding protein, MutS to shield mutant DNA while wild-type DNA is selectively degraded. The second assay is an allele-specific amplification technique called ACB-PCR. By coupling these two techniques, the third assay, MutEx/ACB-PCR was developed. This assay has the sensitivity to detect mutant DNA in the presence of a 10⁷-fold excess of wild-type DNA.

Introduction

Acquiring the capability to precisely assess the risk associated with chemical exposure is vitally important in terms of human health. Many different public health issues revolve around whether or not a chemical exposure causes cancer. In terms of drugs, for instance, the therapeutic value of a particular drug must be weighed against that drug’s potential to induce cancer. In this context, precision in assessing cancer risk could translate into better healthcare decision making. In terms of potentially carcinogenic environmental exposures, the wisdom of reducing or eliminating an exposure can only be judged if the cancer risks associated with the different levels of exposure are reliably known. The approaches used to show that a chemical is, or is not, a cancer hazard must be robust and credible if they are to attach public confidence. Unfortunately, our risk assessment capabilities are not yet viewed as having the level of scientific rigor warranted by the degree to which cancer risk assessment impacts human health (1).

Many of the regulatory responsibilities of the US Food and Drug Administration (FDA) involve assessing the cancer risk associated with chemical exposures (2). Of the five FDA product centers, three of them have regulatory responsibilities that involve cancer risk assessment. The Center for Food Safety and Applied Nutrition (CFSAN) is responsible for determining whether food additives and contaminants of food additives are carcinogenic. The Center for Veterinary Medicine (CVM) is responsible for ensuring that veterinary drugs used in food-producing animals do not leave carcinogenic food residues. The Center for Drug Evaluation and Research (CDER) must determine whether new drugs entering the drug approval process are carcinogenic. Clearly, improving available methods for assessing chemical associated cancer risk is an issue of great importance to the FDA. Because the National Center for Toxicological Research (NCTR) performs fundamental research to support the regulatory needs of the other FDA centers, an important part of the NCTR’s mission is to develop, characterize, and validate new cancer risk assessment approaches.

Current Cancer Risk Assessment Practices

Cancer risk assessment can be described as a three-step process: hazard identification, carcinogenicity determination, and dose-response assessment. In terms of hazard identification, the genotoxicity of a chemical is examined using a battery of tests. Because chemicals that cause DNA damage can be carcinogens, these tests identify chemicals that cause various types of DNA damage, such as point mutations, DNA strand breaks, loss of heterozygosity, or aneuploidy. The test battery includes a bacterial gene mutation assay, an in vivo test for chromosome damage in rodent hematopoietic cells, and either an in vitro cytogenetic test with mammalian cells or the in vitro mouse lymphoma mutation assay that selects for inactivation or loss of the thymidine kinase gene (3). Although short-term transgenic cancer assays are presently being assessed, at the present time carcinogenicity is evaluated using the two species rodent tumor bioassay. In a classic lifetime tumor bioassay, the maximum tolerated dose (MTD) for a particular chemical is determined in a dose-range finding study, then doses close to the MTD are administered to male and female rats and mice for two years (4). A statistically significant, chemical-associated increase in tumor count is taken as evidence of carcinogenicity. Next, a dose-response assessment is used to determine whether there is any dose of the chemical that is likely to be safe for humans. Using a particular mathematical approach, an extrapolation is made from the dose(s) that gave a reliable tumor response to the generally much lower doses relevant to human exposures. For cancer risk assessment, the conservative approach of a linear low dose extrapolation is used (5). Ultimately, the regulatory action taken based on the genotoxicity test battery, the tumor outcome, and the dose-response assessment depends in large part on the intended use of the chemical.

The rodent tumor bioassay has been used for carcinogenicity testing for many years, primarily for four reasons. First, the rodent tumor bioassay fulfills an obvious need for an experimental system. Treating rodents with largely uncharacterized chemicals is a necessary step in drug development; it assures some minimum level of safety even in the use of experimental drugs. Second, in vivo exposure and tumor development are considered more relevant endpoints for assessing human cancer risk than in vitro tests with bacterial or mammalian cells or non-tumor endpoints. Third, the rodent tumor bioassay has been used for many years without disastrous human health consequences. And finally, at the present time there is no other "ideal" assay with which to replace the rodent tumor bioassay.

While the strengths of the rodent tumor bioassay are clear, its drawbacks are equally well known. The use of the rodent tumor bioassay has led to quantitative risk assessment based on tumor counts. This is actually quite problematic because high doses of chemical are needed to induce statistically significant tumor responses in treated rodents (4). This in turn necessitates extrapolating from the measurable tumor response down to the doses relevant to humans but for which there is no actual tumor data (2). Treatment of rodents at the levels of exposure relevant to humans is not done because the absence of an induced tumor response gives no assurance of safety in humans, particularly considering the relatively high background tumor rate of the tumor-sensitive rodent strains used. In other words, the rodent tumor response is not sensitive enough to detect low-dose effects. At the same time, high doses of chemical often lead to a cytotoxic effect in animal tissues. This cytotoxicity may lead to cell proliferation and other promoter-like events (6-8). Many "high-dose" rodent carcinogens have been identified, but the relevance of those findings to human disease is unclear.

