An Assay to Measure CD59 Mutations in CHO A_L Cells using Flow Cytometry^,

Abstract

Background

A sensitive mammalian cell mutation assay was developed previously using a Chinese hamster ovary cell line (CHO A_L) that stably incorporates human chromosome 11. The assay measures mutations in the CD59 gene on chromosome 11 but it requires the use of rabbit complement and colony growth for mutant selection. We have developed a more rapid flow cytometry based mutation assay with CHO A_L cells that uses monoclonal antibodies against CD59 to detect mutants and does not require colony formation.

Methods

CHO-A_L cells were treated with gamma radiation or N-methyl-N'-nitro-N- nitrosoguanidine (MNNG), then allowed to grow for various times for mutant expression. Cells were labeled with monoclonal antibodies against CD59 and analyzed by flow cytometry.

Results

Negative and positive populations were separated by over 100-fold. Mixing various proportions of CD59 positive and negative cells demonstrated that the assay is highly linear (r²=0.9999) and sensitive (<0.05% background mutants). The yield of CD59 inducible mutants was linearly related to dose for both a clastogen (gamma radiation) and point mutagen (MNNG). The mutant yield was both time and treatment specific.

Conclusions

Mutations induced by genotoxic agents can be rapidly and sensitively measured in CHO A_L cells using flow cytometry.

Introduction

Studies of mutagenesis in cultured mammalian cells can help predict carcinogenesis based upon evidence that mutation plays a vital role in carcinogenesis and the correlation between the ability of an agent to induce mutations in vitro and its carcinogenic potency (1). Mutagenesis data are used for heritable risk assessment, as part of the body of information used to make a decision to trigger oncogenicity testing, and as part of the weight-of-evidence for determining a carcinogenicity classification for a chemical when a long-term bioassay has not been performed(2).

Current mammalian cell mutation assay systems, namely the Mouse Lymphoma Assay (MLA) based on the thymidine kinase gene (3-5) and the Chinese hamster ovary hypoxanthine guanine phosphoribosyl transferase (HGPRT) assay (6,7), effectively measure specific types of mutations, but are limited in sensitivity by the requirement that flanking genes on the chromosome remain functional for cell survival (8). If the mutation extends beyond the reporter gene location, it may then cause cell death and the mutation is not scored (9). This is especially true in the HGPRT assay, since the gene is located on the X-chromosome and flanking genes may not be rescued by a homologous chromosome. Large deletions, for example, are likely to kill the cell and alter the accurate mutant yield induced by a genotoxic agent, reducing the assay sensitivity (2).

In light of these difficulties, Puck and co-workers (9,10) designed a mammalian cell mutation assay around a Chinese hamster ovary cell line (CHO A_L) that stably incorporated a single copy of human chromosome 11. The CHO A_L hybrid cells were formed by fusion of a human amniotic fluid fibroblast and a gly^- mutant of the Chinese hamster ovary CHO-K1 cell (11). They retain the normal set of CHO-K1 chromosomes and a single human chromosome 11 (12). The hybrid cells express the CD59 gene on chromosome 11 which encodes a GPI-linked surface protein, CD59, which is not expressed in normal CHO cells. Thus, mutations in CD59 lead to loss of expression of CD59 protein on the surface of the cells. This cell line has been stable for over 30 years with very little rearrangement (13,14). Waldren and co-workers subsequently used this system to assay mutagenesis from a variety of genotoxic compounds (1,8-10,12,15-25). They have found that the mutation assay is a hundred-fold more sensitive than HGPRT and a thousand-fold more sensitive than the bacterial Ames test (1). Those results reflect a major advantage of the A_L system: the cell line does not require chromosome 11 to survive except for an essential gene at the tip of the p arm (26). This makes it possible to quantify the activity of small, non-lethal doses of a mutagen like those to which human populations are likely to be exposed (1).

The original CHO A_L mutation assay system depends on rabbit complement-induced cytotoxicity against cells labeled with monoclonal antibodies against CD59 to detect mutants after clonal growth. Cells which are mutated in the CD59 gene will not bind the antibody and continue to grow in the presence of rabbit complement. The resulting colonies are counted and mutant yield is calculated. When determining exact mutant yield, researchers have to take into account the toxicity of rabbit complement. Furthermore, results vary with different lots of complement.

