Page B. McKinzie

Prospects for applying genotypic selection of somatic oncomutation to chemical risk assessment.

Genotypic selection methods detect rare sequence changes in populations of DNA molecules. These methods have been used to investigate the chemical induction of mutation and for the detection and diagnosis of cancer. The possible use of genotypic selection for improving current risk assessment practices is based on the premise that the frequency of somatic mutation is of critical importance in understanding and modeling carcinogenesis. If genotypic selection can measure the induction of specific mutations that disrupt normal cell/tissue homeostasis, then it could provide key mechanistic information for cancer risk assessment. For example, genotypic selection data might support a particular low-dose extrapolation method or characterize the relationship between rodent and human cancer risk. Strategies for evaluating the use of genotypic selection in cancer risk assessment include the concept of developing a battery of targets that detect a range of agent-specific effects. Ideal targets to examine by genotypic selection are the oncogene and tumor suppressor gene mutations frequently detected in human tumors because these are thought to represent tumor-initiating events. The most commonly occurring basepair (bp) substitutions within the ras and p53 genes are identified. Also, the battery of genotypic selection methods is defined in terms of the most important mutational specificities to include. In theory, the major basepair substitution mutations induced by 29 of 31 chemical carcinogens could be detected by analyzing three different mutations: G:C-->T:A, G:C-->A:T, and A:T-->T:A. Genotypic selection will have the greatest impact on risk assessment if measurement of spontaneous mutation is possible. Data from phenotypic selection assays suggest this corresponds to detection of mutant fractions of approximately 10(-7), and this would necessitate examining DNA samples containing >10(7) target molecules. Despite its apparent potential, considerable development and validation is needed before genotypic selection data can be applied to cancer risk assessment.

ACB-PCR Quantification of K-RAS Codon 12 GAT and GTT Mutant Fraction in Colon Tumor and Non-Tumor Tissue

K-RAS mutation is being developed as a cancer biomarker and tumor K-RAS is being used to predict therapeutic response. Yet, levels of K-RAS mutation in normal and pathological tissue samples have not been determined rigorously, nor inter-individual variation in these levels characterized. Therefore, K-RAS codon 12 GAT and GTT mutant fractions were measured in colonic mucosa of individuals without colon cancer, tumor-distal mucosa, tumor-proximal mucosa, normal tumor-adjacent tissues, colonic adenomas, and carcinomas. The results indicate K-RAS codon 12 GAT mutation is present at measurable levels in normal appearing mucosa. All tumors carried K-RAS mutation, in most cases as a mutant subpopulation.

Measurement of tumor-associated mutations in the nasal mucosa of rats exposed to varying doses of formaldehyde.

This study examined the potential induction of tumor-associated mutations in formaldehyde-exposed rat nasal mucosa using a sensitive method, allele-specific competitive blocker-PCR (ACB-PCR). Levels of p53 codon 271 CGT to CAT and K-Ras codon 12 GGT to GAT mutations were quantified in nasal mucosa of rats exposed to formaldehyde. In addition, nasal mucosa cell proliferation was monitored because regenerative cell proliferation is considered a key event in formaldehyde-induced carcinogenesis. Male F344 rats (6-7 weeks old, 5 rats/group) were exposed to 0, 0.7, 2, 6, 10, and 15 ppm formaldehyde for 13 weeks (6 h/day, 5 days/week). ACB-PCR was used to determine levels of p53 and K-Ras mutations. Although two of five untreated rats had measureable spontaneous p53 mutant fractions (MFs), most nasal mucosa samples had p53 MFs below 10(-5). All K-Ras MF measurements were below 10(-5). No dose-related increases in p53 or K-Ras MF were observed, even though significant increases in bromodeoxyuridine incorporation demonstrated induced cell proliferation in the 10 and 15 ppm formaldehyde-treatment groups. Therefore, induction of tumor-associated p53 mutation likely occurs after several other key events in formaldehyde-induced carcinogenesis.