A basic premise of risk assessment has been that genotoxic and non-genotoxic chemical carcinogens should be regulated differently. Genotoxic carcinogens are those that interact with DNA and cause DNA damage, while non-genotoxic carcinogens operate through other mechanisms. Although challenged recently, it has long been assumed that if a chemical is genotoxic, then even a single molecule of that chemical might cause a tumor and, therefore, linear low dose extrapolation is warranted (5,9). Carcinogens that have non-genotoxic modes of action may be considered to have a threshold below which the chemical is non-carcinogenic. The genotoxicity test battery gives an indication of a chemical’s genotoxicity; but there is currently no clear-cut approach for determining whether a chemical is operating through a genotoxic or non-genotoxic mode in vivo and at low doses. Consequently, the inability to determine when linear low dose extrapolation is warranted, when a different type of dose response extrapolation is warranted (J- or U-shaped), or if a chemical’s effects have a threshold are major deficiencies in the science of cancer risk assessment (10,11).

After the dose-response assessment has been completed, the rodent tumor data must be used to estimate human risk and/or assign a safe level. The rodent data are first transformed using a scaling factor to account for the gross difference in body weight between rodents and humans. Then safety margins or uncertainty factors may be invoked. These are usually ten-fold reductions in allowable exposures that are considered necessary given the lack of scientific rigor in species extrapolation or because of the possible existence of sensitive sub-populations (12). Thus, the lack of scientific rigor in rodent to human extrapolation is another area of risk assessment that could be improved.

How to Improve Cancer Risk Assessment

A generally accepted strategy for improving cancer risk is to incorporate chemical-specific, mechanistic, and biological information. The logic behind this strategy is that specific mechanistic information will define a chemical’s mode of action; information that is essential if "biological realism" is to be incorporated into a scientific judgement of risk (1,13,14). Operationally, the idea is to identify biomarkers that are relevant to mode-of-action and can be used to support a particular low-dose extrapolation method or determine if a threshold exists and relate what takes place in a rodent to the biology of a human. While there is consensus in terms of using this general strategy to improve cancer risk assessment, there is little consensus regarding what specific information should be sought to determine mode-of-action or what biomarkers would be most informative for assessing cancer risk.

In theory, the ideal biomarker(s) to use in evaluating cancer risk would have several characteristics. First, the biomarker should have a direct relationship to cancer; the more direct the relationship between the biomarker and tumor development the more credible that biomarker will be in predicting cancer risk (see Figure 1 following this paragraph). Second, the ideal biomarker would be measurable and have the same relationship to tumor development in both rodents and humans: such circumstances would facilitate translating the cancer risk defined in an experimental rodent model to human risk. And third, a biomarker that appears relatively early in tumor development will be more useful than one that appears later because of the possibility of interceding in tumor development.

Figure 1.  The potential usefulness of different types of biomarkers for cancer risk assessment.

At the NCTR a variety of different endpoints are being developed and evaluated for their potential use in predicting chemical associated cancer risk. These include measurement of DNA adducts, mutation in endogenous or transgenic reporter genes, and the identification of chemically induced changes in gene expression using gene arrays (15-17). Another promising approach for improving chemical risk assessment that is being developed at NCTR is genotypic selection.

Genotypic selection refers to the DNA-based detection of specific, rare mutations. The potential strength of applying genotypic selection to cancer risk assessment is that the actual oncogene and tumor suppressor gene mutations known to be involved in cancer causation could be measured. Figure 1 (see figure above the last paragraph) illustrates why the measurement of oncogene and tumor suppressor gene mutations may represent the ideal metric to use for cancer risk assessment. Such point mutations are known to occur early in the tumorigenic process; they are the initiating mutations detected in tumors and sometimes in pre-neoplastic lesions. Most importantly, these mutations themselves are directly and mechanistically related to carcinogenesis; i.e., cells accumulate these permanent genetic changes that confer the new phenotypes and capabilities of a tumor cell (18). Developing genotypic selection methods to detect these specific and rare mutational events is technically challenging. Also, the idea of using the measurement of specific oncogene or tumor suppressor gene mutations as biomarkers of cancer risk is relatively new. Consequently, the NCTR is taking the lead in investigating the potential use of genotypic selection methods for improving cancer risk assessment.

