Recent developments with the L5178Y TK heterozygote mutagen assay system.

The presently available mutagen test systems are no longer receiving critical acclaim by sole virtue of their positive response to ethyl methanesulfonate (EMS). Rather, they are being asked to yield answers to such questions as: how mutagenic is EMS in this system and, given this number, what does it imply relative to the genetic risk of EMS to man? Both of these questions address themselves to the relevance of the test results, the first in a quantitative manner, the second qualitatively, in that it involves evaluating species' and systems' differences. Because this extrapolation of data must be to a rather complex organism-man-we are immediately'"face&'withagenetic quandary. Or'ganisms -which display thei-r'genetic repertoires in the simplest, most rapid manner-and which, therefore, are easiest to use and most definitive in the answers they yield-are phylogenetic foreigners to many. Used alone, one risks banning penicillin on the basis of toxic effects on gram-positive bacteria, or, to cite an example more oriented to mutagenicity testing, one might be tempted to place 5-bromodeoxyuridine (BUdR) on the GRAS list as a result of testing in thymidine kinase-deficient yeast. Yet, if we choose to eliminate most of this


Introduction
The presently available mutagen test systems are no longer receiving critical acclaim by sole virtue of their positive response to ethyl methanesulfonate (EMS). Rather, they are being asked to yield answers to such questions as: how mutagenic is EMS in this system and, given this number, what does it imply relative to the genetic risk of EMS to man? Both of these questions address themselves to the relevance of the test results, the first in a quantitative manner, the second qualitatively, in that it involves evaluating species' and systems' differences. Because this extrapolation of data must be to a rather complex organism-man-we are immediately'"face&'witha-genetic quandary. Or'ganisms -which display thei-r'genetic repertoires in the simplest, most rapid manner-and which, therefore, are easiest to use and most definitive in the answers they yield-are phylogenetic foreigners to many.
Used alone, one risks banning penicillin on the basis of toxic effects on gram-positive bacteria, or, to cite an example more oriented to mutagenicity testing, one might be tempted to place 5-bromodeoxyuridine (BUdR) on the GRAS list as a result of testing in thymidine kinase-deficient yeast. Yet, if we choose to eliminate most of this *Mutagenesis Branch, National Institute of Environmental Health Sciences, National Institutes of Health, Research Triangle Park, North Carolina 27709. xenomorphism we are faced with expensive, time-consuming mega-chimp experiments which relinquish genetic secrets at such a slow rate that extrapolation to man is done in a near vacuum of supporting studies, thus weakening the advantage of evolutionary proximity.
One compromise to these extremes has been to microbialize the mammal. Growing mammalian somatic cells in tissue culture accomplishes four aims: (1) over 108 genetically expressive genomes are compacted into a volume less than that occupied by one average-size whole mammal; (2) the generation time and hence the duration of the experiment-is reduced from months or years to hours; '(3) definitive genetic studies, including mutagenicity studies akin to those heretofore reserved by the microbial geneticist, can now be performed on mammalian cells; and (4) we have preserved most of the structural and functional peculiarities of the mammalian genome.
However, such cells do not undergo meiosis, nor do they posses the differentiated intricacies which confer upon the whole mammal some degree of protection from a wide variety of toxic agents (but which can also activate innocuous compounds into hazardous ones). For these reasons, such in vitro systems do not constitute a panacea for mutagen testing. Like other systems, their limitations must be defined as ardently as their advantages are expoited.
The TK+'-Heterozygote System sa 2). [Hycanthoneis a highly efficacious antischistosomaldrug which *had earlier been e a e of m shown to induce .frameshift mutations in cells grown in tissue culture are admirably mo by L178 mose ympomacel's.'''Salm-onella-(6) -and to intercalate into DNA illustrated by L5178Y mouse lymphoa cells(7>]. This line of transformed cells was frst isolated in the late 1950's by Fischer (1).
