Effects of Folate Deficiency on the Metastatic Potential of Murine Melanoma Cells1'2

Home , Folate deficiency

[CANCER RESEARCH 48, 4529-4534, August 15. 1988| Effects of Folate Deficiency on the Metastatic Potential of Murine Melanoma Cells1'2

Richard F. Branda,2 John J. McCormack, Carol A. Perlmutter, Linda A. Mathews, and Steven H. Robison3

Departments of Medicine [R. F. B., C. A. P.], Pharmacology [J. J. M.], and Neurology [S. H. R.J. and the Vermont Regional Cancer Center [L. A. M.], University of Vermont, Burlington, Vermont 05405

ABSTRACT acid metabolism critically affects chromosome stability, and vitamin lack is associated with a variety of chromosomal ab Experiments were designed to measure the effect of folie acid defi normalities (6-9). Since tumor metastatic propensity appears ciency on a major determinant of cancer lethality, the propensity to form to correlate with chromosomal instability (10), we investigated mÃ©tastases.Murine B16 melanoma cells (FIO strain) were grown in folate-deficient and -supplemented media. After 3 days, cells in the the effect of cellular folate deficiency on this important deter deficient medium had restricted proliferative capacity, low folate levels minant of tumor lethality. by bioassay, increased cell volume, abnormal deoxyuridine suppression tests, accumulation of cells in S phase by flow cytometry, and increased numbers of DNA strand breaks. These folate-deficient cells consistently MATERIALS AND METHODS initiated more pulmonary mÃ©tastasesthancontrol cells when injected into Murine B16 melanoma cells (FIO strain) were obtained from the host mice. Cell size did not appear to be a major factor in pulmonary Tumor Repository of the Division of Cancer Treatment, National metastasis formation. In vitro growth rates and cloning efficiencies were Cancer Institute, Frederick Cancer Research Facility, Frederick, MD. comparable for cells in both types of medium as was subcutaneous growth The cells were grown in "folate-free" Dulbecco's modified Eagle's of tumors. We conclude that folate deficiency increases the metastatic medium (M. A. Bioproducts, Walkersville, MD) or the same medium potential of cultured melanoma cells. supplemented with 4 /ng/ml of folie acid (Sigma). To these media were added 10% dialyzed horse serum, 2% glutamine, and 1% penicillin- streptomycin (all obtained from Gibco Laboratories, Grand Island, INTRODUCTION NY). The final "low-folate" medium contained 0.26 ng/ml of folate Malnutrition is frequently observed and is often severe in activity by the Lactobacillus casei assay (11, 12). Cell suspensions in 15 ml of medium were placed in T75 flasks (Corning) and incubated at patients with malignant diseases (1). It is due to the combined 37Â°Cwith 5% COz. Cell counts were performed manually, and viability effects of decreased nutrient intake and increased utilization by was assessed by trypan blue exclusion. Cell cultures were divided by tumor metabolism. Losses of adipose tissue and muscle protein treating for 5 min with 2 ml of Versene (1:5000), washing with medium, are common, as is weight loss; the latter may adversely affect and inoculating a new flask with 1 x K)'1cells. In some experiments, survival and response to chemotherapy (1). Deficiencies of flasks were inoculated with 5 x IO5cells. micronutrients are also frequently encountered in cancer pa IMiraceliular folate levels were measured with a L. casei assay (13). tients. For example, some nutritional surveys have reported low Melanoma cells were treated with Versene and washed with cold saline. blood levels of folie acid in 60 to 85% of cases (2, 3). In the Then 1 x 10s cells were suspended in 2 ml of 0.1 M sodium acetate series reported by Magnus, 11 of 64 patients with metastatic buffer (pH 4.5) containing 0.1 % ascorbate and 0.2 ml of dialyzed horse cancer had serum levels of 0 to 0.9 ng/ml, and 19 others had serum. This suspending medium contains no measurable folate activity. Following sonication and incubation of the homogenate for 90 min at levels of 1 to 1.9 ng/ml; 53 of these 64 patients had serum 37Â°C,the protein was denatured by boiling for 5 min and removed by levels below the normal range (3.0 to 11.0 ng/ml) for his centrifugation. The supernatant was assayed as previously described laboratory (2). (11,12). Despite its frequency in patients with malignancies, the effect Deoxyuridine suppression tests were performed by modifications of of folie acid deficiency on the clinical course of neoplasia has our previously reported methods (12). After Versene treatment and not been studied in detail. Early reports by Farber and cowork- washing with medium, melanoma cells were resuspended in appropriate ers of an "acceleration phenomenon" in leukemic patients medium at a concentration of 1 x 10' cells/ml. One ml of suspension receiving folie acid were followed by animal studies suggesting was placed in 12- x 75-mm tubes, to which was added deoxyuridine that tumor growth is retarded in folate-deficient hosts (4, 5). (Sigma) (final concentration, 0.1 or 1 ^M) in phosphate-buffered saline or buffer alone. After l h at 37Â°C,0.1 /Â¿Ciof[3H]thymidine (Research Subsequently clinicians have been reluctant to administer folie acid to deficient cancer patients on the chance that the vitamin Products International Corp., Mt. Prospect, IL) was added, and the mixture was incubated for an additional 3 h. Radioactivity was then might enhance the growth of the malignancy. Since the publi determined in the trichloroacetic acid precipitates. Results are ex cation of these studies 20 or more years ago, a great deal has pressed as thymidine incorporation in samples with deoxyuridine di been learned about folate metabolism. Consequently, we sought vided by incorporation by control cells in deoxyuridine-free cultures x to investigate the influence of this common vitamin deficiency 100. All determinations were done in triplicate. on tumor biology in the light of more recent information. Folk Cell cycle distribution was determined by flow cytometry and ana lyzed by a parametric method (14). Melanoma cells (1 x IO4) were Received 1/11/88; revised 5/12/88; accepted 5/19/88. stained with propidium iodide (50 ^g/ml in 10% horse serum containing The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in 1 mg/ml of RNase and 0.025% Nonidet P-40). Nuclear fluorescence accordance with 18 U.S.C. Section 1734 solely to indicate this fact. intensity was determined with an Ortho Cytofluorograf 50 H/H using 1This work was supported by grants from the Vermont Division of the 488 mn argon ion laser excitation. American Cancer Society, the National Cancer Institute (CA 24543, CA 41843, and IS 10 RRO 1510-01), and the American Institute for Cancer Research. DNA strand breaks were detected by alkaline elution as described by Presented in part at the national meeting of the American Society of Clinical Kohn with several minor modifications (15, 16). Cells were seeded at 1 Investigation, May 1985 (31). x 106/mlin 10 ml of medium containing 0.02 jiCi/ml of [14C]thymidine 2 Recipient of a Research Career Development Award from the NIH (CA- for 24 h. The cells were deposited on a polycarbonate filter (2-^m pore 00946). To whom requests for reprints should be addressed, at Hematology/ size) and lysed with a buffer containing 2% sodium dodecyl sulfate-20 Oncology Unit, 32 N. Prospect Street, University of Vermont, Burlington, VT 05405. IHMEDTA (pH 10.1). The filters were washed with 5 ml of 20 mM 3Supported by a grant from the Sandoz Foundation of America. EDTA (pH 10.1). Elution buffer [25 ml of 20 mM EDTA (free acid)- 4529