Flow Cytometry Mutation Assay

We propose to take this CHO A_L mammalian mutation assay and streamline it using flow cytometry, which we call the flow cytometry mutation assay (FCMA). The cells are stained with a directly-conjugated monoclonal antibody to CD59 and analyzed by a flow cytometer to measure the number of CD59^- mutant cells, avoiding the need for rabbit complement and colony growth. Not only does this eliminate the intrinsic toxicity of the rabbit complement, but it shortens the time required for the assay. Both the clonogenic and flow cytometric methods require a treatment phase (20 min - 24 hrs depending on chemical toxicity), an expression period (5-12 days) and an assay phase. The entire analysis takes 12-22 days for the clonogenic assay but only 7-14 days for the FCMA. The time and labor reduction speeds results from experiments and makes possible the analysis of more variables while allowing for lower laboratory costs.

To validate the flow cytometric mutation assay, we chose both a known clastogen (16,27) (¹³⁷Cs gamma radiation) and a point mutagen (28) (N-methyl-N'-nitro-N-nitrosoguanidine, MNNG). Calibration experiments demonstrate that the FCMA is sensitive and linear. Treatment with either radiation or MNNG results in linear mutation dose response curves, which is due in part to the ability of the assay to detect both small (single locus) and large (multilocus) mutations, allowing one to measure the effect of mutagenic agents at low and high doses (1). We assume that detection of a mutation is contingent on the cell surviving and the CD59 gene not being rescued or repaired. Rescue could occur by a translocation in which the gene is moved from its position on the human chromosome and reintegrated elsewhere in the genome (18).

Methods

Cell Culture

CHO A_L cells were originally obtained from C. A. Waldren. The cells we use contain the neomycin resistance gene integrated into chromosome 11, also known as A_LN (15,23). The cells were cultured in 1:1 ratio of Eagle Dulbecco's Minimal Essential Medium and F12 containing 10% fetal bovine serum (Sigma-Aldrich, St. Louis, MO), penicillin/streptomycin and 20 mM HEPES buffer, pH 7.4, in a humidified atmosphere of 95% air and 5% CO₂ at 37º C in T75 tissue culture flasks (Cellstar, ISC Bioexpress, Kaysville, UT). After several months, cultures were replaced with a new thaw of stock cells. If at all possible, all experiments were carried out with the same continuous culture of cells.

Reducing background mutants

The A_L cell line has a naturally-occurring spontaneous background of around 10-50x10^-5 or less when cells are properly plated and remain at the correct pH (10). Over time, or if cells become confluent or acidic, the background mutation level may increase to 100x10^-5 or above, making the mutation assay system less sensitive. To combat this problem, we employ three techniques for reducing the background: antibiotic treatment, panning, and sorting. Cells are periodically treated with neomycin (800 μM, Sigma-Aldrich, St. Louis, MO) since it effectively reduces the background if it is less than about 100x10^-5. If the background rises higher than this, the cells are panned, which is very effective in reducing background but is time consuming. To pan the cells, sterilized petri plates are labeled with a general goat anti-mouse antibody (Sigma-Aldrich, St. Louis, MO) at 20 μg/ml in Phosphate Buffered Saline (PBS). The cells are labeled with a CD59 monoclonal antibody (E7.1) provided by Dr. Waldren (2 μl anti-CD59 in PBS containing 2% FBS) and then placed in the goat anti-mouse petri plates where they incubate at 4º C for 2 hr. After treatment, the plates are gently washed with cold 2% FBS/PBS and those cells that lack the CD59 antigen are washed away. In one particular experiment, this panning procedure reduced the background from 200x10^-5 to 20x10^-5 in two successive treatments (data not shown). Another method to reduce background is to sort CD59⁺ cells with a cell sorter, with similar results as panning but at a much higher cost. Generally, cells were panned about every 3 months or when cells exhibited a high background.

Monoclonal antibody labeling

Cells were trypsinized, washed once in PBS and then in staining buffer (PBS with 1% BSA, 0.1% sodium azide). Cells were resuspended at 1x10⁶ cells/ml and 1 ml was aliquoted into 1.2 ml microcentrifuge tubes. After centrifugation at 1500 rpm (450g) in an IEC table-top centrifuge, buffer was aspirated and the cells were resuspended in the residual buffer. The PE-conjugated monoclonal antibody to CD59 (Caltag, Burlingame, CA) was added (1:40 dilution) at 50 μl per sample and incubated at 37 º C for 1 hr. Cells were then washed in cold staining buffer and resuspended in 0.5 ml cold buffer, filtered through a 40 μm nylon mesh and kept on ice prior to analysis by flow cytometry.

