The use of transgenic mice for short-term, in vivo mutagenicity testing

212 GATA 7(8): 212-218, 1990 SPECIAL SECTION: TRANSGENIC ANIMALS The Use of Transgenic Mice for Short-Term, in vivo Mutagenicity Testing S T E V E N W. K O H L E R , G. SCOTT PROVOST, PATRICIA L. K R E T Z , A N N A B E T H FLECK, J O S E P H A. S O R G E , and JAY M. S H O R T In order to develop a short-term, in vivo assay to study the mutagenic effects o f chemical exposure, transgenic mice were generated using a lambda shuttle vector containing a lacZ target gene. Following exposure to mutagens, this target can be rescued efficiently from genomic D N A prepared from tissues o f the treated mice using restriction minus, in vitro lambda phage packaging extract and restriction minus Escherichia coli plating cultures. Mutations in the target gene appear as colorless plaques on a background o f blue plaques when plated on indicator agar. Spontaneous background levels were ~1 x 10 -5 in each o f three mouse lineages analyzed. Exposure o f lambda transgenic mice to N-ethyl-N-nitrosourea resulted in as much as a 14-fold induction in detected mutations over background levels. The assay is currently being modified to incorporate lacl as the target f o r ease o f mutation detection as well as in vivo excision properties o f the Lambda Z A P vector, facilitating sequence analysis o f mutant plaques. Introduction The development of techniques designed to study the genetic effects of exposure to environmental mutagens began some 50 years ago when the first evidence of the mutagenic potential of a chemical was demonstrated [1]. The need for increasingly accurate and representative assays has become apparent only in recent years, with increasing numbers of synthetic and naturally occurring chemicals being introduced into the marketplace and the environment. Over 200 different assays From Stratagene, La Jolla, California. Address correspondence to: Dr. Jay M. Short, Stratagene, 11099 North Torrey Pines Road, La Jolla, CA 92037. Received June 29, 1990; revised and accepted September 6, 1990. have been developed to address this need [2], and have been used with varying measures of success depending upon both the compound tested and the assay(s) selected. The A m e s ~ S a l m o n e l l a reverse mutation assay has been used extensively with considerable success as a measure of mutagenicity in bacterial systems [3]. Much of this success can be attributed to the addition of liver microsomal fractions, which provide eukaryotic factors to the system, allowing this bacterial assay to activate promutagens metabolically. Eukaryotic-based methods using yeast have been employed with varied success, probably due to the presence of the yeast cell wall, which inhibits the ability of a genotoxic agent from reaching its target [4]. In vitro assays using mammalian cells such as mouse lymphoma or chinese hamster ovary have been developed with the intent of more closely approximating human exposure to mutagens. These short-term in vitro assays have the advantage of being cost effective, but they are often inefficient predictors of the outcome of the long-term bioassay in whole animals, which is the regulatory standard for mutagenicity testing [5]. Strategies to improve the predictive power of the short-term approach by using a battery of short-term assays have not significantly improved the correlation between short-term and long-term testing. The long-term animal bioassay involves the treatment of an animal with a test compound and the subsequent observation of tumor formation to assess long-term effects of exposure to chemicals. It may be difficult, however, to assess thoroughly the mutagenic effects of a particular compound from tumor formation alone, since carcinogenesis is considered to be a multistep phenomenon [6]. In addition, the long-term bioassay is not practical for preliminary screening of all compounds due to the time required (up to 2 years) and the costs, which range from 1 to 2 million dollars per compound [4]. The challenge for genetic toxicologists, given the wide variety of tests available, is to design a program of various assays that will expediently and accurately assess the mutagenicity of a test chemical. This program