2% tetrapropyl ammonium hydroxide (pH 12.15)] was applied to the filter and pumped through at a rate of 0.035 ml/min. Fractions of 3 ml were collected at 90-min intervals. Afterward, the filter was dissolved in 1 ml of protosol and prepared for liquid scintillation counting. Elution fractions and a wash of the pump lines were suspended as a gel using Aquasol-2 containing 0.7% glacial acetic acid and assayed for radioactivity. Cloning efficiency was measured by growing cells for 3 days in standard or low folate medium and then plating at a density of 100 or 200 cells/plate in 60-mm culture dishes containing standard culture medium. After 1 wk, the plates were stained with crystal violet and colonies were counted. Mean channel volume was measured on a Model ZBI Coulter Coun ter and Model 256 channelyzer (Coulter Electronics, Inc., Hialeah, FL). The mean channel volume was converted into mean volume by calibrating the machine with microspheres of 9.77, 14.75, and 20.01 ^m in diameter (Epics Division of Coulter Corp., Hialeah, FL). Ã›J 2 Chromosome counts were performed on cells which were incubated U with Colcemid (0.02 Â¿ig/ml)for 2 h prior to harvesting. Following Versene treatment, the suspended cells were treated with 0.075 M KCI for 15 min at 37Â°C,fixed with methanol:acetic acid, 3:1 by volume, for 10 min, spread on wet slides, and then air dried. Slides were stained with 2% Giemsa solution. Metastasis formation was determined by injecting murine melanoma cells into the tail veins of male C57BL/6 mice which were 6 to 8 wk old. The cell suspension contained 1 x 10' viable cells in 0.1 ml of medium. Cell viabilities were 95.7 Â±3.4% (mean Â±SD) for control cells and 95.4 Â±3.0% for low folate cells in 13 experiments.