Oncomutations as biomarkers of cancer risk

Cancer risk assessment impacts a range of societal needs, from the regulation of chemicals to achieving the best possible human health outcomes. Because oncogene and tumor suppressor gene mutations are necessary for the development of cancer, such mutations are ideal biomarkers to use in cancer risk assessment. Consequently, DNA‐based methods to quantify particular tumor‐associated hotspot point mutations (i.e., oncomutations) have been developed, including allele‐specific competitive blocker‐PCR (ACB‐PCR). Several studies using ACB‐PCR and model mutagens have demonstrated that significant induction of tumor‐associated oncomutations are measureable at earlier time points than are used to score tumors in a bioassay. In the particular case of benzo[a]pyrene induction of K‐Ras codon 12 TGT mutation in the A/J mouse lung, measurement of tumor‐associated oncomutation was shown to be an earlier and more sensitive endpoint than tumor response. The measurement of oncomutation by ACB‐PCR led to two unexpected findings. First, oncomutations are present in various tissues of control rodents and “normal” human colonic mucosa samples at relatively high frequencies. Approximately 60% of such samples (88/146) have mutant fractions (MFs) >10−5, and some have MFs as high as 10−3 or 10−4. Second, preliminary data indicate that oncomutations are present frequently as subpopulations in tumors. These findings are integrated into a hypothesis that the predominant preexisting mutations in particular tissues may be useful as generic reporters of carcinogenesis. Future research opportunities using oncomutation as an endpoint are described, including rodent to human extrapolation, dose‐response assessment, and personalized medicine. Environ. Mol. Mutagen., 2010. Published 2010 Wiley‐Liss, Inc.

Accumulation of K-Ras Codon 12 Mutations in the F344 Rat Distal Colon Following Azoxymethane Exposure

Azoxymethane (AOM) administration to F344 male rats is a widely used model of human colon carcinogenesis. The current study investigates quantitatively the accumulation of K‐Ras codon 12 mutations following AOM exposure. Male, 6‐week‐old F344 rats were treated subcutaneously with 30 mg/kg body weight of AOM, and colon tissue was collected at 1, 8, 24, and 32 weeks after treatment. The K‐Ras codon 12 GGT to GAT and GGT to GTT mutant fractions (MFs) were measured using allele‐specific competitive blocker polymerase chain reaction (ACB‐PCR). Between 1 and 32 weeks after AOM treatment, the K‐Ras codon 12 GGT to GAT geometric mean MF in the rat colon increased significantly from 12.9 × 10−5 to 145 × 10−5, and the GGT to GTT geometric mean MF increased significantly from 5.26 × 10−5 to 180 × 10−5. K‐Ras codon 12 GGT to GAT MF also increased significantly in AOM‐treated rat colon tissue at 1 week compared to controls (4.44 × 10−5). The accumulation of the GGT to GAT MF long after the DNA adduct repair phase suggests that a portion of cells containing this mutation have a proliferative advantage, allowing them to accumulate as nascent tumors progress. Also, the GGT to GAT background MF increased in untreated rats, indicating that there is accumulation with age. The ACB–PCR assay generates quantitative data of cancer‐related mutations and thus provides insight into pathological processes following carcinogen exposure. Environ. Mol. Mutagen., 2011.

Spectrum of benzo[a]pyrene-induced mutations in the Pig-a gene of L5178YTk+/− cells identified with next generation sequencing

We used Sanger sequencing and next generation sequencing (NGS) for analysis of mutations in the endogenous X-linked Pig-a gene of clonally expanded L5178YTk+/- cells. The clones developed from single cells that were sorted on a flow cytometer based upon the expression pattern of the GPI-anchored marker, CD90, on their surface. CD90-deficient and CD90-proficient cells were sorted from untreated cultures and CD90-deficient cells were sorted from cultures treated with benzo[a]pyrene (B[a]P). Pig-a mutations were identified in all clones developed from CD90-deficient cells; no Pig-a mutations were found in clones of CD90-proficient cells. The spectrum of B[a]P-induced Pig-a mutations was dominated by basepair substitutions, small insertions and deletions at G:C, or at sequences rich in G:C content. We observed high concordance between Pig-a mutations determined by Sanger sequencing and by NGS, but NGS was able to identify mutations in samples that were difficult to analyze by Sanger sequencing (e.g., mixtures of two mutant clones). Overall, the NGS method is a cost and labor efficient high throughput approach for analysis of a large number of mutant clones.