may vary with each test chemical, depending on the suspected mode of action, the potential utility for society, and how broad the projected exposure. Several general schemes have been proposed and used [1, 4]. One method of testing is the battery approach [4], which involves simultaneously conducting a © 1990 Elsevier Science Publishing Co., Inc., 655 Avenue of the Americas, New York, NY 10010 1050-3862/90/$03.50 213 GATA 7(8): 212-218, 1990 Transgenic Mice in Mutagenicity Testing variety of inexpensive, short-term assays on the test chemical to generate a relatively large data base with which to compare mutagenicity results, and to determine whether the more expensive animal bioassay is necessary. A tiered approach [7] has been proposed using several assays in a sequence with decision points at each step in the scheme. Only if the test chemical scores positive in the first tier (usually an inexpensive, short-term test) would testing proceed to a more definitive, but also more costly, in vivo assay. Currently, however, there is no widely accepted short-term in vivo assay. Compounds passing the second tier testing would proceed to long-term testing. The advantage of the tiered approach is a savings of time and money, but decisions must be made from the examination of a smaller data base. The problems associated with mutagenicity testing would be greatly simplified if a single assay could combine the time savings and ease of short-term in vitro assays with the predictive power of longterm whole animal assays. An area of considerable concern is that of germ cell mutagenicity, which is not directly measured by the long-term rodent bioassay. Mutations may accumulate in germ cells without tumor induction in the long-term bioassay, leading to an improper classification of the mutagenic potential of a chemical. Currently, it is estimated that each phenotypically normal individual carries 5-8 deleterious alleles, which are heritable [8], and that - 6 % of neonates have some sort of genetic anomaly [9]. In terms of risk to the gene pool, mutations in the germ line are much more serious than those in somatic cells. Germ cell mutations can be transmitted to the descendants of those exposed, whereas somatic cell mutations have an endpoint in the affected individual. Thus, germ cell mutations have the potential of remaining in Figure 1. The lambda shuttle vector used in the generation of ~sgenic mice is - 4 6 . 5 kb. Cesium-banded lambda C2B DNA was cos ligated, dialyzed, and diluted to 5 ixg/ml, and - 3 0 0 copies of the vector were injected into the male pronucleus of fertilized B6CBA/J F1 embryos. Lamtxla C2B I I~, I art I II ¢1857 II the gene pool indefinitely, thereby increasing the total genetic load of the species. It is for these reasons that the EPA Guidelines for Mutagenicity Risk Assessment place greater emphasis on those mutagenicity assays that are performed in germ cells rather than somatic cells, on tests performed in vivo rather than in vitro, and on mammalian species rather than submammalian species. In the case of male germ cells, spermatogonial stem cells are of greater relevance, since their life span constitutes the majority of the total reproductive cycle [10]. The stage of germ cell development is also important, since it has been shown that primordial germ cells at various stages react differently to chemical exposure [11]. For most compounds, mammalian germ line mutation data do not exist. There have been attempts to utilize non-germ-line data to estimate the likelihood of inducement of heritable genetic damage [10]. These types of extrapolations would be substantiated by a system capable of assessing mutagenicity in a variety of somatic tissues as well as at various stages of germ cell development. The transgenic mouse/lambda shuttle vector system described here has been designed to address these needs. The basis of this short-term, in vivo mutagenesis assay is the use of a bacteriophage lambda shuttle vector (Figure 1) that has been integrated into the mouse genome via microinjection. The lambda vector has a l a c Z target gene to score for mutations using a color assay. To determine the mutation frequency of a suspected mutagen, mice containing the transgene are