DAY Fig. 1. Effect of incubation in folate-defÃ¯cient medium on melanoma cell RESULTS proliferation. Murine B16 (FIO strain) melanoma cells were cultured in standard, folate-containing medium (â€¢)or folate-deficient medium (O). After 3 days, cells Murine B16 melanoma cells were grown in folate-deficient from the folate-deficient medium were split and inoculated into either the medium Eagle's medium or the same medium supplemented with folie or into standard medium (A). Points, mean of at least 6 determinations; hnr\. SEM. acid. In both types of medium, there is a 1-day lag in growth, while the cells adhere to the flask, followed by a 2- or 3-day period of exponential growth (Fig. 1). If the cells remain con fluent they quickly lose viability over the next 1 or 2 days. If, however, the cells in standard medium are divided and cultured in fresh medium on Day 3, they will again grow exponentially after a 1-day lag. Cells grown in folate-deficient medium prolif erate at the same rate as control cells until confluence (Fig. 1, Days 1 to 3). If these cells from folate-deficient medium are divided into folate-supplemented medium after 3 days of cul Fig. 2. The distribution of cellular volume in murine melanoma cultures after various periods of incubation. The profile of cells cultured in folate-containing ture, they will proliferate at the same rate, or slightly faster, medium is shown with a continuous line. The volume distribution of melanoma than control cells (Fig. 1, A). On the other hand, if they are cells cultured in low folate medium is represented by the interrupted lines. Cell volume was derived from channel number by calibration with microspheres as split into fresh, low folate medium, they do not proliferate (Fig. described in the text. Each histogram consisted of approximately 50,000 cells. 1, O; Days 3 to 6). These results suggest that folate deficiency The results shown are for one representative experiment of 6 experiments. reversibly limits cell growth after approximately 3 days in low folate medium. The reversible nature of this inhibition is further To assess the folate status of these cells more directly, intra- substantiated by the observation that, in 10 experiments, the cellular folate activity was determined with a /,. casei assay, and cloning efficiencies of cells after 3 days in control or low folate deoxyuridine suppression tests were performed (Table 1). The medium were similar, 70.5 Â±7.9 and 58.4 Â±6.5 (mean Â±SE), latter test measures the availability of folate compounds for respectively. This difference was not statistically significant. conversion of deoxyuridine to thymidine. Homogenates of cells As shown in Fig. 2, mean cell size increased progressively grown in standard medium contained L. casei activity which when melanoma cells were grown in culture for 3 to 5 days, but fluctuated between 0.38 and 0.95 ng/106 cells. These values this increase was greatly magnified by growing the cells in represent the mean of 6 determinations. The levels of folate folate-deficient medium. The latter cells also manifested a activity appeared to vary with growth rate, being lowest during greater variation in cell size, indicated by the width of the the exponential phase. In contrast, melanoma cells grown in curves. After 3 days in culture, the differences were rather low folate medium had a drop in folate activity during the modest. In 6 experiments the mean cell volume Â±SDfor control culture period to a nadir of 0.28 ng/106 cells. Deoxyuridine cells was 1627 Â±96 fl, while for folate deficient cells the mean suppression tests were performed as paired observations in at was 2023 Â±206 fl. After 4 days of incubation, the differences least 6 experiments, using two different concentrations of deox were more substantial: 1825 Â±165 fl (control); 2539 Â±439 yuridine (0.1 and 1 pM). After 48 h of culture, deoxyuridine (low folate). Cells which survived for 5 days in folate-deficient suppression in cells from low-folate medium tended to be medium were very large: 1912 Â±238 fi (control) compared to reduced compared to control cells, but these differences were 2759 Â±799 fl (low folate). not statistically significant. However, after 72 h in folate- 4530