Genotypic Selection Methods

Genotypic selection can be used to detect a variety of mutations including translocation, deletion, expansion of microsatellites, and point mutation. The value of techniques that can detect specific mutations in pools of DNA molecules has been recognized for almost 20 years. The first genotypic selection methods were developed in order to detect oncogene mutations in tumor tissues. More recently, genotypic selection has also been used in the diagnosis and management of cancer patients, for the identification of specific microbes, and in pool screening for genetic polymorphisms (19).

A variety of different molecular approaches have been developed for the detection of rare point mutation. These approaches span a large range in terms of their sensitivity. In the context of genotypic selection, sensitivity refers to the level of mutant DNA sequence that can be detected amongst wild-type DNA; the lower the mutant fraction detected by a particular genotypic selection, the more sensitive the assay. The sensitivity of a genotypic selection method determines the applications for which they will be useful (19). Measurement of the rare chemically induced mutations in oncogene and tumor suppressor gene targets is an application that requires a great deal of sensitivity. Fortunately, the development of new tools and methodologies has improved the sensitivity of genotypic selection to a point where the potential use of these methods for cancer risk assessment has become an important avenue of research.

Development of Genotypic Selection Methods in the Division of Genetic and Reproductive Toxicology at the National Center for Toxicological Research

Information from a number of different research areas had to be analyzed before this research could proceed in an effective manner. Specifically, a theoretical framework was needed in terms of what experimental designs should be used (see Figure 2, which follows this paragraph), what mutational targets should be investigated, and what sensitivity might be required. The necessary information was recently assembled in the form of a review article (20). It was concluded that eventually a small battery of genotypic selection assays should be developed. Initially, development of genotypic selection methods should focus on the most common mutations that occur in human tumors, and it should be determined empirically which of those behave similarly between rodent and human. The most important mutational specificities to include in such a battery are G:C to T:A, G:C to A:T, and A:T to T:A because analysis of 31 different carcinogens showed that the mutagenic effects of 29 of these were encompassed by these three mutational specificities. It was also concluded that the information provided by genotypic selection would be most valuable for risk assessment if the ability to detect background levels of spontaneous mutation was possible (see Figure 2, which follows this paragraph). It was determined that a spontaneous mutation frequency of ~10^-7 per basepair is expected based on data from phenotypic selection assays (spontaneous mutation frequency per locus divided by target size in basepairs gives a value of ~10^-7 per basepair). This information was used to define the goal of this research; to develop a genotypic selection method that could detect mutant DNA sequence in the presence of a 10⁷-fold excess of wild-type DNA sequence.

Figure 2.  Experimental strategies for applying genotypic selection methods to cancer risk assessment.

Identifying the most useful mutational targets is a key issue in successfully applying genotypic selection to cancer risk assessment and is the first step in assay development. Ras was identified as a prototype for oncogene targets. The ras family of proteins (H-ras, N-ras, and K-ras) function as molecular switches in signal transduction pathways. The ras protein is a small (21 kDa) GTPase, which is in its active conformation when bound to GTP. The most common mutations in the ras gene result in protein species that bind GTP, but not GDP, and are, therefore, continuously active. Constitutive ras activity can cause continuous cell proliferation via the raf-MAPK pathway (see Figure 3, following this paragraph). However, this is not the only pathway affected by ras activity. It is becoming clear that there is extensive cross-talk between GTPase signaling pathways. Thus, the effect of continuous ras activity will depend on cell background and sometimes results in apoptosis.

Figure 3. Consequences of ras activation are dependent on cell-type specific signal transduction effector molecules.  Two different outcomes of ras activation are depicted.  In one pathway, active ras recruits raf to the membrane, facilitating the phosphorylation (activation) of raf.  Raf then phosphorylates MEK (mitogen activated and extracellular response kinase), which phosphorylates ERK (extracellular response kinase), which translocates into the nucleus where it can activate transcription factors to increase cyclin D1 levels in the cell. Cyclin D1 (with its cofactors) phosphorylates pRB, which releases E2F, allowing E2F regulated proteins to be transcribed.  At this point, the cell is committed to the S phase and cell proliferation (44).  However, ras can also activate the expression of p16ARF (alternative reading frame), which is considered a tumor suppressor.  P16ARF indirectly affects the activity of p53 by sequestering Mdm2 (an inhibitor of p53).   In this case, the ras initiated pathway can lead to apoptosis (45).

Ras mutations frequently occur in both spontaneous and chemically induced rodent tumors and are localized to a few specific DNA regions, codons 12, 13, and 61 (21). Similar patterns of ras mutation are present in a number of different human tumor types. Ras mutation is found in 90% of pancreatic tumors, 50% of colon and thyroid tumors, and 30% of lung tumors and leukemias (22). The frequencies with which the most commonly occurring ras gene base substitutions have been detected in human tumors are given in Figure 4 (following this paragraph) (22-38). Thus, ras mutations are valuable targets to use in the development and application of genotypic selection methods.

