A method to distinguish between the de novo induction of thymidine kinase mutants and the selection of pre-existing thymidine kinase mutants in the mouse lymphoma assay

The mouse lymphoma assay (MLA) is the most widely used in vitro mammalian gene mutation assay. It detects various mutation events involving the thymidine kinase (Tk) gene in L5178Y/Tk+/- -3.7.2C mouse lymphoma cells. Mutants are detected using a thymidine analogue that arrests the growth of cells containing a functional Tk gene. However, there are a number of potential test chemicals that are thymidine analogues, and there is a problem when using the MLA to evaluate the mutagenicity of these chemicals. Thymidine analogues are activated by Tk before eliciting their toxicity. Therefore, any pre-existing Tk-/- mutants may avoid the toxicity of the test chemical and obtain a growth advantage over the Tk+/- cells, increasing the Tk mutant frequency (MF) in the culture via a selection mechanism. This potential mutant selection effect needs to be distinguished from de novo mutant induction in order to properly evaluate the mutagenicity of these chemicals. Here we describe a simple MLA study design that can differentiate between the selection of pre-existing mutants and de novo mutant induction. Trifluorothymidine (TFT), a thymidine analogue and the selection agent normally used in the MLA, and 4-nitroquinoline-1-oxide (4-NQO), a potent mutagen, were used to treat cells from two different Tk+/- mouse lymphoma cell cultures with different background MFs (approximately 112 and 305x10(-6)). Both agents significantly increased the Tk MFs in both the normal and high background cultures (p<0.01). In 4-NQO-treated cultures, the induced MFs (MF of treated culture-MF of control) for the cultures with different background MFs were about the same (p>0.1), while in TFT-treated cultures, they were significantly different (p<0.01). In TFT-treated cultures, the fold-increases of MF (MF of treated culture/MF of control) for the cultures with different background MFs were about the same (p>0.1), while in 4-NQO-treated cultures, they were significantly different (p<0.01). This study confirms that, when de novo mutations are induced, the induced MF is the same for cultures with normal and artificially high background MFs. In situations where the increase in MF is due solely to selection of pre-existing mutants, the "induced" MF will be a multiple of the background MF and the magnitude of the increase of the induced MF will depend upon the magnitude of the background MF. Our results demonstrate that it is possible, using this experimental design, to distinguish between chemicals acting primarily via the selection of pre-existing Tk mutants and those inducing de novo mutants in the MLA.