Table 1 Inlracellular folate activity of melanoma cells (FIO strain) during phase, and increased fragility of DNA compared to cells in incubation in folate-deficient or standard medium folate-supplemented medium. casei activity (ng/10* cells) in the following Murine melanoma cells were again grown for 3 days in mediaStandard Day of standard and low folate media and injected as a suspension (1 culture0 medium0.72 folatemedium0.46 x 10s viable cells in 0.1 ml of medium) into the tail veins of Â±0.11Â° normal C57BL/6 mice. Control (standard medium) cells were 1 0.95 Â±0.10 Â±0.07 2 0.38 Â±0.07 0.32 Â±0.06 always injected first (18). Pigmented pulmonary and hepatic 3 0.47 Â±0.07 0.28 Â±0.10 mÃ©tastaseswere scored 10 to 14 days later. Folate-deficient 4L. 0.72 Â±0.21Low 0.32 Â±0.06 cells initiated many more pulmonary (Table 3) and hepatic mÃ©tastases.For example, in Experiment 1, only a single hepatic Deoxyuridine suppression (% of control) metastatic nodule was found in 8 animals given injections of UM"NDC8.0 flMND23.2 MM9.0 llM19.3 control FIO cells, while 9 of 11 animals receiving folate-de 12 Â±2.3Â°15.7 Â±5.837.5 pleted cells had liver mÃ©tastases(P < 0.01). In 12 additional Â±2.6 Â±1.1 Â±2.2 Â±9.7 experiments performed over a 3-yr period, folate-deficient cells 31 8.7 Â±3.20.1 28.7 Â±7.81 33.7 Â±7.20.1 66.4 Â±17.6 â€¢MeanÂ±SEM. consistently initiated more pulmonary mÃ©tastasesthan mela * Deoxyuridine concentration. noma cells grown in folate-supplemented medium. In many c ND, not done. cases, the difference was striking and easily recognizable as soon as the chest was opened. The difference was statistically Table 2 Study of cell cycle distribution byflow cytometry of melanoma cells significant, usually highly so, when tested by both parametric (FIO strain) cultured in folate-deficient or standard medium and nonparametric methods (Table 3) in 10 of 13 experiments. (%)Go-Gi69.9 cycle distribution In three other experiments, the number of mÃ©tastasesiniti Day123CulturemediumStandard ated by folate-deficient cells was not statistically greater than Â±6.6Â° Â±4.4 Â±2.3 the number derived from control cells. The explanation for this LowfolateStandard 67.3 Â±7.962.4 16.0Â±3.219.3 16.7 Â±5.518.3 may be multifactorial. Clearly, the assay system is highly vari Â±4.6 Â±5.1 Â±5.1 able both within an experiment and from experiment to exper LowfolateStandard 62.7 Â±5.679.6 21.0Â±8.610.6 16.3 Â±5.09.7 iment, and this variability complicates statistical analysis. We found that using early passage cells helped reduce this variabil Â±1.6 Â±0.7 Â±1.7 Low folateCell 63.2 Â±9.0*S14.626.6 Â±6.9CGj15.510.2 Â±2.9 ity among experiments, because metastasis formation decreased Â°Mean Â±SD of 5 determinations. with longer passage in vitro (data not shown). Confirming the * P < 0.005. observations of others (19), we noted a direct relationship ' P< 0.001. between cell culture density and number of mÃ©tastasesformed, but the correlation coefficient for 8 observations was only 0.63. deficient medium, melanoma cells manifested abnormal deox- In addition, different lots of horse serum appeared to influence yuridine suppression tests at both concentrations of deoxyuri- the metastatic capacity of tumor cells. Finally, it was important dine (P < 0.05 by the paired t test) (17). Thus, by 72 h of culture to incubate the tumor cells for at least 3 days in folate-deficient in deficient medium, melanoma cells had sufficiently low intra- medium, since cells cultured for 2 days did not show enhanced cellular levels of folate activity to limit formation of thymidylate metastasis formation (92 versus 66, and 5.8 versus 5.6, control for DNA synthesis (Table 1). versus cells in low folate medium; these differences were not As an additional index of folate activity and its effect on statistically significant). We conclude that attention to culture DNA metabolism, cell cycle distribution in 5 separate cultures conditions can reduce but not eliminate the variability in assay was determined by flow cytometry (Table 2). Distribution was results. similar until 72 h of culture, when highly significant differences Cell size alone did not appear to have an important effect on were noted in the numbers of cells in G0-G| and S in low folate metastasis formation. As shown in Fig. 2, mean cell size in medium. Therefore, by 3 days of incubation, cells in low folate creased progressively during incubation in folate-deficient me medium have an accumulation of DNA synthesis in S phase dium. Nevertheless, there was not a coincident change in met compared to control cells. astatic capacity. Tumor cells from this experiment in folate- The effect of this folate deficiency and impaired DNA syn deficient medium formed 126 (85 to 180), 96 (47 to 300), and thesis on melanoma chromosomal stability was assessed by the 132 (68 to 243) pulmonary mÃ©tastases(median, range) after 3, alkaline elution test. In this assay, decreased retention of DNA 4, and 5 days in culture, respectively. Control cells in parallel on the filter indicates the presence of DNA strand breaks. After cultures produced 47 (28 to 108), 42 (16 to 61), and 39 (20 to 2 days of incubation, cells in low-folate and standard media had 60) pulmonary mÃ©tastases.Analysis of the correlation between similar elution rates (P > 0.05). However, by 3 days there was mean cell volume of folate-deficient cells and number of mÃ©tas a statistically significant decrease (P < 0.01) in retention of tases counted in 7 experiments produced a coefficient of only DNA from folate-depleted cells, and this effect became much 0.27. more pronounced by 4 days in low folate medium (Fig. 3). Subcutaneous tumor growth rates were measured with folate- Karyotypic analysis indicated that the FIO cell line was highly depleted and control cells. Following a 3-day incubation in aneuploid. The mean chromosome number in 30 metaphases either folate-deficient or standard medium, melanoma cells was 68.1, with a range of 42 to 78. Results in folate-deficient were injected s.c. into the flanks of C57BL/6 host mice. Tumor cells were nearly identical: mean, 68.0; range, 48 to 77. volume was determined by caliper measurements. As shown in Taken together, these studies indicate that, after 3 days of Fig. 4, growth rates were comparable for both types of cells, culture in folate-deficient medium, murine B16 melanoma cells confirming our in vitro observation that the 3-day incubation have a reduced proliferative capacity, increased cell volume, in folate-deficient medium did not permanently alter prolifera low intracellular folate levels, impaired utilization of folate tive characteristics. compounds for thymidine synthesis, accumulation of cells in S Finally, the ability of these deficient cells to recover and 4531

