Ubiquitous, heritable damage in cell populations that survive treatment with methotrexate

Abstract

A permanent line of mouse embryo fibroblasts was treated with concentrations of the anticancer drug methotrexate (MTX) that left 20–50% surviving colonies. The surviving population initially multiplied at a much slower rate than controls after subculture in the absence of the drug, and required 9–12 days of serial subculture, with selective growth of the faster growing cells, to approximate the control rate. To determine the distribution of growth rates of cells in the original posttreatment populations, many single cells were isolated in multiwell plates immediately after the treatment period, and the resulting clones were serially subcultured. Most of the control clones underwent about 2 population doublings per day (PD/D). Almost all the survivors of MTX treatment multiplied at heterogeneously reduced rates, ranging from 0.6 PD/D to as high as control rates for a very few clones. They maintained the reduced rates through many subcultivations. The heritability of the reduced growth rates indicates that most cells that retain proliferative capacity after treatment with MTX carry random genetic damage that is perpetuated through many divisions of their progeny. Similar results have been described for cells that survive x-irradiation, and suggest random genetic damage is a common occurrence among cells in rapidly growing tissues that survive cytotoxic treatment. It also occurs in serial subcultures of cells that had been held under the constraint of confluence for extended periods, which suggests that the accumulation of random genetic damage to somatic cells during aging of mammals underlies the reduction of growth rate and function of the cells that characterizes the aging process.

The classical method for evaluating the effect of deleterious treatments on cells is based on the proportion of colony forming survivors from a number of single cells plated out during or after the treatment (1). Variations of this method have been widely applied in studies of toxicology, genetics, and aging of animal cells. A common assumption arising from this methodology is that the great majority of colony forming survivors of deleterious treatments are unchanged by the treatment except for a small fraction of mutants. Sinclair (2), however, found that a high proportion of clones that survive x-irradiation multiply at a slower rate than the original population and continue to do so indefinitely. It was then pointed out that the usual parameter of scoring survivors of radiation as a certain minimum clone size is an arbitrary definition of reproductive integrity, and it is more meaningful to study the full clone size distribution (3). In addition to a reduced growth rate, a high proportion of the survivors of x-irradiation exhibit a persistent incidence of delayed reproductive death (4, 5), and increases in frequency of mutations (6), sensitivity to mutagenic treatment (7), and probability of neoplastic transformation (8). The findings with x-irradiation have not been extended to deleterious chemicals. However, we discovered that some apoptotic cell death occurred in populations of NIH 3T3 cells that were maintained under the growth constraint of confluence for prolonged periods without chemical treatment, and the bulk of the surviving population multiplied upon subculture at a reduced rate for many cell generations (9, 10). As in the case of x-irradiation, there was a reduction in size of many of the colonies produced by cells that were seeded after several days of passage at low population density to allow recovery from the inhibitory effects of confluence (9, 11). Clones isolated after reseeding long-term confluent cultures maintained their reduced growth rates indefinitely (12). However, when a new calf serum was employed during the period of confluence and subsequent subculture, the reduced growth rate exhibited quantitative fluctuations (13).

To minimize the vagaries of the purely physiological variations associated with population density, batch of serum, and secular drift among the cells, we decided to examine heritable effects on growth of cells that survived treatment with cytotoxic drugs. For this work, we chose the anticancer drug, methotrexate (MTX), because much information was available regarding its effects on cells from studies of gene amplification and apoptosis (14–16), and from information about its toxicity in humans treated for cancer, psoriasis and rheumatoid arthritis (17, 18). However, no systematic accounts have appeared about lasting effects of methotrexate on the growth rate of cells that survive single treatment with MTX or other cytotoxic drugs. Our results show that there are heterogeneous, heritable reductions in growth rate distributed throughout the population of cells that survive MTX treatment. We discuss the significance of these results for understanding the contribution of population-wide, sublethal genetic damage to long-term problems that arise in treatment of cancer, psoriasis, and rheumatoid arthritis, as well as the continuous decline in cell function that accompanies aging in mammals.