Whole genome sequencing of mouse lymphoma L5178Y-3.7.2C (TK+/−) reveals millions of mutations and genetic markers

The mouse lymphoma L5178Y-3.7.2C (TK+/-) cell line is extensively used in genetic toxicology to conduct the mouse lymphoma assay (MLA). The MLA is used to establish the mutagenic and clastogenic effects of chemicals and pharmaceuticals, and is one of the few genetic tests widely accepted by regulatory agencies throughout the world. Despite the extensive use and regulatory impact of L5178Y-3.7.2C (TK+/-) cells, little is known about their genetic composition or how it affects the outcome of the MLA. To determine the genetic background of this cell line, we sequenced and analyzed its entire genome. Our results confirm the existence of previously described mutations in the Tk1 and Trp53 genes and catalog millions of other mutations, many of which impair the function of genes with key roles in cell physiology and genetic toxicology.

Whole genome and normalized mRNA sequencing reveal genetic status of TK6, WTK1, and NH32 human B-lymphoblastoid cell lines.

Closely related TK6, WTK1, and NH32 human B-lymphoblastoid cell lines differ in their p53 functional status. These lines are used frequently in genotoxicity studies and in studies aimed at understanding the role of p53 in DNA repair. Despite their routine use, little is known about the genetic status of these cells. To provide insight into their genetic composition, we sequenced and analyzed the entire genome of TK6 cells, as well as the normalized transcriptomes of TK6, WTK1, and NH32 cells. Whole genome sequencing (WGS) identified 21,561 genes and 5.17×10(6) small variants. Within the small variants, 50.54% were naturally occurring single nucleotide polymorphisms (SNPs) and 49.46% were mutations. The mutations were comprised of 92.97% single base-pair substitutions and 7.03% insertions or deletions (indels). The number of predicted genes, SNPs, and small mutations are similar to frequencies observed in the human population in general. Normalized mRNA-seq analysis identified the expression of transcripts bearing SNPs or mutations for TK6, WTK1, and NH32 as 2.88%, 2.04%, and 1.71%, respectively, and several of the variant transcripts identified appear to have important implications in genetic toxicology. These include a single base deletion mutation in the ferritin heavy chain gene (FTH1) resulting in a frame shift and protein truncation in TK6 that impairs iron metabolism. SNPs in the thiopurine S-methyltransferase (TPMT) gene (TPMT*3A SNP), and in the xenobiotic metabolizing enzyme, NADPH quinine oxidoreductase 1 (NQO1) gene (NQO1*2 SNP), are both associated with decreased enzyme activity. The clinically relevant TPMT*3A and NQO1*2 SNPs can make these cell lines useful in pharmacogenetic studies aimed at improving or tailoring drug treatment regimens that minimize toxicity and enhance efficacy.

Glycosylphosphatidylinositol (GPI) anchored protein deficiency serves as a reliable reporter of Pig‐a gene Mutation: Support from an in vitro assay based on L5178Y/Tk+/− cells and the CD90.2 antigen

Lack of cell surface glycosylphosphatidylinositol (GPI)‐anchored protein(s) has been used as a reporter of Pig‐a gene mutation in several model systems. As an extension of this work, our laboratory initiated development of an in vitro mutation assay based on the flow cytometric assessment of CD90.2 expression on the cell surface of the mouse lymphoma cell line L5178Y/Tk+/−. Cells were exposed to mutagenic and nonmutagenic compounds for 24 hr followed by washout and incubation for an additional 7 days. Following this mutant manifestation time, cells were labeled with fluorescent antibodies against CD90.2 and CD45 antigens. These reagents indicated the presence of GPI‐anchored proteins and general cell surface membrane receptor integrity, respectively. Instrument set‐up was aided by parallel processing of a GPI anchor‐deficient subclone. Results show that the mutagens reproducibly caused increased frequencies of mutant phenotype cells, while the nonmutagens did not. Further modifications to the method, including application of a viability dye and an isotype control for instrument set‐up, were investigated. As a means to verify that the GPI‐anchored protein‐negative phenotype reflects bona fide Pig‐a gene mutation, sequencing was performed on 38 CD90.2‐negative L5178Y/Tk+/− clones derived from cultures treated with ethyl methanesulfonate. All clones were found to have mutation(s) within the Pig‐a gene. The continued investigation of L5178Y/Tk+/− cells, CD90.2 labeling, and flow cytometric analysis as the basis of an in vitro mutation assay is clearly supported by this work. These data also provide evidence of the reliability of using GPI anchor‐deficiency as a valid reporter of Pig‐a gene mutation. Environ. Mol. Mutagen. 59:18–29, 2018.