z o Fig. 3. DNA strand breaks detected by al kaline filter elution in melanoma cells (FIO strain). Cells were cultured in standard (â€¢)or folatc-deficient (O) medium for the indicated â€¢ WITH FOLATE number of days. Then cells were deposited on O WITHOUT FOLATE a filter and lysed, and the DNA eluted as described in the text. At the indicated number z of hours of elution. retention of radiolabeled Q DNA was determined. Points, mean of quad ruplicate determinations; bars, SEM. DAY 5

3 6 9 12

HOURS OF ELUTION

Table 3 Melastatic potential of B16 murine melanoma cells after incubation in resume normal DNA synthesis when placed in standard me standard or low folate medium dium was evaluated by flow cytometry. The panels on the left nodules/lungExperiment12345678910111213StandardMedian no. of of Fig. 5 show analyses of FIO strain cells grown for 3 days in folate P6P*11 standard medium, then split into fresh standard medium. DNA medium14(10-198)''23 medium388 distribution was determined 24 and 48 h later. The middle set (53-604)179(22-370)285<0.001<0.00213 of panels represents similar studies in cells grown in low folate (9-78)210(68-318)1 <0.001<0.00210 (82-624)7 <0.025<0.0513 medium for 3 days and then split into standard medium. By 48 (0-9)14 (0-35)35(4-125)23(11-180)85<0.05<0.0513 h these cells have achieved a normal DNA distribution. In NS'NS12 (0-54)17(2-53)25 contrast, cells grown for 5 days in low folate medium (right NSNS9 (5-134)54(3-401)10 (20-425)76(18-301)69 <0.01<0.058 panels) were unable to recover in standard medium. Cellular NSNS7 proliferation was impaired, and most cells were dead by 72 h (0-43)38 (23-84)238(113-292)126(85-180)143<0.001<0.028 (data not shown). (23-75)47(28-108)15(2-28)51 <0.001