exposed to the test chemical (Figure 2). Genomic DNA is then prepared from a variety of tissues. The lambda vector can then easily be recovered from the treated mouse by mixing the DNA with in vitro lambda phage packaging extract, which recognizes the cos ends of the bacteriophage within the genomic DNA and specifically packages the bacteriophage DNA into an infectious phage particle. Each packaged phage, which represents a single rescued target gene, can SL A I J I I ; ,;Lae5. I II L © 1990 Elsevier Science Publishing Co., Inc., 655 Avenue of the Americas, New York, NY 10010 214 GATA 7(8): 212-218, 1990 S. W. Kohler et al. ANY TISSUf ( , I N ( ) M I C DNA Heart . RESCUEI) IARG[: [ GENE PI A(.)LJES NOII-IllLIIdI/[I)laqtl(,s • Mutant Plaques 0 Spleen ~ Kidney grain DAY ONE then infect host Escherichia coli cells to form plaques on the bacterial lawn. When plated on indicator agar plates containing X-gal and IPTG, mutations in the lacZ target gene render the gene inactive and thus cause formation of colorless plaques, whereas intact lacZ targets result in blue plaques. The ratio of colorless plaques to background blue plaques is, therefore, a measure of the mutation frequency of the test chemical. The nature of the mutations can then be analyzed by sequencing through the target gene region. Materials and Methods Exposure of Mice to the Alkylating Agent, N-Ethyl-N-Nitrosourea (EtNU) Compounds can be administered to the animal based on the anticipated level and route of human exposure [4]. In previous studies [12], 6- to 12week-old transgenic B6CBA/F1 J hybrid mice were treated by intraperitoneal injection with two concentrations of EtNU (125 mg/kg body weight and 250 mg/kg body weight). Control animals were similarly injected with 10 ml phosphate buffer per kilogram body weight. Separate doses were administered on days 1 and 4 and animals were sacrificed 2 h after the final dose. Tissues were collected, flash frozen in liquid nitrogen and stored until use at -80°C. I Two I DA T,, FE I Figure 2. Flow chart of the temporal sequence of the assay procea-fi-re. After dosing, the assay can be completed in 3 - 4 days. Isolation of Genomic DNA Frozen tissue (100-500 rag) was transferred to a 7-ml Wheaton tissue grinder dounce containing 3 ml douncing buffer (6 mM NazHPO4 • 7 H20, 130 mM NaC1, 13 mM KCI, 1.5 mM KHzPO 4, and 10 mM Na 2 • EDTA, pH 8.0). The tissue was disaggregated using a Wheaton pestle B (0.0004-in. clearance) and transferred to a 50-ml conical tube. Three milliliters of Proteinase K Solution (800 txg/ml Proteinase K [Stratagene Cloning S y s t e m s ( S C S ) ] , 2% S D S , and I00 mM Na2 • EDTA) was quickly added and mixed by inverting the tube several times. This mixture was then incubated at 50°C for 4 h. An equal volume of phenol/chloroform (Mallinkrodt) saturated with TE (10 mM Tris, pH 8.0, 1 mM EDTA) was added and the mixture was inverted until an emulsion formed. The emulsions were then centrifuged at 2000 g for 15 minutes at 4°C and the aqueous phase was transferred to a new tube with a largebore transfer pipet. This phenol/chloroform extraction was repeated, followed by a chloroform extraction. Two volumes of 100% ethanol were added to the final aqueous phase and mixed by inversion until a visible DNA precipitate formed. The DNA precipitate was then transferred to a © 1990 Elsevier Science Publishing Co., Inc., 655 Avenue of the Americas, New York, NY 10010 215 GATA 7(8): 212-218, 1990 Transgenic Mice in Mutagenicity Testing fresh 15-ml tube, washed with 80% ethanol, and resuspended in 0.5-1.0 ml TE (pH 8.0) over a period of 24 h at 4°C. Rescue of Transgene from Genomic DNA The lambda shuttle vector was recovered from the genomic DNA using Transpack in vitro lambda packaging extract (SCS). Typically, 20 txl of genomic DNA were incubated with packaging extract. As a control, 500 ng of lambda ci857 Sam7 DNA were packaged to determine the efficiency of the packaging extract. These reactions were then incubated for 2 hours at room temperature and subsequently terminated by addition of 2.5 ml of SM Buffer [I0 mM NaCI, 8 mM MgSO4, 50 mM Tris-HC1 (pH 7.5), 0.01% gelatin]. The terminated reactions were kept on ice. Preparation of E. coli Plating Cultures Twenty milliliters of