Figure 4. Frequency of the most common ras mutations in human tumors.   The frequency of each particular basepair substitution mutation was plotted.   Only mutations that were detectable in less than or equal to 2% of the tumors of a particular tissue origin are included.

MutEx and ACB-PCR Genotypic Selection

Some of the most powerful tools that have been used for genotypic selection are restriction enzymes. Digestion of DNA, with a restriction enzyme is used to selectively destroy wild-type DNA sequences. This effectively enriches for sequences not digested because they carried a mutation in the restriction enzyme cleavage site (19). A report by Ellis et al. suggested that the E. coli mismatch binding protein, MutS could be used similarly as a tool for the selective destruction of wild-type DNA sequence while avoiding the limitation of only being able to analyze restriction enzyme cleavage sites (39). In their "MutEx" assay, PCR products were synthesized, denatured, and reannealed; thereby creating heteroduplex molecules in the DNA from individuals that were heterozygous for a germline mutation. The E. coli MutS protein was incubated with the PCR products and the 3’ - 5’ exonuclease activity of T7 DNA polymerase was used to digest the heteroduplex DNA. The bound MutS protein blocked this digestion so that the length of the protected DNA fragment defined the position of the germline mutation in the PCR product being analyzed. Because homoduplex DNA would be degraded in such an assay, it was realized that this approach might also be used for genotypic selection; to selectively degrade a large excess of wild-type sequence while preserving mutant sequence. Consequently, this type of MutS selection was the basis for the first genotypic selection method developed at the NCTR.

The H-ras codon 61 CAA to AAA mutation was used as the model system in the development of the MutEx approach as a genotypic selection method. This mutation was selected because: 1) It is the most frequent mutation detected in mouse liver tumors, 2) its occurrence in tumors can be increased by chemical treatment, and 3) mouse strain differences in the frequency of this mutation might eventually be used for method validation (21). Therefore, mutant and wild-type mouse H-ras sequences were cloned and restriction fragments corresponding to each were isolated and quantified. These restriction fragments were used in reconstruction experiments; meaning that DNA mixtures with known mutant fractions were prepared and used in the analysis of different experimental procedures. The molecular events occurring in each step of the MutEx genotypic selection that was developed are depicted in Figure 5 (following this paragraph). The result of this genotypic selection is that a large proportion of the wild-type DNA molecules is destroyed while the mutant sequences are preserved. Mutant sequence was then detected using single nucleotide primer extension (SNuPE) (40). In SNuPE, the extension of a primer adjacent to the base being interrogated is carried out in the presence of a single nucleotide complementary to either the mutant or wild-type base. Using this approach it was determined that mutant fractions between 0.5 and 2 x 10^-5 were detectable (41). In addition, it was determined that SNuPE alone had a sensitivity of less than or equal to 2 x 10^-2. From this information it was concluded that the MutEx assay was providing an ~1,000-fold enrichment of mutant DNA sequences.

Figure 5. The MutEx/SNuPE genotypic selection method.

The second genotypic selection method that was developed at the NCTR was based on a completely different type of selection, allele-specific amplification. In an allele-specific amplification, a PCR primer that has more mismatches to the wild-type sequence than the mutant sequence is used to selectively amplify mutant DNA (19). Allele-specific competitive blocker PCR (ACB-PCR) is an allele-specific amplification method that was reported to have a sensitivity of 10^-4 (42). The assay uses three different primers, a mutant-specific primer that amplifies the mutant sequence, a blocker primer that obstructs PCR amplification from the wild-type sequence, and an upstream PCR primer (see Figure 6 following this paragraph). At the NCTR, this approach was adapted to the detection of the H-ras codon 61 CAA to AAA mutation (43). The assay was modified in a number of ways, including the use of the Stoeffel fragment of Taq DNA polymerase and PerfectMatch PCR Enhancer. These modifications resulted in an increase in the assay sensitivity with mutant fractions as low as 10^-5 being detectable.

Figure 6. Primer design used in allele-specific competitive blocker PCR (ACB-PCR). Three PCR primers are shown: the blocker primer (BP), mutant specific primer (MSP), and upstream primer (UP). The selective annealing of the MSP to mutant sequence and BP to wild-type sequence is depicted. These primer-template pairings result in single 3'-penultimate mismatches. These pairs are favored over annealing of the MSP to wild-type template or BP to mutant template, which would result in double 3'-terminal mismatches.  The blocker primer carries a 3'-terminal dideoxy nucleotide and cannot be extended.