200 40O 600 2OO 400 600 200 400 600

B16F10Line o "o

24 Mrs 24 Hrs Later I Lat

ys CD __i^^^^l& 200 400 600 200 400 6OO 200 400 600

48 Hrs 48 Hrs 48 Hrs Later i Later Later

^__LJ^â€”I ?00 400 600 200 400 600 200 400 600 8 10 12 14 16 18 X 22 Fluorescence Intensity (Channel Number) Days Fig. 5. Cell cycle distribution in cells cultured in folate-replete medium follow Fig. 4. Subcutaneous growth of B16 melanoma (FIO strain) cells following a ing incubation in either standard or folate-deficient medium. Melanoma cells 3-day incubation in either standard or folate-deficient medium. Points, mean of 9 (FIO strain) were cultured for 3 (left and middle) or 5 days (right) in either animals given injections of cells from either low folate (O) or control (â€¢)medium. standard medium (/*â€¢//(orfolate-deficient medium (middle and right). All cells Bars, SF.M. The differences between points on the two curves were not statistically were then transferred to standard (folate-replete) medium. Cell cycle distribution significant. was determined 24 and 48 h later. 4532

DISCUSSION (23). Since one of the DNA polymerases responsible for the repair of DNA damage is thought to be involved with DNA The studies reported here indicate that folate-deficient mel replication, it is likely that the repair process is also affected. anoma cells have an enhanced capacity to colonize and form Finally, DNA ligase, an enzyme common to the repair and mÃ©tastases.The mechanism by which folate deficiency increases replication processes, is inhibited by a rise of the intracellular metastatic capacity or, more specifically, organ colonization is pool of dATP as occurs in folate deficiency (22). These meta unclear at present. bolic alterations could account for the chromosome abnormal Our studies confirm the work of others showing that folate ities observed in patients with megaloblastic anemias. Chro deficiency restricts the proliferation of rapidly dividing cells mosome breaks, gaps, despiralization, and increased sister chro- (20). Initially, this restriction is reversible, but after 5 days in matid exchanges have been reported (6-8). That normal folie low folate medium, the melanoma cells are unable to regain acid metabolism is necessary for the integrity of chromosomes proliferative capacity with folate supplementation. Thus, severe is further suggested by reports that folate-deficient medium folate deficiency resembles treatment with antifols, such as enhances expression of fragile sites and that methotrexate methotrexate, and eventually leads to cell death. On the other causes chromosomal abnormalities (9, 25). Taken together, hand, there was only a slight tendency, which was not statisti these observations indicate that folate deficiency is associated cally significant, for folate replacement to accelerate the growth with chromosomal instability. of deficient tumor cells (Fig. 4). Therefore, it appears that Increased genetic instability contributes to the generation of replacing folate levels of deficient cells restores proliferative tumor cells with the capacity to metastasize, presumably by capacity but does not increase the rate of proliferation above fostering diversity. Only a few cells in a primary tumor have the usual doubling time of the tumor (13). Based upon this the combination of characteristics required to progress through observation, it seems unlikely that the increased metastatic the complex series of steps necessary to form a metastasis (26). potential of folate-deficient cells is due simply to accelerated These variants arise due to an acquired genetic variability or growth in a folate-replete host and earlier recognition of lung instability in tumor cells which allows diversity to develop with nodules. time (27). In vivo host selection pressures then determine which Metastasis formation is a multistep process, and folate defi variants survive and grow. Recent studies suggest that cell lines ciency could conceivably alter one or more of these steps. Folie with greater metastatic ability have a higher rate of generation acid is necessary for 1-carbon transfers and participates in of metastatic variants (10). Metastatic potential can be further amino acid, purine, and pyrimidine synthesis. Deficiency of the enhanced by exposure to mutagens such as ultraviolet radiation vitamin produces large cells with delayed nuclear maturation (28). Pretreatment