Flow Cytometry

Cells were analyzed with an EPICS V flow cytometer (Beckman Coulter, Fullerton, CA) using 488 nm excitation with a 515 nm long pass filter and a 575 nm band pass filter. Cellular debris was removed by gating on Forward Scatter vs. Side Scatter. This gating is critical for obtaining low background. A total of 100,000 cells were analyzed for each sample. Cells which expressed CD59 were bright (positive staining) whereas mutants were dim (negative staining) (Figure 1). Gates were set so that 97% of the negative parental cells which lacked human chromosome 11 were included in Region 1. The percentage of negative or mutant cells was determined by calculating the fraction of cells in Region 1 compared to the total number using Cicero Software (DakoCytomation, Inc., Fort Collins, CO).

Figure 1.

Flow cytometry histograms of CHO A_L cells either unstained (top) or stained with antibodies against CD59 (bottom). Region 1 is set to include 97% of the negative cells and is used for counting mutants. This gives a spontaneous background of 0.03% or about 30x10^-5 analyzed cells.

Calibration

The linearity of the FCMA was determined by mixing parental CHO cells (negative for CD59) with the CD59⁺ CHO A_L cells in two-fold serial dilutions from 4% to 0.125%. The cell mixtures were then stained with anti-CD59 antibodies and analyzed with the flow cytometer.

Mutation Assay

Cells were treated with neomycin (800 μM) for 5 days prior to treatment. The day before treatment, 4x10⁵ cells were plated in a T75 flask, giving 8x10⁵ at time of treatment. The medium was aspirated, then medium containing MNNG (Sigma-Aldrich, St. Louis, MO) in various concentrations (0.0625, 0.125, 0.25, 0.5 μg/ml) was added to the flasks and the cells were incubated at 37 ºC in 5% CO₂ for 3 hrs. The clonogenic survival after the highest dose was 20% (data not shown). The flasks were then washed twice with PBS and fresh medium was returned to the flasks. The cells were returned to the incubator for growth and mutant expression. Cells were grown for 2 days and then at least 1.5x10⁵ cells were passed, the actual number depending on the relative growth compared to the control flasks. The same passing procedure occurred on days 6, 9, and 12. Cells were assayed with flow cytometry every 2 days to determine the day of maximum mutant expression after treatment with 0.5 μg/ml MNNG. Untreated control cells were passed in parallel and were processed along with the treated samples. After determining the mutant expression period, a dose response curve was generated by treating cells with various doses of MNNG that caused 20-80% survival and measuring the mutant yield on the day of maximum expression (day 12).

Radiation Treatment

Cells were also treated with ¹³⁷Cs γ radiation (J.L. Shepherd and Associates, Glendale, CA), a well-known clastogen, to test the robust nature of the assay. Cells were plated similarly to MNNG, then irradiated with doses of 0-4 Gy (dose rate of 0.93 Gy/min at 22º C) and returned to the incubator for 2 days. Cells were passed with a minimum of 1.5x10⁵ cells per T75 flask. After determining the day of maximum mutant expression, a dose response curve was generated using doses of 1-4 Gy and assaying on the day of maximum mutant expression (day 9) .

Results

Separation of CD59 Positive and Negative Cells

In order to determine whether flow cytometry could adequately distinguish CD59⁺ CHO A_L cells from CD59^- mutants, we compared histograms of antibody-labeled CHO A_L cells with unstained CHO A_L cells as a negative control (Figure 1). We also used antibody-labeled parental CHO cells which do not contain chromosome 11 (and thus do not express CD59) as negative controls, but they had slightly less fluorescence than unstained CHO A_L cells (data not shown). This is probably because the CHO A_L cells are somewhat larger and thus the unstained cells are a better control. There was more than a 100-fold separation in fluorescence intensity. A region was set that included 97% of the negative cells. This was done to minimize the inclusion of false- negatives but to maximize the number of true mutants. The background proportion of negative cells (presumably mutants) was typically 0.05% or less (50x10^-5). These results demonstrate that the assay has sufficient sensitivity to be able to measure low levels of mutants in a CD59⁺ background.