Abstract 1739: The prevalence of KRAS, PIK3CA, and BRAF mutant subpopulations in tumors may be impacting the success of personalized cancer treatment

Proceedings: AACR 103rd Annual Meeting 2012‐‐ Mar 31‐Apr 4, 2012; Chicago, IL Tumor mutations are being used as predictive biomarkers of response, in order to select the most effective treatment for individual cancer patients. Currently, this is being done without sufficient characterization of relevant oncogene mutations as quantitative biomarkers. The goal of the current study was to define normal and pathological levels of the most prevalent hotspot mutations in the KRAS, PIK3CA, and BRAF genes, to establish the frequency with which the mutations occur as subpopulations, and what diagnostic sensitivity is needed to detect defined percentages of tumors carrying mutant subpopulations. Therefore, the sensitive Allele-specific Competitive Blocker-PCR (ACB-PCR) was used to quantify the levels of specific hotspot point mutations in a panel of normal human tissues and tumors. The mutations examined have established significance in terms of personalized cancer treatment, specifically KRAS G12D, KRAS G12V, BRAF V600E, and PIK3CA H1047R. The tissues examined included lung, colon, pancreas, and thyroid. In colon tumors, the 5th, 25th, 50th, 75th, and 95th percentiles of KRAS G12D mutant fraction (MF) are 1.7 x 10−5, 7.4 x 10−5, 3.0 x 10−4, 2.8 x 10−2, and 8.4 x 10−1, respectively. In lung tumors, the 5th, 25th, 50th, 75th, and 95th percentiles of KRAS G12V MF are 7.0 x 10−6, 1.1 x 10−5, 3.3 x 10−5, 2.3 x 10−2, and 1.2 x 10−1, respectively. Based on the data across these tissue types, 67.5% of tumors carry KRAS G12D or G12V mutation at a subpopulation frequency higher than that observed in normal tissue. Only 18.1% of tumors had a KRAS MF α10−1 (i.e., that detectable by DNA sequencing). From these data it was determined a diagnostic with a sensitivity of 10−2 or 10−3 would detect 27.7% or 43.4% of these tumors, respectively. Surprisingly, analysis of KRAS mutation in papillary thyroid tumors showed KRAS G12V mutations were present above normal thyroid levels, but as subpopulations in 42.1% of papillary thyroid tumors, even though the COSMIC database indicates this mutation occurs in only 0.15% of papillary thyroid tumors. The occurrence of these KRAS G12V mutations was positively correlated with percent tumor necrosis. For PIK3CA H1047R mutation in colon tumors, the 5th, 25th, 50th, 75th, and 95th percentiles are 1.2 x 10−6, 5.3 x 10−4, 7.6 x 10−4, 1.1 x 10−3, and 4.2 x 10−2, respectively. Data on BRAF V600E shows it occurs primarily as large subpopulations in papillary thyroid tumors. For effective development of personalized cancer treatment, quantitative and sensitive analyses of tumor mutations are needed to establish the effect of mutant subpopulations on patient response and/or relapse. Because so many tumors carry KRAS mutation, therapies targeting KRAS mutant cells are needed for use in conjunction with therapies directed against other targets. The views presented do not necessarily reflect those of the US FDA. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 1739. doi:1538-7445.AM2012-1739