Quantitative and molecular analyses of genetic risk: a study with ionizing radiation.

Mammalian cells in culture have been used to study the genetic effects of physical and chemical agents. We have used Chinese hamster ovary (CHO) cells, clone K1-BH4, to quantify mutations at the X-linked, large (35 kb) hypoxanthine-guanine phosphoribosyltransferase (hprt) locus (the CHO/HPRT assay) induced by environmental agents. By transfecting an hprt-deletion mutant CHO cell line with the plasmid vector pSV2gpt, we isolated a transformant, AS52. AS52 cells carry a single functional copy of an autosomal, small (456 bp) xanthine-guanine phosphoribosyltransferase (gpt) gene (the bacterial equivalent of the mammalian hprt gene; AS52/GPT assay). We found that ionizing radiations such as X-rays and neutrons and oxidative genotoxic chemicals such as Adriamycin, bleomycin, hydrogen peroxide, and potassium superoxide are much more mutagenic to the gpt gene in AS52 cells than to the hprt locus in K1-BH4 cells. The hypermutability of the gpt gene probably results from a higher recovery of multilocus deletion mutants in AS52 cells than in K1-BH4 cells, rather than a higher yield of induced mutants. These results demonstrate that the use of the hprt locus alone could lead to an underestimate of the genetic risk of these agents. Analyses of the mutation spectrum using a polymerase chain reaction-based deletion screening and DNA sequencing procedure showed that a high proportion of HPRT- and GPT- mutants induced by X-rays carry deletion mutations. Thus, both the mutant frequency and mutation spectrum need to be considered in assessing the genetic risk of ionizing radiation and oxidative genotoxic chemicals.


Introduction
Sixty-five years ago, Muller discovered that X-rays induce mutations in Drosophila rnelanogaster (1). Since then, the genetic effects of radiation have been extensively studied in a wide spectrum of biological systems (2,3). Over the past quarter of a century, several mammalian cell mutational systems have been developed for studying chemical and radiation mutagenesis (4,5). These include the endogenous hypoxanthine-guanine phosphoribosyltransferase (hprt), adenine phosphoribosyltransferase (aprt), thymidine kinase (tk), and the transgenic xanthineguanine phosphoribosyltransferase (gpt) gene.
Most studies on the mutagenic effects of ionizing radiation are concerned with the induction of mutant frequency 'Department of Preventive Medicine and Community Health, University of Texas Medical Branch, Galveston, TX 77555-1010. 2Department of Radiology, University of Texas Health Science Center, San Antonio, TX 78284.
'Biological and Medical Division, Argonne National Laboratory, at these genetic loci. Quantitative analyses of radiationinduced mutant frequency have been used to assess the potential genetic risk imposed by radiation. Recent advances of mammalian molecular genetics has enabled analyses of radiation-induced gene mutations at the DNA sequence level.
In this paper, we present aspects of our quantitative and molecular studies of radiation-induced mutations in the hprt locus in Chinese hamster ovary (CHO) cells clone Ki-BH4 (CHO/HPRT assay [6]) and in the gpt gene in a CHO cell derivative, AS52 (AS52/GPT assay [7]). Evidence is presented to demonstrate that both the mutant frequency and the molecular spectrum of gene mutations need to be considered in the assessment of radiation-induced genetic risk.

Materials and Methods
Cell Lines, Culture Conditions, and Radiation Mutagenesis K1-BH4 and AS52, a subelone and derivative of the CHO-Ki cell line, respectively, were used in all the experiments to be described. Whereas K1-BH4 cells contain an endogenous hprt gene (6), the AS52 cells contain a transgenic gpt gene (7). Cell cultures were maintained in Ham's F12 medium containing 5% fetal calf serum (F12FCM5) in a 5% CO2-95% air incubator at 370C with 100% humidity. To decrease the background of HPRTmutants, K1-BH4 cells were treated with F12 medium containing aminopterin (HAT medium), and to decrease of the background of GPTmutants, AS52 cells were treated with F12 medium containing adenine, aminopterin, and mycophenolic acid (MPA medium) for 2 days as we described earlier (6,7). Both cell types were then grown in F12FCM5 medium for another 2 days before irradiating these cells with X-rays (8,9).
Our previously published procedure was used to determine both the radiation-induced cytotoxicity and mutagenicity (10). Cytotoxicity was expressed as percent of surviving clonable cells relative to that of untreated control(s). Mutant frequency was calculated as the number of 6-thioguanine-resistant (TGr) colonies per 106 clonable cells at the end of 7 of days phenotypic expression time. Independent TGr mutants were isolated (9,11) for the analysis of the molecular spectrum of gene mutations described below.