Keeping in mind that the goal of this work was to develop an assay that could detect spontaneous mutation (estimated at 10^-7) neither the MutEx/SNuPE assay nor ACB-PCR alone had sufficient sensitivity. In an attempt to reach this sensitivity, the relatively insensitive mutation detection step of the MutEx/SNuPE assay was replaced by ACB-PCR. In other words, the MutEx mutant DNA enrichment was coupled with the sensitive ACB-PCR mutation detection method. This combined assay, named MutEx/ACB-PCR, was found to have a sensitivity of 10^-7 in reconstruction experiments (see Figure 7 following this paragraph) (40). As a means of validating the use of this assay in the measurement of very low mutant fractions, the level of H-ras mutation induced by Pfu DNA polymerase during PCR amplification was determined. Pfu DNA polymerase was selected because it has the highest fidelity in replicating DNA sequences of any known thermostable polymerase and the most common error it produces is C to A transversion. The results from three replicate MutEx/ACB-PCR experiments measured the Pfu DNA polymerase-generated mutant fractions as 10 ± 3 x 10^-7, from which a polymerase error rate of 8 ± 3 x 10^-7 was calculated (40). This value is in good agreement with the published reports regarding Pfu DNA polymerase error rate and, therefore, substantiates the accuracy of the MutEx/ACB-PCR assay in the measurement of low mutant fractions.

Figure 7. The MutEx/ACB-PCR assay developed at NCTR has a sensitivity of 10^-7. In the reconstruction experiment shown, each reaction contained 300 nanograms of genomic DNA and 3 x 10⁷ copies of wild-type H-ras restriction fragment. Addition of different amounts of mutant restriction fragment was used to generate the mutant fraction standards analyzed (10^-4 – 10^-7). The signal in the 10^-7 lanes corresponds to the detection of three mutant H-ras molecules in the presence of 3 x 10⁷ wild-type alleles.

Summary

Genotypic selection is being developed as an approach for improving chemical risk assessment. This is primarily because the biomarkers that will be measured by genotypic selection, oncogene and tumor suppressor gene mutations, have a direct relationship to cancer. At the NCTR, work toward this goal has proceeded in two areas: 1) identifying the theoretical issues that should be considered in the development of new assays and 2) the assay development itself. For example, evidence that an average spontaneous mutation frequency of 10^-7 might be expected was important for setting a goal for assay sensitivity. That goal, the detection of mutant allele in the presence of a 10⁷-fold excess of wild-type allele is now achievable using the MutEx/ACB-PCR assay. However, additional challenges remain. In order to take advantage of MutEx/ACB-PCR sensitivity, a pool containing >10⁷ molecules must be analyzed. This corresponds to a genomic DNA sample of >100 micrograms, a mass that would inhibit the sensitive MutEx/ACB-PCR approach. PCR amplification of target DNA cannot be used to generate the necessary DNA pool because the error rate of even the most reliable thermostable polymerase is higher than the level of mutation that needs to be detected (10^-7). Thus, the development of gene-specific enrichment techniques is viewed as necessary, and development of such techniques is currently underway at the NCTR. Ultimately, the measurement of a small battery of oncogene and tumor suppressor gene mutations will be necessary to understand chemical-specific effects. Therefore, the genotypic selection assays already developed are being adapted to new mutational targets, namely human and rodent K-ras mutations. Eventually, the information provided by these sensitive assays should support a more scientifically rigorous approach to cancer risk assessment.

Acknowledgements

We wish to thank Martha Moore and Robert Heflich for their helpful suggestions in preparing this manuscript.

References

1. Conolly, R. B., Beck, B. D., and Goodman, J. I. 1999. Stimulating research to improve the scientific basis of risk assessment. Toxicological Sciences 49:1-4.

2. Gaylor, D. W., Axelrad, J. A., Brown, R. P., Cavagnaro, J. A., Cyr, W. H., Hulebak, K. L., Lorentzen, R. J., Miller, M. A., Mulligan, L. T., and Schwetz, B. A. 1997. Health risk assessment practices in the U.S. Food and Drug Administration. Regulatory Toxicology & Pharmacology 26:307-21.

3. Zeiger, E. 1998. Identification of rodent carcinogens and noncarcinogens using genetic toxicity tests: premises, promises, and performance. Regulatory Toxicology & Pharmacology 28:85-95.

4. Gold, L. S., Slone, T. H., and Ames, B. N. 1998. What do animal cancer tests tell us about human cancer risk?: Overview of analyses of the carcinogenic potency database. Drug Metabolism Reviews 30:359-404.

5. Golden, R. J., Holm, S. E., Robinson, D. E., Julkunen, P. H., and Reese, E. A. 1997. Chloroform mode of action: implications for cancer risk assessment. Regulatory Toxicology & Pharmacology 26:142-55.

6. Ames, B. N. and Gold, L. S. 1990. Too many rodent carcinogens: mitogenesis increases mutagenesis. Science 249:970-1.

7. Goodman, J. I. 1998. The traditional toxicologic paradigm is correct: dose influences mechanism. Environmental Health Perspectives 106:285-8.

8. Preston-Martin, S., Pike, M. C., Ross, R. K., Jones, P. A., and Henderson, B. E. 1990. Increased cell division as a cause of human cancer. Cancer Research 50:7415-21.