Calibration: System Sensitivity

We then investigated whether we could use flow cytometry to accurately measure known proportions of negative cells in a positive population. Parental CHO cells lacking CD59 were mixed with CHO A_L cells in two-fold serial dilutions from 4.0% to 0.125% and stained with a monoclonal antibody to CD59, then measured with flow cytometry (Figure 2). The results demonstrate that flow cytometry can accurately count the proportion of CD59 negative cells in a highly linear fashion (r²=0.9999).

Figure 2.

Negative CD59 parental cells were mixed with CHO A_L cells in varying proportions and then labeled with anti-CD59 monoclonal antibodies and measured with the EPICS V. The X- axis represents the percentage of negative cells in the mixture and the Y-axis represents the measured percentage. The results are from 3 independent experiments. Error bars representing the standard error of the mean (SEM) are smaller than the data points.

MNNG mutagenesis

After the FCMA linearity was determined with mixed populations, we used a well-known mutagen, MNNG, to determine whether mutants could reliably be measured. It was first necessary to determine the kinetics of mutant expression. Cells were treated with 0.5 μg/ml MNNG for 3 hr, a dose giving 20% survival, and assayed for mutants at various times later (Figure 3). No excess mutants were detected by day 7, but a high level was detected on day 10 with peak expression on day 12. The mutant yield subsequently decreased by day 15.

Figure 3.

Temporal expression of mutants after treatment with 0.5 μg/ml MNNG for 3 hr. The data points are corrected for the background level of mutants from parallel untreated samples. The mutant yield is the number of mutants per 10⁵ cells measured in R1 (see Figure 1). Symbols are the mean values from 3 independent experiments and error bars represent the SEM. Results were obtained with the EPICS V flow cytometer.

Cells were then treated with various doses of MNNG that gave surviving fractions of 20%, 40%, 60%, and 80% and assayed on day 12 (Figure 4). The MNNG dose response curve increased linearly from 0.0625 μg/ml to 0.5 μg/ml (r²=0.9980) and was statistically significant from control at 0.125 μg/ml (p<0.0025).

Figure 4.

Dose response curve for MNNG treatment. Cells were treated with various doses of MNNG for 3hr, and then assayed on day 12. The minimum dose resulted in 80% survival while the maximum dose gave 20% survival. Symbols represent the mean values of 3 independent experiments and error bars represent SEM. *Significantly different from control at p<0.0025.

Radiation mutagenesis

We also analyzed the ability of the FCMA to measure mutations induced by a well-known clastogen, ¹³⁷Cs gamma radiation. The maximum day of mutant expression was evaluated by irradiating cells with 4 Gy, a dose giving 50% survival, and assaying the mutant fraction with flow cytometry at various times (Figure 5). The mutant yield remained at background levels for 4 days, but was high at 7 days, peaked at 9 days and gradually declined over the next 6 days.

Figure 5.

Temporal expression of mutants after treatment with 4 Gy gamma radiation. The data points are corrected for the background level of mutants from parallel untreated samples. Symbols are the mean values from 3 independent experiments and error bars represent the SEM.

Cells were irradiated at doses that gave surviving fractions of 50%, 65%, 75% and 90% and assayed on day 9 (Figure 5). The γ radiation dose response curve increased in a linear manner from 1 to 4 Gy (r² = 0.9958) and was statistically significant from control at 1 Gy (p<0.05) (Figure 6).

Figure 6.

Dose response curve for gamma radiation. Cells were treated with various doses of gamma radiation and then assayed on day 9. The minimum dose resulted in 90% survival while the maximum dose gave 50% survival. Symbols represent the mean values of 3 independent experiments and error bars represent SEM. *Significantly different from control at p<0.0005.

Discussion

The FCMA method was developed to enable a rapid and quantitative analysis of mutations induced by genotoxic agents that would not require the time and labor of a clonogenic assay. It was first necessary to demonstrate that flow cytometry could measure a very small proportion of negative cells in a positive background, which is opposite the usual case with flow cytometry where positive cells are measured in a background of negative cells. The separation between the CD59 positive and CD59 negative cells in Fig.1 clearly demonstrates that there is sufficient resolution to measure mutants in a background of CD59 positive cells. Background levels of mutants could be measured down to <0.05% (50x10^-5), which is as low as the background is measured using the standard complement-based cytotoxicity assay for CD59 mutants (Charles Waldren, personal communication). The mixing experiments demonstrate that the FCMA effectively measures CD59 mutants in a linear and sensitive manner.