LB medium supplemented with 12.5 mM MgSO4 and 2.5% maltose were innoculated with a single colony of E. coli containing the delta minute 99 deletion [13] [(SCS-7: mcrA-, A(mrr-hsd-mcrB)] and grown at 37°C with shaking at 250 rpm until OD6oo -~ 1.5. E. coli DP50 (hsd-) was grown as indicated above for use with the lambda ci857 Sam7 control. The cultures were centrifuged at 2000 g for 10 min, the supernatant discarded, and the cell pellets resuspended in 10 mM MgSO 4 at OD600 = 0.5. Cultures were stored at 4°C. Plating Rescued Phage Two milliliters of OD60o = 0.5 SCS-7 cells were aliquoted into a 50-ml conical tube for each packaging reaction to be plated. Five hundred microliters of packaging reaction containing rescued phage from mouse genomic DNA were added to each tube of host cells and incubated at 37°C for 15 minutes. Twenty-five milliliters of 50°C molten top agarose [0.35% agarose containing 3 mM IPTG (SCS) and 3 mM Xgal (SCS)] was added to each tube, mixed by swirling, and immediately plated onto 25-cm x 25-cm assay dishes containing NZY agar. These dishes were then incubated overnight at 37°C. Similarly, 500 ~1 of OD6oo = 0.5 DP50 cells were aliquoted into a 15-ml tube, infected with 1 ~1 of a 1 × 10 -4 dilution of the lambda CI857 Sam7 packaging reaction, incubated at 37°C for 15 min, plated with 8 ml of molten top agarose onto a 150-mM NZY agar plate, and incubated at 37°C overnight. Results Effects of Methylation on Lambda Rescue Initial attempts at rescue of the lambda phage from the genome of transgenic mice resulted in efficiencies too low for mutation analysis. The effects of eukaryotic methylation were investigated with the cytosine analog, 5-azacytidine, through treatment of fibroblast cultures derived from fetuses of the transgenic mice [12]. This chemical treatment reduces the level of methylation within the DNA. Rescue efficiencies of demethylated cells demonstrated a 50-fold increase in rescue efficiency over nontreated controls [12, 14]. The inhibitory effects of eukaryotic methylation supported by this set of experiments, combined with the identification of several prokaryotic restriction systems as being inhibitory to methylated eukaryotic DNA [15], supported the decision to remove as many sources of restriction activity as possible. Mutation of the mcrA locus and deletion of the minute 99 region containing the mrr, hsd, and mcrB,C loci in E. coli K-12 packaging extract strains and plating cultures [13, 16] resulted in rescue efficiencies >30% of theoretical levels without 5-azacytidine (Figure 3) [12, 14]. Mutagenesis Testing The high rescue efficiency obtained with restriction free extracts and strains allowed determination of spontaneous background rates. These rates were compared in three lineages: AL, BA, and LU [12] (Table I). Spontaneous mutation rates ranged from 6.8 x 10 -6 to 6.3 × 10 -5, depending upon the tissue analyzed. These levels were sufficiently low to allow mutagenicity testing. Initial testing was performed with EtNU at 125 mg/kg in four mice and 250 mg/kg in five mice. A group of five control mice were treated with phosphate buffer in the same manner as the test animals. Animals were sacrificed and genomic DNA was prepared from the spleens of the animals. Spontaneous background mutation rates derived from the phosphate buffer-treated control mice were - I x 10 -5 (Figure 4). Treatment of the mice with EtNU resulted in as much as a 14fold increase in the number of mutant plaques obtained over background levels [12] (Figure 4). © 1990 Elsevier Science Publishing Co., Inc., 655 Avenue of the Americas, New York, NY 10010 216 G A T A 7(8): 2 1 2 - 2 1 8 , 1990 S. W. Kohler et al. 40% -- (-) nl Z iii 31.9% ¢3 (D Z 30% -- < n-" ii © >r'r iii > 0 0 W Figure 3. B a r graphs showing effect of p r o k a r y o t i c ~ o n systems on lambda phage recovery from genome of transgenic mice. Theoretical rec o v e r y is defined as the maximum recovery/copy/ cell of the shuttle vector, within the efficiency of lambda phage packaging. The ( + ) symbol denotes K - 1 2 (mcrA + ,rncrB- ) packaging extracts and the (-) denotes K-12 p a c k a g i n g extracts. 