9. Lutz, W. K. and Kopp-Schneider, A. 1999. Threshold dose response for tumor induction by genotoxic carcinogens modeled via cell-cycle delay. Toxicological Sciences 49:110-5.

10. Sielken, R. L., Bretzlaff, R. S., and Stevenson, D. E. 1995. Challenges to default assumptions stimulate comprehensive realism as a new tier in quantitative cancer risk assessment. Regulatory Toxicology & Pharmacology 21:270-80.

11. Lutz, W. K. 2000. A true threshold dose in chemical carcinogenesis cannot be defined for a population, irrespective of the mode of action. Human & Experimental Toxicology 19:566-8; discussion 571-2.

12. Gaylor, D. W., Chen, J. J., and Sheehan, D. M. 1993. Uncertainty in cancer risk estimates. Risk Analysis 13:149-54.

13. Dellarco, V. L. and Wiltse, J. A. 1998. US Environmental Protection Agency's revised guidelines for Carcinogen Risk Assessment: incorporating mode of action data. Mutation Research 405:273-7.

14. Andersen, M. E., Meek, M. E., Boorman, G. A., Brusick, D. J., Cohen, S. M., Dragan, Y. P., Frederick, C. B., Goodman, J. I., Hard, G. C., O'Flaherty, E. J., and Robinson, D. E. 2000. Lessons learned in applying the U.S. EPA proposed cancer guidelines to specific compounds. Toxicological Sciences 53:159-72.

15. Beland, F. A., Hamilton, L. P., and Marques, M. M. 1999. Comparison of the DNA adducts formed by tamoxifen and 4-hydroxytamoxifen in vivo. Carcinogenesis 20:471-477.

16. Casciano, D. A., Aidoo, A., Chen, T., Mittelstaedt, R. A., Manjanatha, M. G., and Heflich, R. H. 1999. Hprt mutant frequency and molecular analysis of Hprt mutations in rats treated with mutagenic carcinogens. Mutation Research 431:389-95.

17. Dobrovolsky, V. N., Casciano, D. A., and Heflich, R. H. 1999. Tk+/- mouse model for detecting in vivo mutation in an endogenous, autosomal gene. Mutation Research 423:125-36.

18. Hanahan, D. and Weinberg, R. A. 2000. The hallmarks of cancer. Cell 100:57-70.

19. Parsons, B. L. and Heflich, R. H. 1997. Genotypic selection methods for the direct analysis of point mutations. Mutation Research 387:97-121.

20. McKinzie, P. B., Delongchamp, R. R., Heflich, R. H., and Parsons, B. L. 2001. Prospects for applying genotypic selection of somatic oncomutation to chemical risk assessment. Mutation Research in press.

21. Maronpot, R. R., Fox, T., Malarkey, D. E., and Goldsworthy, T. L. 1995. Mutations in the ras proto-oncogene: clues to etiology and molecular pathogenesis of mouse liver tumors. Toxicology 101:125-56.

22. Bos, J. L. 1989. ras oncogenes in human cancer: a review. Cancer Research 49:4682-9.

23. Ooi, A., Herz, F., Ii, S., Cordon-Cardo, C., Fradet, Y., and Mayall, B. H. 1994. Ha-ras codon 12 mutation in papillary tumors of the urinary bladder. A retrospective study. International Journal of Oncology 4:85-90.

24. Knowles, M. A. and Williamson, M. 1993. Mutation of H-ras is infrequent in bladder cancer: confirmation by single-strand conformation polymorphism analysis, designed restriction fragment length polymorphisms, and direct sequencing. Cancer Research 53:133-9.

25. Czerniak, B., Cohen, G. L., Etkind, P., Deitch, D., Simmons, H., Herz, F., and Koss, L. G. 1992. Concurrent mutations of coding and regulatory sequences of the Ha-ras gene in urinary bladder carcinomas. Human Pathology 23:1199-204.

26. Czerniak, B., Deitch, D., Simmons, H., Etkind, P., Herz, F., and Koss, L. G. 1990. Ha-ras gene codon 12 mutation and DNA ploidy in urinary bladder carcinoma. British Journal of Cancer 62:762-3.

27. Zhang, H., Nordenskjold, B., Dufmats, M., Soderkvist, P., and Sun, X. F. 1998. K-ras mutations in colorectal adenocarcinomas and neighbouring transitional mucosa. European Journal of Cancer 34:2053-7.

28. el Sebai, H., Ged, C., Bonichon, F., de Verneuil, H., and Longy, M. 1998. Genetic alterations in colorectal cancer, comparative analysis of deletion events, and point mutations. Cancer Genetics & Cytogenetics 104:32-8.

29. Smit, V. T., Boot, A. J., Smits, A. M., Fleuren, G. J., Cornelisse, C. J., and Bos, J. L. 1988. KRAS codon 12 mutations occur very frequently in pancreatic adenocarcinomas. Nucleic Acids Research 16:7773-82.