Molecular Analyses of the Mutation Spectrum at the hprt and gpt Loci
For the analysis of the molecular spectrum of radiationinduced mutations at the hprt locus, our recently developed polymerase chain reaction (PCR)-based comprehensive procedure was used (9,11). This procedure includes direct sequencing ofPCR-amplified hprt cDNA for locating point mutations in the expressed coding sequences, multiplex PCR-amplification of all nine hprt exons for screening large deletions, and direct sequencing of PCR-amplified hprt exons and their flanking regions for detecting intronic mutations resulting in mRNA splicing errors. To analyze the mutation spectrum of the gpt gene in AS52 cells, we have adapted the nested PCR amplification method (9,12).

CHO/HPRT Mutation Assay
In 1974, we used CHO cells, clone K1-BH4, to develop a quantitative mutation assay at the hprt locus, the CHO/ HPRT assay (6). The quantitative nature of this assay has been used to determine the mutagenicity of radiation and chemicals and to study mechanisms of mammalian cell mutagenesis (4,13,14).
In our studies using the CHO/HPRT assay with physical agents, we found that ultraviolet (UV) light is a strong mutagen; it causes a linear dose-dependent increase in mutagenicity (15). However, X-rays are relatively weakly mutagenic. The mutant frequency was found to increase from approximately 30 x 10-6 mutants per cell at 100 rads to 85 x 10-6 mutants per cell at 600 rads (with a spontaneous mutant frequency of 9 x 10-6 mutants per cell).
The shape of the dose-response curve could not be defined adequately (16). One possible explanation for the apparent weak mutagenic response of CHO K1-BH4 cells to X-rays could be a low recovery of multilocus deletion mutants induced by X-rays rather than a weak mutation induction per se (17).

AS52/GPT Assay
In 1982, we transformed an X-ray-induced, hprtdeletion subelone K1-BH4 cells with a plasmid vector, pSV2gpt. The gpt gene, which codes for the enzyme GPT, is the bacterial equivalent of the mammalian hprt gene. We obtained a transformant, AS52, that carries a single functional copy of the gpt gene stably integrated into the highmolecular-weight DNA of the host (7). The AS52 cell line fulfills the classical requirements for specific-locus mutation assays in mammalian cells (18).
We found that X-irradiation is equally toxic to both AS52 cells and the parental K1-BH4 cells. However, X-rays are approximately 10 times more mutagenic to the gpt gene in AS52 cells than to the hprt locus K1-BH4 cells (Fig. 1). Thus, AS52 cells are hypersensitive to mutations induced by X-rays. Southern blot analysis showed that the proportion of the TG-resistant mutants induced by X-rays that carry a total loss of the gpt gene in AS52 cells is much higher than that of the hprt gene in K1-BH4 cells. Small deletions are the most frequent type of deletion in spontaneous mutants (8,9,12).
Because X-irradiation induced predominantly deletion mutations in these cells as analyzed by Southern blot hybridization (8), the observed low mutagenic activity of X-irradiation in K1-BH4 cells, where the hprt gene exists in a hemizygous state, is likely due to lethal events associated with multilocus deletions, reducing the viability of the induced TG-resistant mutants. The gpt gene is most likely integrated in the heterozygous state in one of the autosomes of AS52 cells. In AS52 cells, a multilocus deletion induced by X-irradiation could produce only a hemizygous state in a vital gene that flanks the gpt gene, resulting in a higher probability of recovering induced TG-resistant mutants rather than to an actual reduced mutation induction (1 7,19,20).
We found that neutrons are also mutagenic to the hprt gene in CHO cells. The mutant frequency was found to increase from approximately 35 x 10-6 mutants per cell at 20 rads to 100 x 10-6 mutants per cell at 100 rads (with a spontaneous mutant frequency of 12 x 10-6mutants per cell). The pattern of X-ray-induced differential mutagenic response between the hprt locus in K1-BH4 cells and gpt gene in AS52 cells was also produced by irradiation with neutrons. Neutron irradiation resulted in nearly identical toxicity to both cell types; however, neutrons were approximately 10 times more mutagenic to the gpt gene than to the hprt locus (17).