of murine melanoma cells with chemother- and an imbalance between DNA and RNA synthesis. These apeutic agents has been associated with increased metastatic megaloblastic cells might be more likely to arrest in the pul capacity. Hydroxyurea, l-/3-D-arabinofuranosylcytosine, 5-aza- monary capillary bed, but this mechanical alteration would not cytidine, and methotrexate all produced significant increases in explain the increased colonization of other organs by folate- the number of lung mÃ©tastases(21, 29). The latter drug is deficient cells. We did not find that metastatic capacity in particularly relevant to our studies, because of the well-known creased substantially when cells were grown for as long as 5 effects of methotrexate on folate metabolism. We propose that days in folate-deficient medium, despite the fact that these cells folate deficiency, by increasing chromosomal instability, might were much larger, on average, and with a greater variability in similarly promote tumor heterogeneity and increase the likeli size, than folate-replete cells. Moreover, the correlation between hood of developing cells with the capacity to form mÃ©tastases. mean cell volume and metastasis formation was weak when We did not find that major karyotypic abnormalities were analyzed in 7 experiments. These observations are consistent induced by folate deficiency in this highly aneuploid cell line. with studies in other laboratories of hydroxyurea-treated tumor This observation is consistent with studies in other laboratories cells. Pretreatment with hydroxyurea of 3 murine tumor cell which found no apparent relationship between generation of lines (B16-F1, B16-F10, and K-1735-clone 19) brought about metacentric marker chromosomal changes and metastatic po a marked increase in the median number of lung tumor nodules. tential. These authors proposed that more subtle nonrandom Treatment with 0.1 mM hydroxyurea was associated with little genetic or molecular changes may be determining factors for change in cell volume but significantly increased metastatic malignant potential (30). Our observation that folate-deficient propensity. In other experiments, hydroxyurea-treated cells melanoma cells have more DNA strand breaks than folate- were larger than controls, but they were not trapped in the replete cells (Fig. 3) may implicate these chromosomal changes lungs to a greater extent than control cells. These authors in the process which eventuates in metastatic variants. These concluded that neither cell size nor the position of the majority associations do not prove a causal relationship, but they provide of the cells in the cell cycle was a major factor in pulmonary a hypothesis that can be tested further. metastasis formation (21). It seems more likely that the increased metastatic potential REFERENCES of folate-deficient cells is related to the effects of folate com 1. Kisner, D. I... and DeWys, W. D. Anorexia and cachexia in malignant pounds on nucleic acid metabolism and chromosome stability. disease. In: G. R. Newell and N. M. Ellison (eds.). Nutrition and Cancer: Decreased intracellular folate activity results in profound alter Etiology and Treatment, pp. 355-365. New York: Raven Press, 1981. ations in nucleotide pools (22, 23). Under these conditions, 2. Magnus, E. M. Folate activity in serum and red cells of patients with cancer. Cancer Res., 27:490-497, 1967. there is a reduction in the rate of DNA replication fork move 3. Somayaji, B. N., Nelson, R. S., and McGregor, R. F. Small intestinal function ment and inhibition of the gap-filling step and/or joining of in malignant neoplasia. Cancer (Phila.), 29: 1215-1222, 1972. Okazaki pieces, resulting in DNA with persistent single- 4. Potter, M.. and Briggs, G. M. Inhibition of growth of amethopterin-sensitive and amethopterin-resistant pairs of lymphocytic neoplasms by dietary folie stranded regions (24). In addition, dUMP tends to be incorpo acid deficiency in mice. J. Nati. Cancer Inst., 28: 341-351, 1962. rated during gap repair, leading to its cyclic incorporation and 5. Farber, S., Diamond, L. K.. Mercer, R. D., Sylvester, R. F., and Wolff, J. A. Temporary' remissions in acute leukemia in children produced by folie acid removal. antagonist. 4-aminopteroyl-glutamic acid (aminopterin). N. Engl. J. Mod.. This process may result in extensive degradation of DNA 238: 787-793, 1948. 4533