MATERIALS AND METHODS

The NIH 3T3 line of mouse embryo fibroblasts (19) was maintained in a continuous state of exponential multiplication by three-times weekly subcultivation at low population density in 100-mm plastic Petri dishes as described (9, 10). The growth was in molecular, cellular, and developmental biology 402 medium (MCDB 402) (20) with 10% calf serum (vol/vol). Cells were enumerated electronically at each subculture in a Coulter counter. To initiate experiments, 10⁴ cells were transferred to 60-mm culture dishes in growth medium containing either 0, 200, or 1,000 nM MTX (Sigma) and incubated at 37°C for 72 h. They were then washed with 5 ml Tris-saline buffer, trypsinized, counted, and subcultured in the absence of MTX. To determine directly the effect of pretreatment with MTX on the growth rate of the population, the cells were seeded at 5 × 10³ cells per 60-mm dish and serially subcultured in the same way every 3 days. Population doublings per day (PD/D) were determined from the ratio of cell number at 3 days to the number seeded. Immediately after the 3-day treatment period, the cells were also subcultured at sufficiently small numbers to produce discrete, countable colonies in 6.5 days. The colonies were then washed with Tris-saline buffer, fixed in Bouin’s reagent, washed again, and stained with 4% Giemsa buffered at pH 7.0.

To determine the growth rates of individual cells and avoid cross-contamination by other cells in control and MTX-treated populations, they were seeded immediately after the treatment period at 1 cell per well in 96-well plates in 0.07 ml of growth medium. The morphology and multiplication of each clone were observed microscopically from 4 to 8 days after seeding in one experiment and from 4 to 14 days in a second experiment. Beginning at 4 days, those continuing to multiply were considered for expansion with emphasis placed on selecting a representative cross section of growth rates. The cells were trypsinized in the wells in subsequent days, depending on their observed growth rate, transferred to 60-mm dishes, and passaged until they reached populations of 5 × 10⁵–10⁶ cells. They were then started on four successive passages of 10⁴ cells at 3-day intervals, the cell count at each passage serving to quantitate growth rates. At the fifth passage, samples of each clone were seeded at 50 cells on each of four 60-mm dishes for 6.5 days when two were stained and two were counted for another estimate of growth rate. We then selected one or two fast, medium, and slow growing clones from each group for four additional subcultures at 3-day intervals for further determinations of growth rate. These were also used to compare the growth inhibitory effects of treatment with 200 nM MTX of cells with different growth rates and prior histories of exposure to MTX.

It should be noted that MCDB 402 medium contains 1.0 μM each of adenine, thymidine, and folinic acid that circumvent the deleterious effects of MTX (21, 22) and require the use of relatively high concentrations of the drug to produce those effects. Similar effects to those described here are produced in the absence of the three additives, but require only about 1/50th the concentrations of MTX used here (J. Koo, M.C., and H.R., unpublished data).

RESULTS

Growth Rates of Cell Populations During and After MTX Treatment.

Fig. 1 shows the rate of multiplication of untreated control and MTX-treated cells during the 3-day treatment period and in successive 3-day subcultures up to 15 days after removal of MTX. The untreated cells underwent about 2 PD/D in each 3-day interval. During treatment with 200 or 1,000 nM MTX, the cells multiplied, respectively, at about 60% and 25% the rate of the untreated cells, and successively faster in each subculture. Both sets reached almost to the control rate 9 days after removal of the drug, and stayed at that level in later subcultures. In a second experiment, it required 12 days for cells treated with 1,000 nM MTX to approximate the growth rate of control cells (data not shown). Both results indicated that only a small fraction of the cells that survived MTX treatment retained the rapid multiplication rate of the untreated cells.

Figure 1.

Effect of MTX treatment on the rate of cell proliferation. The PD/D of cells were determined during 3-day treatment with 0, 200, and 1,000 nM MTX, and in five successive 3-day subcultures without MTX. The −3 day point on the abscissa represents the treatment period, and each of the other values represents the first day of serial 3-day subcultures without MTX. ○, Untreated control; ▵, 200 nM MTX; □, 1,000 nM MTX. The error bars represent the range of cell counts of duplicate dishes.

Isolation and Characterization of Clones Initiated Immediately After MTX Treatment.