Evidence for Reactive Oxygen Species Inducing Mutations
Reactive oxygen species (ROS) are being implicated in the toxic action of ionizing radiation and oxidative chemicals (21). If ROS were to mediate the mutagenic effects of  radiation such as X-rays and neutrons, then oxidative genotoxic chemicals such as streptonigrin, Adriamycin, and bleomycin, which are known to produce superoxide and hydroxy radicals, would be expected to be equitoxic to both cell types and more mutagenic to the gpt gene in AS52 cells than to the hprt locus in K1-BH4 cells. Likewise, potassium superoxide and hydrogen peroxide, which are ROS themselves, should be more mutagenic to the gpt gene in AS52 cells and equitoxic to both cell types. The results of our experiments fulfilled such expectations (Fig.  1). We have previously shown that agents such as ethyl methanesulfonate (EMS), ICR-191, and W-light, which do not produce ROS, do not elicit differential mutagenic response in both cell types (8,22) (Fig. 1). Taken together, these experiments support the view that one of the mechanisms responsible for the mutagenic effects of ionizing radiation and oxidative genotoxic chemicals is mediated through ROS (17,19,20).

Molecular Analyses of Gene Mutations
Southern blot analysis shows that ionizing radiation induces predominantly deletions at the hprt locus in CHO cells clone . Mutations undetectable by Southern blot analysis have largely been uncharacterized because appropriate large-scale molecular genetic pro-  cedures to analyze the mutation spectrum ofthe hprt locus at the DNA sequence level have been developed only recently (9,11). A preliminary analysis of spontaneous mutations of the gpt gene in CHO cell subclone AS52 showed that a high proportion of the spontaneously arising mutations are deletions (12).
Recently, we established a PCR-based comprehensive procedure for molecular analysis of mutation spectrum at the hprt locus in CHO cells, clone K1-BH4 (9,11).This procedure includes a) multiplex PCR amplification of all nine hprt exons for deletion screening, b) direct sequencing of PCR-amplified hprt cDNA for identification of point mutations at the coding region, and c) direct sequencing of PCR-amplified individual exon and flanking regions for determining splicing mutations (Fig. 2).
Results from the sequence analysis of 43 spontaneous GPT mutants in AS52 cells showed that the great majority (68%) of mutations is deletion. Among these deletion mutants, 76% have lost the entire gpt gene. Other mutation types include 14% single base substitution, 9% frameshift (±+ 1 bp) and 9% insertion (9). We found a 3-bp deletion hot spot as reported by others (12).
We screened for X-ray-induced deletions and found that X-rays induce primarily deletions (9). Among 41 HPRTmutants induced by X-rays at 400 rads, 51% of the mutants had lost the entire gpt gene. Analyzing 28 mutants induced by 600 rads of X-rays, the proportion of total deletion increased to 68%. It appears that the induction of deletiontype mutations could be dose dependent. Seven percent of the HPRTmutants induced by X-rays at 600 rads exhibits partial deletion ( Table 2). Analyses of 25 GPTmutants induced by X-rays (400 rads) showed that 80% of them have lost the entire gpt gene and another 4% exhibits a partial deletion ( Table 2).
These data on X-ray-induced mutation spectrum of both the hprt and gpt gene (9) are consistent with our earlier findings from cellular mutagenesis in which we found that the majority of X-ray-induced mutants at these two loci were found to be deletion as analyzed at the cellular level (17,19,20).

Quantitative and Molecular Analyses of Genetic Risk
Our studies demonstrate that ionizing radiations and oxidative genotoxic chemicals, both of which are known to generate ROS as a major mechanism, are approximately 5-10 times more mutagenic to the heterozygous gpt gene than to the hemizygous hprt locus. Thus, these agents could be considered a potent mutagen to one genetic marker (gpt) and a weak mutagen to another (hprt).
Demonstration that a high proportion of mutants induced by radiation and ROS producers may not survive to be scored as mutants in the K1-BH4 cells (CHO/HPRT assay) implies that assessment of genetic risk for these agents using hemizygous genetic markers such as hprt might be underestimated. This also leads to consider a need to reevaluate assessment of genetic risk relative to ionizing radiation. Our studies on the relationship between mutation induction and mutant recovery as analyzed by Southern blot hybridization and PCR procedure suggests a role of molecular genetics in the quantitative assessment of the mutagenic risk.