6. Menzies, R. C, Crossen, P. E., Fitzgerald, P. H.. and Gunz, R. W. Cytoge- intravenous injection of CT 26 colon carcinoma, Lewis lung carcinoma, or nelic and cytochemical studies on marrow cells in B,2 and folate deficiency. B16 amelanotic melanoma cells. Cancer Res., 44: 3880-3883, 1984. Blood, 2Â«:581-594, 1966. 19. Honn, K. V., Steinert, B. W., Pilarski, S. M., Nelson, K. K., Sloane, B. F., 7. Heath, C. W. Cytogenic observations in vitamin B|2 and folate deficiency. and Onoda, J. M. Effect of culture conditions on metastatic potential of Blood. 27:800-815, 1966. BI6a tumor cells. Proc. Am. Assoc. Cancer Res., 27: 30, 1986. 8. Knuutila, S., Helminen, E., Vuopio, P., and de la Chapelle, A. Increased 20. Rosen. R., and Nichol. C. A. Inhibition of the growth of an amethopterin- sister chromatid exchange in megaloblastic anaemia-studies on bone marrow refractory tumor by dietary restriction of folie acid. Cancer Res., 22: 495- cells and lymphocytes. Hereditas, 89: 175-181, 1978. 500. 1962. 9. Sutherland, G. R. Heritable fragile sites on human chromosomes. I. Factors 21. McMillan, T. J., Rao, J., and Hart, I. R. Enhancement of experimental afTccting expression in lymphocyte culture. Am. J. Hum. Genet., 31: 125- metastasis by pretreatment of tumor cells with hydroxyurea. Int. J. Cancer, 135, 1979. JÂ«:61-65, 1986. 10. Hill, R. P., Chambers, A. F., Ling, V., and Harris, J. F. Dynamic heteroge 22. Hoffbrand, A. V., and Tripp. E. Unbalanced deoxyribonucleotide synthesis neity: rapid generation of metastatic variants in mouse B16 melanoma cells. caused by methotrexate. Br. Med. J., 2: 140-142, 1972. Science (Wash. DC), 224: 998-1001, 1984. 23. Goulain, M., Bleile, B., and Tseng, B. Y. Methotrexate-induced misincor- 11. Waters, A. H., and Mollin, D. L. Studies on the folie acid activity of human poration of uracil into DNA. Proc. Nati. Acad. Sci. USA, 77: 1956-1960, serum. J. Clin. Pathol.. 14: 335-344, 1961. 1980. 12. Branda, R. F.. Moldow, C. F., MacArthur, J. R., Wintrobe. M. M., Anthony, 24. Wichremasinghe, R. G.. and Hoffbrand, A. V. Reduced rate of DNA repli B. K., and Jacob. H. S. Folate-induced remission in aplastic anemia with cation fork movement in megaloblastic anemia. J. Clin. Invest., 65: 26-36, familial defect of cellular folate uptake. N. Engl. J. Med., 298: 469-475, 1980. 1978. 25. Jensen, M. K., and Nyfors, A. Cytogenetic effect of methotrexate on human 13. Steinberg, S. E., Fonda, S., Campbell, C. L.. and Hillman. R. S. Folate cells in vivo. MutÃ¢t.Res., 64: 339-343, 1979. utilization in Friend erythroleukemia cells. J. Cell. Physiol., 114: 252-256, 26. Fidler, I. J. Tumor heterogeneity and the biology of cancer invasion and 1983. metastasis. Cancer Res., 38: 2651-2660, 1978. 14. Dean, P. N., Gray, J. W., and Dolbeare, F. A. The analysis and interpretation 27. Nowell, P. C. The clonal evolution of tumor cell populations. Science (Wash. of DNA distributions measured by flow cytometry. Cytomelry, 3: 188-195, DC), / 94: 23-28, 1976. 1982. 28. Fisher, M. S., and Cifone, M. A. Enhanced metastatic potential of murine 15. Kohn. K. W.. Ewig, R. A. G., Erickson, L. C, and Zwelling, L. A. DNA fibrosarcoma treated in vitro with ultraviolet radiation. Cancer Res., 41: Repair: A Laboratory Manual of Research Procedures, pp. 279-402. New 3018-3023,1981. York: Marcel Dekker, 1981. 29. McMillan, T. J., and Hart, I. R. Enhanced experimental metastatic capacity 16. Robison, S. H.. and Bradley, W. G. Impaired DNA repair replication in of a murine melanoma following pre-treatment with anticancer drugs. Clin. Alzheimer's disease cells. In: i. T. Hutton and A. D. Kenney (eds.). Senile Exp. Metastasis, 4: 285-292, 1986. Dementia of the Alzheimer Type. pp. 205-218. New York: Alan R. Liss, 30. Kendal. W. S., Wang, R-Y., Hsu, T. C., and Frost, P. Rate of generation of 1985. major karyotypic abnormalities in relationship to the metastatic potential of 17. Armitage, P. Statistical Methods in Medical Research, pp. 116-118. New B16 murine melanoma. Cancer Res., 47: 3835-3841, 1987. York: John Wiley & Sons, 1971. 31. Branda. R. F.. McCormack, J. J., Perlmutter, C. A., and Matnews, L. A. 18. Karpatkin. S., Ambrozio, C., and Pearlstein, S. Lack of effect of in vivo Folate deficiency increases the metastatic potential of melanoma. Clin. Res., prostacyclin on the development of pulmonary mÃ©tastasesinmice following Â¿J.-576A,1985.

4534

Richard F. Branda, John J. McCormack, Carol A. Perlmutter, et al.

Cancer Res 1988;48:4529-4534.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/48/16/4529

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/48/16/4529. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.