One important issue is the criteria for setting the mutant region on the histograms. We chose to include 97% of the negative control population in the region. This is a trade-off that was chosen so that only 3% of mutants would be missed but the false-positive rate of non-mutants that have a very low fluorescence would be minimized. It is possible to use other values, 98% or 99%, for example, but the effect is to give a larger background. Using 97% was a reasonable compromise. The main issue is to use a consistent value so that all samples will be analyzed in the same way. A previous report using flow cytometry to measure mutations in the CD59 gene had a much higher background level of mutations of 0.2% (27). The separation between the positive and negative cells was much less than the results reported here. Furthermore, they used a cutoff to detect mutants at 10% of the mean value of the positive cell population. This resulted in artificially high mutant induction rates for ionizing radiation.

Mutations were induced both with ionizing radiation, a clastogen, and MNNG, a point mutagen, to determine whether the FCMA could measure a dose response for mutagenesis. Both agents resulted in linear dose response mutation curves over the range of doses studied. This shows that the FCMA assay is able to detect both large and small mutations, similar to what has been shown with the complement-based cytotoxicity assay (18,23,29). In order to demonstrate the widespread applicability of this assay for measuring genotoxicity, we are currently studying a panel of genotoxic agents and appropriate control agents.

As with all mutation assays, an expression period is necessary for mutants to be measured. We have shown that the mutant expression is time-dependent for both radiation and MNNG, but that the kinetics are quite different. The time for maximum expression was 12 days for MNNG but 9 days for radiation. One difference in this assay compared to a clonogenic assay is that dead cells will be present in the cell population for a period of time after mutagenesis and then will eventually be overgrown by the viable cells, including mutants. This will potentially affect the temporal expression of mutants. The expression time is also affected by cell cycle alterations, which can be quite different for different agents. Preliminary cell cycle analysis demonstrates that for MNNG, expression does not occur until the mutated cells begin a normal cell cycle (data not shown). An additional factor that may be reflected in the time of mutant expression with different genotoxic agents is that other physiological processes may be affected by toxic drugs that could affect the recycling of existing CD59 from the cell surface. Another important observation is that the mutant yield is reduced after the peak expression time. We do not know why it is reduced, but it may be due to a somewhat reduced viability or slower growth rate of the mutant cells. A detailed understanding of the kinetics of mutant expression requires further study, which is currently underway.

The FCMA, using the monosomic CD59 gene in CHO A_L cells, gives a higher mutant yield for 4 Gy of radiation than other mutation assays, including the A_L clonogenic assay (23), MLA (3), and HPRT (30). Both the MLA and HPRT are limited in sensitivity by having flanking genes near the analysis gene that may be deleted or mutated, causing cell death and artificially low mutant measurements. With the large chromosome 11 as a target, the A_L cell line is immune from that problem. The CHO A_L clonogenic assay is somewhat less sensitive than the FCMA, which also utilizes the CHO A_L hybrid cell line. The reasons for this are unknown but it may be due to the very different methods for detecting mutants. The rabbit complement system is much more complex and may kill some of the mutants if they have a low expression of CD59. Wedemeyer et al. (27) reported much higher mutation frequencies using CHO A_L cells (about 2.8% after 4 Gy), but their results are much higher than any previous reports. Furthermore, they reported an induction of mutants within 2 days after 3 Gy of X-rays. We do not see any mutants before 4 days after gamma-irradiation Both of their results may be due to the fact that they had less separation between mutant and non-mutant cells and set the gating to identify mutants as cells that had 10% or less fluorescence intensity than the peak of the CD59-positive cells. This is a much less stringent condition than we use and would give a much higher mutant fraction.

We have also compared our results for the mutant yield induced by MNNG with the MLA (31), the HPRT assay (3), the CHO A_L clonogenic assay (10). The FCMA measures a substantially higher mutant yield after treatment with MNNG than the other assays. In fact, we could only measure results up to a concentration of 0.5 μg/ml (80% killing) while others used doses up to 4 μg/ml. In our hands, that concentration of MNNG would lead to unacceptable cell killing.

In summary, we have demonstrated that the FCMA is robust, linear, and more sensitive than other currently used mutation assay systems. Results can be obtained more rapidly and easier than with conventional clonogenic assays. It is necessary to determine the optimal mutant expression period, which varies depending on the mutagenic agent, to get the most accurate results.