Subsequent studies onra 'different hetero-They grow well in suspension culture with zygote (TK+/--3.7.2) with a higher TK+'a generation time of 10-11 hr; they dfone i -TK; spontaneous mutation rate (2 x with 80-100%o efficiency in a soft agar me-+ 10 mutations / locus / generation) show dium (2,3). We have found (3,4) these cells the wide range of induced mutant frequento be diploid at the thymidine kinase (TK) cies to which this system can respond.- Figure  locus and have developed selection 'media 1 illustrates the relative mutagenicities of for isolating TK-competent or TK-deficient the three alkylating agents, EMS, MMS mutants occurring at low frequencies among (methyl methanesulfonate), and EDB (ethycells of the opposite phenotype. Reconstruclene dibromide) as a function of their molar tion experiments show recovery of 80-100%o of plated mutants under the conditions used. Table 1 summarizes the genetic and phenotypic characteristics of the two homozygotes I0 (TK+/+ and TK-/-) and of the heterozygote 2 (TK+/-) clonal lines that were initially in-. vestigated (5). Spontaneous mutation rates, 90 TK activities, and mode of isolation are all l0 &MMS M/ compatible with the assigned genotypes.
80i The TK+/--+TK-'-(our basic mutagen as-8 say system) mutation rate was greatly stim-S 70' ulated by x-irradiation, EMS, and 6 hycanthone-methanesulfonate ( Thymidine kinase activity, pmole thymidine phosphorylated/10' Spontaneous mutation rate, (Induced mutant frequencies are obtained by subtracting the mutant frequency observed in the control culture from that of each of the treated cultures and in all instances is expressed in terms of number of TK-/mutants per surviving cell. The vertical and horizontal scales are exaggerated in the inset to display more prominently the dose response curves of MMIS and EDB. Though much less mutagenic than EMS they still represent a mutagenic equivalent of over 600 R of x-irradiation at the higher concentrations (cf. Table 2.
This system is capable of still better sensitivity, as is shown in Figure 2. Here the mutagenicity of hycanthone as a function of concentration is compared with the alkylating agents on still further exaggerated axes. The highest hycanthone concentration used is slightly more mutagenic than 600 R of x-irradiation.
So far we have accomplished little except demonstrate that the TK+'--+TK-'-system yields quantitative results and that different external concentrations of different mutagens effect different dose: response curves. This latter result can give seemingly contradictory results. Thus, by comparing the curves for hycanthone and EMS (Figs. 2 and 3) one can claim (a) hycanthone is about 40 times as mutagenic as EMS at 104M concentration, and (b) EMS is 25 times as mutagenic as hycanthone at the highest concentrations examined. Which, if either, of these two statements is to be used when extrapolating these data to man's genetic risk, or are they  Figure 1, comparing dose-mutagenic response curve of hycanthone methanesulfonate with those of the alkylating agents. 600 R of x-irradiation induces approximately 3 x 10-4 mutations/locus, less than the two highest responses to hycanthone.
both relevant? What do such comparisons mean with respect to the relative efficiencies with which alkylation or intercalation events in the DNA can induce mutations? Partial relief from such questions is obtained by replacing extracellular molar concentration of the mutagen as the independent variable in these figures by some parameter related to intracellular dose. In this manner the contribution to mutagenic response of   induced by three alkylating agents, expressed as a function of growth inhibition. The curves of Figures 1 and 2 are made to overlap in their abscissal range by relating mutagenic response to a biological (growth inhibition), rather than to a chemical (molarity) measure. differences in membrane permeability to and/ or transport of various mutagens would be minimized. Ideally, this should be number of mutagen molecules physically associated with the TK gene, but except for EMS (see below) such dosimetric relationships have not yet been established. Instead, we have chosen growth inhibition as a more easily assessible index of intracellular-and possible genomal-concentration.
Growth inhibition is measured in two steps. First, the rate of growth in suspension culture is determined for each treated cul-ture for the combined treatment (2 hr) and expression (48-72 hr) times (shamtreated control = 100). Then cloning efficiencies (relative to Coulter counts) are assessed for each culture (control = 1.00). The product of these two figures for each culture represents the overall growth of that culture relative to the control (= 100%o); growth inhibition is obtained by subtracting these relative growth figures from 100. The previous data can now be plotted over a common abscissal range (0-100%o growth inhibition) as in Figures 3 and 4. Note (Fig.  4) that EMS appears to be a much more potent mutagen over the whole range of growth inhibitions covered, while MMS, hycanthone, and possibly even EDB fall fairly closely together.