30. Keohavong, P., DeMichele, M. A., Melacrinos, A. C., Landreneau, R. J., Weyant, R. J., and Siegfried, J. M. 1996. Detection of K-ras mutations in lung carcinomas: relationship to prognosis. Clinical Cancer Research 2:411-8.

31. Vachtenheim, J., Horakova, I., Novotna, H., Opaalka, P., and Roubkova, H. 1995. Mutations of K-ras oncogene and absence of H-ras mutations in squamous cell carcinomas of the lung. Clinical Cancer Research 1:359-65.

32. Hruban, R. H., van Mansfeld, A. D., Offerhaus, G. J., van Weering, D. H., Allison, D. C., Goodman, S. N., Kensler, T. W., Bose, K. K., Cameron, J. L., and Bos, J. L. 1993. K-ras oncogene activation in adenocarcinoma of the human pancreas. A study of 82 carcinomas using a combination of mutant-enriched polymerase chain reaction analysis and allele-specific oligonucleotide hybridization. American Journal of Pathology 143:545-54.

33. Grunewald, K., Lyons, J., Frohlich, A., Feichtinger, H., Weger, R. A., Schwab, G., Janssen, J. W., and Bartram, C. R. 1989. High frequency of Ki-ras codon 12 mutations in pancreatic adenocarcinomas. International Journal of Cancer 43:1037-41.

34. Kondo, H., Sugano, K., Fukayama, N., Kyogoku, A., Nose, H., Shimada, K., Ohkura, H., Ohtsu, A., Yoshida, S., and Shimosato, Y. 1994. Detection of point mutations in the K-ras oncogene at codon 12 in pure pancreatic juice for diagnosis of pancreatic carcinoma. Cancer 73:1589-94.

35. Tamagawa, E., Ueda, M., Takahashi, S., Sugano, K., Uematsu, S., Mukai, M., Ogata, Y., and Kitajima, M. 1997. Pancreatic lymph nodal and plexus micrometastases detected by enriched polymerase chain reaction and nonradioisotopic single-strand conformation polymorphism analysis: a new predictive factor for recurrent pancreatic carcinoma. Clinical Cancer Research 3:2143-9.

36. Enomoto, T., Inoue, M., Perantoni, A. O., Buzard, G. S., Miki, H., Tanizawa, O., and Rice, J. M. 1991. K-ras activation in premalignant and malignant epithelial lesions of the human uterus. Cancer Research 51:5308-14.

37. Enomoto, T., Inoue, M., Perantoni, A. O., Terakawa, N., Tanizawa, O., and Rice, J. M. 1990. K-ras activation in neoplasms of the human female reproductive tract. Cancer Research 50:6139-45.

38. Mizuuchi, H., Nasim, S., Kudo, R., Silverberg, S. G., Greenhouse, S., and Garrett, C. T. 1992. Clinical implications of K-ras mutations in malignant epithelial tumors of the endometrium. Cancer Research 52:2777-81.

39. Ellis, L. A., Taylor, G. R., Banks, R., and Baumberg, S. 1994. MutS binding protects heteroduplex DNA from exonuclease digestion in vitro: a simple method for detecting mutations. Nucleic Acids Research 22:2710-1.

40. Parsons, B. L. and Heflich, R. H. 1998. Detection of basepair substitution mutation at a frequency of 1 x 10(-7) by combining two genotypic selection methods, MutEx enrichment and allele-specific competitive blocker PCR. Environmental & Molecular Mutagenesis 32:200-11.

41. Parsons, B. L. and Heflich, R. H. 1997. Evaluation of MutS as a tool for direct measurement of point mutations in genomic DNA. Mutation Research 374:277-85.

42. Orou, A., Fechner, B., Utermann, G., and Menzel, H. J. 1995. Allele-specific competitive blocker PCR: a one-step method with applicability to pool screening. Human Mutation 6:163-9.

43. Parsons, B. L. and Heflich, R. H. 1998. Detection of a mouse H-ras codon 61 mutation using a modified allele-specific competitive blocker PCR genotypic selection method. Mutagenesis 13:581-8.

44. Pincus, M. R., Brandt-Rauf, P. W., Michl, J., Carty, R. P., and Friedman, F. K. 2000. ras-p21-induced cell transformation: unique signal transduction pathways and implications for the design of new chemotherapeutic agents. Cancer Investigation 18:39-50.