To properly characterize the treated populations, it is necessary to clonally isolate cells immediately after treatment, and allow each clone to multiply in isolation before transferring each to a separate dish to expand their numbers. This was done by seeding an average of one cell per well in multiwell plates. Each growing clone was trypsinized from its well and expanded to numbers large enough for frequent serial passage at low density to obtain repeated determinations of their growth rates. The fraction of cells that could multiply after treatment with 200 and 1,000 nM MTX was 0.45 and 0.23 of the control as determined by the number of positives in the multiwells. Clones were chosen at random to ensure that there was a representative cross section of clonal growth rates from each group. We expanded to large populations 22 control clones, 24 that had been treated with 200 nM MTX and 8 treated with 1,000 nM MTX. The averages of growth rates from four successive passages from every clone of each of the three groups are graphically displayed in Fig. 2A. Nineteen of the control clones averaged between 1.8 and 2.1 PD/D with a sharp peak at 2.0 PD/D, but 3 clones multiplied at distinctly lower rates. Clones from the group treated with 200 nM MTX were heterogeneously distributed over a wide range of growth rates, all but 2 of which were lower than the modal value for the control clones. All 8 of the surviving clones derived after treatment with 1,000 nM MTX multiplied at lower rates than the control peak and were heterogeneously distributed down to very low values.

Figure 2.

Proliferation rates of clones isolated immediately after the MTX treatment period. After a 3-day treatment, an average of 1 cell per well from each treatment group was seeded in 96-well plates as described. Each data point represents the average PD/D of four serial subcultures of a clone, categorized by intervals of 0.1 PD/D. (A) ○, Untreated, 22 clones; ▵, 200 nM MTX, 24 clones; •, 1,000 nM MTX, 8 clones. (B) ○, Untreated, 19 clones; •, 1,000 nM MTX, 27 clones

A second cloning experiment was done in which cells were treated with 1,000 nM MTX. In this experiment, all 19 of the untreated clones multiplied between 1.6 and 2.1 PD/D with a peak at 1.9 PD/D (Fig. 2B). The 27 clones from the MTX-treated cultures multiplied between 0.6 and 1.8 PD/D, with none multiplying as fast as the modal value of the untreated clones. Most colonies from the untreated population were uniformly dense in appearance in contrast to the heterogeneously lighter colonies seeded after MTX treatment (Fig. 3). The cloning efficiency of the MTX-treated cultures was about 0.2 that of the controls. The occasional dense colony of MTX-treated cells represents a fast growing clone that would become the dominant element in subculturing an uncloned population, as seen in Fig. 1. The clonal isolation procedure reveals that virtually all the MTX survivors have a persistently reduced growth rate, indicating genetic damage.

Figure 3.

Colonial morphology of untreated and MTX-treated cells. Cells were untreated or treated with 1,000 nM MTX for 3 days, washed free of MTX, and subcultured at 50 cells (untreated) or 100 cells (MTX-treated) per 60-mm dish for 6.5 days in the absence of MTX before fixation and staining. The dishes on the left originated from an untreated culture, those on the right from an MTX-treated culture. Note the uniformity in size and high density of the untreated colonies in contrast to the heterogeneously reduced size and density of the MTX-treated colonies.

Thirteen of the clones representing high, moderate, and low growth rates in each of the three experimental groups of the first experiment (Fig. 2A) were chosen for extended study. Fig. 4 shows the morphology of 12 of the clones seeded at the second passage after the initial clonal expansion. (The slowest growing clone derived from the original 200 nM MTX treatment produced barely visible colonies and is not shown.) It can be seen that the fastest growing clones selected from each group produced dense colonies, but those from the untreated control were somewhat larger than those that had been treated with MTX. The colonial morphology of 15 of the 22 control clones was very similar to that of the 2 fastest growing controls of Fig. 4. The remaining 4 control clones in Fig. 4 were the slowest of the 22 clones of this group. The slow growing MTX clones, of course, constituted a much larger proportion of their group (see Fig. 2A). All 13 clones were maintained in serial passages for a total of 43 days after the initial treatment. Their growth over the entire period is shown in Fig. 5. The three slowest growing clones of the control group maintained their initial low growth rate for the first five measurements, but grew at faster rates in later measurements. The late increase in growth rate of the slow growers suggested that faster growing variants arose during clonal growth and eventually dominated the population. By contrast, the slower growing clones induced by MTX treatment maintained their reduced growth rate through all successive measurements. All the clones from the second experiment (Fig. 2B) were given at least 6 serial passages, and a representative selection of them were given up to a total of 16 passages (data not shown). Most MTX-treated clones were speeded up slightly in the course of the passage, though this was not obvious in the first 4 passages. They were still behind the control growth rate by 0.2–0.5 PD/D at the end of 8 weeks. Presumably the increase in growth rates, as in the first experiment, resulted from the appearance of faster growing variants in the clonal populations. The general stability of the reduced clonal growth rates reinforces the evidence that they were of genetic origin.