These data are now in a format which is both quantitative and easily compared with  Environmental Health Perspectives other systems. If such comparisons show that the mutagenicity-toxicity relationships for a given compound are not too highly speciesdependent-or, if they are, and if wellestablished biochemical or physical difference in the more xenomorphic systems can adequately explain such differences-then the data can be directly extrapolated to man and an estimate of the genetic risk can be hazarded. The emphasis is, of course, on high correlations among various test systems after allowing for their respective pitfalls. As a beginning in putting our mutagenicity test systems to positive social advantage we have been studying the mutagenicity of a number of hycanthone analogs possessing varying degrees of antischistosomal activity. The mutagenicity-or lack thereof-of one of these, IA-3 (the 6-chloroindazole derivative of lucanthone), is compared with that of hycanthone in Figure 5 as a function of both molar concentration and growth inhibition. At up to 80%o inhibition of growth this compound is no more than 10%oas mutagenic as hycanthone. Of course, further testing in other systems is essential before draw- ing firm conclusions relative to clinical potential. Especially, the question of whether hydroxylation of the 4-methyl group occurs in whole mammals (as it does for lucanthone) must be considered, since the hydroxymethyl derivative of IA-3 is 3 times as mutagenic as IA-3 itself (i.e., it is 30%o as mutagenic as hycanthone). As alluded to above, these cells readily lend themselves to dosimetry studies at the genomal level. Specificially, Patterson (8) has determined the extent of ethylation of the DNA over a tenfold range in EMS concentration; by comparing these data with the EMS mutagenicity data an estimated 8 x 10-5 mutations per alkylation occur which is constant over a 5-fold concentration range (see Fig. 6). Allowance must still be made for the unknown fraction of all possible mutations at the TK locus which are not amorphic and hence not detectable in this system. Even assuming as little as 1%o of all TK mutants to be amorphic, we still have a figure of approximately 1 mutation per 100 alkylations, raising the possibility of a 99 %o level of repair efficiency. It would appear from the foregoing that the TK+/-.JTK-/-L5178Y mouse lymphoma mutagen assay system is (1) faster than a breeding mammal; (2) able to eke much data from a simple experiment; (3) more relevant than Salmonella; and (4) able to distinguish quantitatively among the mutagenicities of closely related analogs (e.g., EMS, MMS; hycanthone, IA-3). Additionally, an attempt has been made to present these data in a form which encourages comparison with other systems (since toxicity is an easily quantitated parameter) and of extrapolation to man. Weaknesses do exist and these have been briefly touched upon. But some of these weaknesses might be surmountable. Certainly the absence of host metabolism can be at least partially overcome by coupling the test system as described to a microsomal activation system such as that described by Malling and Frantz (9) or by utilizing these cells as indicator organisms in a host-mediated assay, as discussed by Capizzi (10) and Lee (11).

Conclusions
The mammalian cell culture system just described helps bridge the gulf between microbial-especially prokaryotic-systems and whole mammal systems presently in use for mutagenicity studies. Such a claim follows from the fact that these lymphoma cells retain both a typically mammalian architecture (down to and including the genome) and certain facets of mammalian metabolism. Quantitative aspects of mutagenesis have been emphasized and mutagens can be classified as strong, weak, or of intermediate potency on the basis of their mutagenicity: growth inhibition curves. In an attempt to improve upon growth inhibition as a dosimetric measure, we have determined the relationship between extent of EMS-induced alkylation and resultant mutant frequency, and plan on extending such studies to other mutagens. Finally, a number of hycanthone analogs are being studied; one of these has been shown to exert possibly negligible mutagenicity at concentrations which are highly inhibitory to growth. It is hoped that such studies can lead to a safe hycanthone substitute for the control of schistosomiasis.
This workshop has demonstrated that basic agreement presently exists on the desirability of multiple test systems; multiple test systems exist, each with their characteristic strengths and weaknesses, for screening and safety testing programs; a possible hierarchv of such programs has been described by Bridges (12). We can move on to the real world where mutagens do not attack in single file but en masse; where the target cells are not intraperitoneal, but gonadal; where the issue is not only dead fetuses, but an imperceptibly debilitating population; and where Salmonella sapiens does not exist.