45. Sherr, C. J. 2000. The Pezcoller lecture: cancer cell cycles revisited. Cancer Research 60:3689-95.

The Authors
Pictured left to right: 
Page B. McKinzie, Ph.D. and Barbara L. Parsons, Ph.D.,
September 18, 2001
(NCTR Photo: Virginia B. Taylor)

Barbara L. Parsons is a FDA Staff Fellow in the Division of Genetic and Reproductive Toxicology at the National Center for Toxicological Research (NCTR), Jefferson, Arkansas. Dr. Parsons began her scientific career as a technician at Cold Spring Harbor Laboratory, Cold Spring Harbor, NY where she was involved in sequencing the Adenovirus genome. She entered the Department of Microbiology and Immunology and Interdisciplinary Program in Genetics at Duke University in 1982 and received her Ph.D. in 1988. During this time, her research was focused on animal virology; investigating the structure and function of Orthopoxvirus telomeres. Her first post-doctoral position was at the Beltsville Agricultural Research Center in Beltsville, Maryland where she studied changes in tomato fruit gene expression induced by wounding and the plant hormone, ethylene. Dr. Parsons began her work at the NCTR in 1994 when she was hired as a post-doctoral fellow through the Oak Ridge Institute for Science and Engineering (ORISE); work she continues now as an FDA Staff Fellow. Dr. Parsons was recently elected as a Councilor of the Environmental Mutagen Society. Her research interests lie in the development of DNA-based mutation detection methods and their application to cancer risk assessment.

Page B. McKinzie is an ORISE post-doctoral fellow in the Division of Genetic and Reproductive Toxicology at NCTR. Dr. McKinzie entered the graduate program of the Department of Biochemistry and Molecular Genetics in 1987 at the University of Alabama at Birmingham (UAB) and received her Ph.D. degree in 1993. Her graduate work was on the use of liposomes that are sensitive to low pH as a delivery vehicle for superoxide dismutase into fetal lung epithelial cells as an approach for relieving symptoms of bronchopulmonary dysplasia. Her subsequent post-doctoral position was with the Gene Therapy Program at UAB, with work focusing on: 1) using single-chain antibodies to knock out protein functions associated with carcinogenesis and 2) using replication deficient adenovirus as a vehicle for delivery of DNA into HPV-18 infected cells, with the goal of killing those cells. She began her work with the NCTR in 1999 and is currently adapting the DNA-based mutation detection methods developed at the NCTR to the detection of human and rat K-ras mutations.

Glossary

• allele – one of a series of possible forms of a gene that differ in DNA sequence
• aneuploidy – having an abnormal set of chromosomes
• deletion – loss of a segment of DNA
• expansion of microsattelites – an increase in the length of a cluster of repetitive DNA where each repeating unit is a particular DNA sequence and the increase in length corresponds to the addition of repeat units
• loss of heterozygosity – only a single allele of a gene is present in a cell when previously two distinguishable alleles had been present
• mutation – a heritable change in the DNA sequence of a gene
• oncogene – a gene that causes uncontrolled cell proliferation
• PCR primer – a short segment of single stranded DNA that binds to a longer, complementary DNA strand. A PCR primer is used to direct the in vitro replication of DNA that occurs in a Polymerase Chain Reaction (PCR)
• phenotype – a manifest trait or ability exhibited by a cell or organism that is determined by that cell or organism’s genes
• restriction enzymes –proteins that cleave DNA at positions determined by the specific order or sequence of DNA bases
• tumor suppressor gene – a gene that controls normal growth and cell division; loss or inactivation of such genes are known to contribute to tumor development
• translocation – transfer of a segment of one chromosome to a different position within the same chromosome or to a different chromosome

Regulatory Research Perspectives
Editorial Board

Norris Alderson, Ph.D. - Office of the Commissioner (OC)
Daniel A. Casciano, Ph.D. – National Center for Toxicological Research (NCTR)
Thomas A. Cebula, Ph.D. – Center for Food Safety and Applied Nutrition (CFSAN)
Lireka P. Joseph, Ph.D. – Center for Devices and Radiological Health (CDRH)
Joanne N. Locke – Office of the Commissioner (OC)
Edward E. Max, Ph.D. – Center for Biologics Evaluation and Research (CBER)
Michael C. Olson – Office of Regulatory Affairs (ORA)
Frank D. Sistare, Ph.D. – Center for Drug Evaluation and Research (CDER)
Mary S. Wolfe, Ph.D. – National Institute of Environmental Health Science (NIEHS)
Linda D. Youngman, Ph.D. – Center for Veterinary Medicine (CVM)
Hal Zenick , Ph.D. – Environmental Protection Agency (EPA)

Note: New members from ATSDR and NIOSH will join the Regulatory Research Perspectives Editorial Board in future issues.

Editorial Matters:

Address for editorial matters:   
DHHS/FDA/Jefferson Labs
National Center for Toxicological Research
3900 NCTR Road, HFT-1
Jefferson, Arkansas 72079-9502
Telephone: (870) 543-7516
web site: http://www.fda.gov/nctr
Article may be republished without permission.  Credit to Regulatory Research Perspectives as the source is appreciated.
Managing Editor/Layout: Virginia B. Taylor  (email:  virginia.taylor@fda.hhs.gov )