Figure 4.

Colonial morphology of selected clones. In the second subculture after clonal expansion, each clone was seeded at 50 cells per 60-mm dish and grown for 6.5 days before fixation in Bouin’s reagent and staining with Giemsa. The clones shown were those selected to illustrate rapid, intermediate, and slow proliferation in the three groups, and their number is not representative of the actual proportion of clones with those rates as seen in Fig. 2A. The average rates (in PD/D) of proliferation of designated clones averaged over four determinations are as follows: untreated, top two rows: 2C, 2.07; 8H, 2.00; 3F, 1.77; 8D, 1.56; 4A, 1.17; 6C, 1.22. At 200 nM MTX, third row: 9E, 1.85; 1G, 1.90; 8D, 1.37. At 1,000 nM MTX, bottom row: 11H, 1.73; 12A, 1.38; 6C, 0.95.

Figure 5.

Cumulative growth rates of clones selected to undergo extended periods of proliferation. The 13 clones shown are those selected in the first experiment for determination of growth rates over a 6-week period. See legend to Fig. 4 for average PD/D for the first four passages, except for the 200 nM MTX clone 2B = 1.21 PD/D. (Upper) Untreated control: ▪, 2C; ○, 8H; •, 3F; □, 8D; ▴, 4A; ◊, 6C. (Lower) At 200 nM MTX: ▴, 9E; □, 1G; ⧫, 8D; ○, 2B. At 1,000 nM MTX: •, 11H; ▵, 12A; ▪, 6C.

DISCUSSION

Judging from the clonal growth rates in Fig. 2, most of the cells treated with 200 nM MTX sustained damage sufficient to irreversibly reduce their growth rates. In terms of lethality, the treatment was relatively mild because almost half the proportion of the control cells that grew continuously retained that capacity after treatment with 200 nm MTX. Although MTX blocks the enzymatic activity of dihydrofolate reductase, it is most likely that the damage to our cells was ultimately to the genetic material because it remained manifest in reduced growth rates over a 6-week period, or more than 40 divisions, even for the slowest growing clones. By blocking the reduction of folate, MTX treatment is equivalent to starving cells for folate which markedly increases the number of single-strand breaks in DNA (23). Folate starvation also produces chromosome breaks as indicated by micronuclear formation (24), and increases the metastatic potential of murine melanoma cells (23). MTX treatment itself has been shown to produce single- and double-strand breaks in a variant of the NIH 3T3 cells used here (25). The ubiquitous nature of the random genetic damage we find after low lethality treatment with MTX is reminiscent of the high proportion of cells with chromosome aberrations after ionizing radiation at doses an order of magnitude lower than the mean lethal dose (26, 27). These aberrations are detected in the mitoses that appear within 2 h after irradiation, and represent damage induced during the G₂ period of the cell cycle. Many of the aberrations are chromatid and chromosome gaps that are likely to be repaired in the next round of DNA synthesis in viable cells (28). If the repair is imperfect, however, the damage becomes permanent and could underlie a reduction in growth rate of a clone. Such perpetuated genetic damage could lead to a variety of pathologies. In this sense, intermediate doses of a harmful agent can be more damaging in the long run than large doses, because the former may leave more mutated cells (26, 29). This is an important consideration in the therapeutic use of MTX where serious side effects are often encountered not only in the “high dose” treatment of cancer (17) but in the “low dose” treatment of chronic inflammatory conditions like psoriasis and rheumatoid arthritis (18). Such effects as myeloid suppression, gastrointestinal symptoms, hepatotoxicity, alopecia, and osteopathy might be perpetuated in chronic form long after drug administration as a result of subnormal performance of heritably damaged cells. The well-known occurrence of chromosome aberrations in MTX-treated cells (14, 30) might explain the occurrence of lymphoma in rheumatoid arthritis after low dose treatment with MTX (31). The damage produced in culture by MTX is proportional to the growth rate of cells at the time of exposure to the drug (ref. 32; P. Ng and H.R., unpublished data). Therefore, it is to be expected that renewing tissues in the body with fast growing stem cells such as intestine and bone marrow would be the most severely damaged, whereas nerve and striated muscle would suffer minimal effects.