20% -- rr" A(mrr-hsdRMS-mcrB),mcrA- (+) 10%-- W n" o UJ 7- (+) (4 0.0001% 0.002% ( mcr 2.3% ( mcr A-,B+) Discussion The ability to recover lambda shuttle vectors efficiently from the genomic DNA of transgenic mice using restriction minus E. coli host strains and packaging extracts has enabled the system to be used successfully as a short-term, in vivo mutagenesis assay. A dose-dependent induction of detected lacZ mutations over background was seen after EtNU treatment. Current studies are underway to determine the sensitivity of this assay for detection of weaker mutagens in somatic cells. The spontaneous background level of 1 x 10 -5 makes this assay more sensitive than existing germ cell tests, where the sensitivity is limited by the number of animals used. One such assay, the specific locus test [17], utilizes 12,000-14,000 offTable 1. Spontaneous mutation levels Line LU Figure 4. G r a p h showing increase in number of mutant ~etected upon treatment with E t N U . A n average of 1 x 105 p l a q u e s w e r e recovered from each animal. 25-- DOSE RESPONSEIN SPLEEN 20- .~ Tissue Brain Kidney Spleen Testis AL spring of treated animals in a typical assay [4]. In contrast, over 1 million target genes can be analyzed from a single mouse using the transgenic mouse lacZ system. The availability of a short-term, in vivo assay capable of providing metabolizing activities may help to resolve discrepancies between short-term assays and long-term carcinogenicity tests. For example, carcinogenicity studies using the compounds furan and furfural suggest an increase in liver tumor formation even though neither com- ~= Mouse A-,B-) E.coli E.coli Kidney Spleen Frequency <3.5 <2.8 1.3 8.1 × x × x 10 -5 10 -5 10 -5 10 - 6 ~ o 15- 10 4.8 x 10 -5 < 1 . 4 x 10 -5 m BA Brain Kidney Liver Spleen Testis <3.4 1.1 6.3 6.8 <1.0 x x x x x 10 -5 10 -5 10 -5 10 - 6 10 -5 Spontaneous mutation rates were determined for three lineages: LU, AL, and BA. For tissues where no mutant plaques were recovered, data points are listed with a (<) symbol and the number of plaques rescued. "5 E 5- I ] PHOSPHATE BUFFER contro( © 1990 Elsevier Science Publishing Co., Inc., 655 Avenue of the Americas, New York, NY 10010 EtNU 125mg/kg body w t EtNU 250mg/kg body w t 217 GATA 7(8): 212-218, 1990 Transgenic Mice in Mutagenicity Testing pound induced mutations in the Ames~Salmonella assay [18]. This is despite the fact that liver tumors induced by these compounds were found to contain point mutations within the ras oncogene. In vivo testing of the mutagenicity of furanlike compounds in the transgenic system will address whether these compounds may act as tumor promoters by activating previously initiated cells. Alternatively, studies may show that such nongenotoxic carcinogens are actually mutagenic, and that in vitro assays are inappropriate for detection of their mutagenicity. Although spontaneous background rates for the lineages described in Table 1 do not vary significantly, it is expected that for some lineages generated, spontaneous background rates may vary with several parameters. These include integration site of the transgene, the transcriptional activity of the transgene, copy number of the lambda transgene, the strain of mouse used to generate the transgenics, and whether the mouse is homozygous for the transgene locus. The actual effect of each of these parameters has yet to be evaluated fully. In addition, spontaneous background rates may vary within a single transgenic lineage. For instance, if a mouse undergoes a spontaneous mutation in the lacZ locus early in development, then by clonal expansion that cell will confer its mutant phenotype to progeny cells, giving rise to an artificially high spontaneous background rate. It will be important, therefore, to rely on data from not one, but several animals in each test group when evaluating mutagenesis, and to use statistical methods to eliminate aberrant data points, should this situation occur. In an effort to derive data more quickly and easily from this assay, testing has begun on a new system, which has several modifications to the original assay (S.W.K., G.S.P., A . E , P.L.K., J.M.S., manuscript in preparation). First, the new lambda shuttle vector contains the E. coli lacI gene as the mutagenic target controlling lacZ. This serves two functions: (a) Screening for mutations in lacI allows mutated plaques to appear as an easily detectable blue on a background of colorless plaques; and (b) lacI is -1290 bp in length, whereas lacZ is -3200 bp, enabling easier characterization of mutations in lacI by sequence analysis. The other important modification to the shuttle vector is the addition of the Lambda ZAP vector functions, which permit in vivo excision of the target gene into a plasmid. This facilitates rapid sequence analysis of the target gene without the need for subcloning [19]. The existence of the lacI and lacZ target gene systems will now allow the comparison of mutation spectra in each of two different genes. In addition, several transgenic lineages exist with the target gene located at different chromosomal sites of integration, allowing comparison of mutation spectra as a function of chromosomal position. Data generated from these assays may also be correlated directly with long-term carcinogenesis studies. To allow these systems to detect mutations specifically in the germ line of the transgenic mouse, the DNA isolation protocol described here may be modified to permit isolation at various stages of development of male germ cells that have been purified away from the surrounding somatic tissue (testis capsule, connective tissue). Standard techniques of density gradient centrifugation can be applied to isolate specific precursor lineages for analysis of germ cells that may exhibit increased mutagen sensitivity. Genomic DNA can be isolated using the previously described method, and mutations in specific male germ cell lineages can then be monitored without somatic cell contamination. This procedure should greatly simplify determination of germ cell mutagenesis rates, making it practical to assess the genetic risk of most suspected compounds. The ability to use this assay to detect and characterize lesions in a chromosomal target while analyzing a variety of tissues should enable this system to provide information that cannot currently be obtained by any single assay system. In addition, the short-term nature of the assay should enable more cost-effective screening of compounds and perhaps reduce the number of assays that are required in either battery or tiered testing strategies. The authors wish to acknowledge Dr. Jane Moores and Heidi Short for critical reading of the manuscript. This work was supported in part by NIEHS SBIR grant 2R44ES04484-02 and in part by NIEHS grant 1R01ES04728-01A1. References 1. Sobels FH: Mutat Res 181:299-310, 1987 2. Nesnow S, Argus M, Bergman H, et al.: Mutat Res 1985:1-195, 1987 © 1990 Elsevier Science PublishingCo., Inc., 655 Avenue of the Americas, New York, NY 10010 218 GATA 7(8): 212-218, 1990 3. Ames BN, Lee FD, Durston WE: Proc Natl Acad Sci USA 70:782-786, 1973 4. Brusick D: In Principles of Genetic Toxicology, 2nd ed. New York, Plenum Press, 1987, p 196 5. IARC Monographs on the Evaluation of Carcinogenicity: An Update of IARC Monographs 1 to 42 Suppl. 7. Lyon, International Agency for Research on Cancer, 1987 6. Fearon ER, Vogelstein B: Cell 61:759-767, 1990 7. Flamm WG: Mutat Res 26:329, 1974 8. Fabricant JD, Legator MS: Environ Health Perspect 41:189-193, 1981 9. Venitt S, Parry JM: In Mutagenicity Testing, A Practical Approach. Oxford, IRL Press, 1984, pp 14-15 10. Russell LB, Aaron CS, de Serres F, et al.: Mutat Res 134:143-157, 1984 11. Sega GA, Owens JG: Environ Mol Mutagen 10:119-127, 1987 S. W. Kohler et al. 12. Kohler SW, Provost GS, Kretz PL, Dycaico MJ, Sorge JA, Short JM: Nucleic Acids Res 18(10):3007-3013, 1990 13. Kretz PL, Kohler SW, Short JM: Submitted, 1990 14. Short JM, Kohler SW, Provost GS, Fieck A, Kretz PL: In Mutation and the Environment, part A. New York, WileyLiss, 1990, pp 355-367 15. Raleigh EA, Wilson G: Proc Natl Acad Sci USA 83:90709074, 1986 16. Kretz PL, Reid CH, Greener A, Short JM: Nucleic Acids Res 17:5409, 1989 17. Russell WL: Cold Spring Harbor Symp Quant Biol 16:327-335, 1951 18. Reynolds SH, Stowers S J, Patterson RM, Maronpot RR, Aaronson SA, Anderson MW: Science 237:1309-1316, 1987 19. Short JM, Fernandez JM, Sorge JA, Huse WD: Nucleic Acids Res 16:7583-7600, 1988 © 1990 Elsevier Science Publishing Co., Inc., 655 Avenue of the Americas, New York, NY 10010