The evidence presented here for heritable damage in MTX-treated cells raises the question of whether similar effects are produced by other cytotoxic drugs that are commonly used in cancer and other conditions. Hydroxyurea, for example, which is used in cancer chemotherapy (17), is known to produce chromosome aberrations and gene amplification in cultured cells (33, 34). A number of other anticancer drugs produce strand breaks in DNA (35, 36). It would therefore be of great interest to determine whether such agents produce the type of population-wide, heritable reduction in growth rate of cells described here for MTX. The high frequency of heritably reduced growth rate in an MTX-treated cell population is consistent with the recent finding that insertion of a transposable element into any one of more than half of 268 genes tested in yeast results in a growth disadvantage to the cells carrying the altered gene (37). Other processes shown to be altered by local mutation over a large fraction of the genome are sporulation in Bacillus subtilis (38), penicillin production by Aspergillus nidulans (39) and the development of the several organs in Drosophila melanogaster (40). In addition, extensive DNA rearrangement is found after exposure to MTX, and the DNA sequences seem to differ in each cell line (30). They include nondihydrofolate reductase as well as dihydrofolate reductase sequences. These aberrations are likely to result in reduced growth rate in affected cells. As already noted, they occur at extremely high frequency in cells treated with ionizing radiation even at doses of low lethality (26, 27). The implication for the MTX-treated cells is that heritable changes at sites distributed throughout the genome, as well as chromosomal aberrations, have the capacity to reduce the growth rate of the affected cells.

The evidence that a variety of treatments to cells—i.e., x-irradiation (2, 3), long-term growth inhibition under crowded conditions (12), and low lethality MTX treatment—heritably lower the growth rates of most cells in a population may have relevance for the age-related reduction in growth rate of cells in many tissues of mammals (41–45). The frequency of somatic mutation in single genes of young animals and humans is in the neighborhood of 10⁻⁵ per cell and increases with age (46). Assuming that the observed mutation frequencies are similar for all genes, it is likely that a large fraction of cells in aging animals carry one or more mutations, which would account for the reduction in growth rate of cells with age. Unlike the strong selection for fast growing cells seen here in populations kept continuously growing at maximal rates (Fig. 1), damaged cells might persist indefinitely in tissues where growth is regulated. It seems likely that such cells would be genetically unstable. Many survivors of x-irradiation have highly increased mutation frequencies (6) and are hypersensitive to mutagenic treatment (7). Cells selected from a population for amplified dihydrofolate reductase genes undergo increases or decreases in gene copy number much more rapidly than the parental population (47). As a result, cells exposed to stepwise increases in MTX concentration develop resistance to high doses of the drug 7 times faster than those exposed to a single high dose (48). The occurrence of lymphoma in MTX-treated patients (31) and the increased frequency of metastasis in MTX-treated mouse murine melanoma cells (23) may therefore be related to the genetic instability induced by the treatment.

It should be emphasized that the detection of the full extent of heritably reduced growth rate after MTX treatment depends on clonal isolation of cells immediately after the treatment. If the treated cells are left in large populations, even a small minority of faster growing cells soon overshadows the slow growing cells. The multiwell method of clonal isolation also avoids contamination of the slow growing ones by the occasional cell that detaches from the fast growing clones, an ever present danger in the conventional ring cloning procedure. An advantage of low lethality treatment is its similarity to most environmental hazards. If recent experience with radiation-induced chromosome aberrations (26, 27) is any guide, population-wide, heritable reduction of growth rate in mammalian cells will prove to be orders of magnitude more sensitive to potentially mutagenic and carcinogenic agents than single gene bacterial tests (49).

ABBREVIATIONS

MTX

methotrexate

PD/D

population doublings per day

MCDB 402

molecular, cellular, and developmental biology medium 402