Proc. Nat. Acad. Sci. USA Vol. 72, No. 12, pp. 4994-4998, December 1975 Cell Biology
Selective killing of transformed baby hamster kidney (BHK) cells (anti-cancer agents/cell cycle/virus transformation)
ARTHUR B. PARDEE* AND LYNNE J. JAMES Department of Biochemical Sciences, Princeton University, Princeton, New Jersey 08540
Contributed by Arthur B. Pardee, September 22, 1975
ABSTRACT We report here that certain drugs can protect Syrian baby hamster kidney cells (BHK) in culture against the lethal agents cytosine arabinonucleoside, hydroxyurea, and colcemid. Polyoma virus-transformed BHK cells (PyBHK) are killed under the same conditions. The protective drugs include caffeine and streptovitacin A. Kinetic studies show that these drugs act specifically in GI, and that they shift BHK cells from GI into the Go state at the restriction point, similar to the effects of high cell density or serum deprivation. These drugs do not block the growth of PyBHK cels nearly as effectively, consistent with a reduced effectiveness of restriction point control in virus-transformed cells. Consequently, the transformed cells continue around
their cycle and
killed by the cell cycle phase-specific the arrested BHK cells. These findings provide a model for studies on differential killing of tumor versus normal cells in vivo. in
Successful cancer chemotherapy will depend upon finding drugs which, alone or in combination, kill cancer cell populations effectively, but which do not irreversibly harm normal cells. A large differential effect against cancer cells relative to all essential normal cells in the body is required. In this paper we report experiments which show, with a model system in culture, that the differential killing by several anti-cancer drugs can be greatly increased by pretreatment of the cells with certain nonlethal drugs. This report is based on the earlier finding that normal cells in culture can be shifted between proliferation and quiescence (1), depending on whether or not the cells can accomplish some specific event in G1. We call this phase in the cell cycle the restriction point (2). A weaker control at the restriction point could be responsible for the continued proliferation of tumor cells. If a specific event restricts normal cell proliferation, then drugs should exist which specifically inhibit this event. Indeed, there are reports in the literature that cells of normal lines are arrested in the G1 part of the cycle by caffeine (3) or puromycin aminonucleoside (4). We report here that these drugs and others arrest the cells in a very similar way to serum deprivation, nutritional insufficiency, cell crowding, or high 3':5'-cyclic AMP. The drugs arrest baby hamster kidney (BHK) cells in their cycle, but do not similarly stop growth of polyoma-transformed BHK (PyBHK) cells. This difference provides the basis for selective protection of the normal cells against lethal anti-cancer agents that cycle phase-specific.
(5). New cultures were grown from frozen stocks at monthly intervals. The cells were maintained in Dulbecco's modification of Eagle's medium with 10% calf serum, at 370 under 10% CO2 pressure. Experiments on thymidine (dThd) incorporation were performed in Linbro FB-16-24-TC trays. Approximately 3 X 103 cells in 1 ml of medium were placed into each well of a Linbro tray and were allowed to grow into a loose network during 2-3 days. The cells remained attached to the bottoms of the wells throughout the experiments. Media were changed by aspiration and rinsing with phosphate-buffered saline (pH 7.2) or medium. The cells were allowed to incorporate [3H]dThd, added to 0.05 1AM and 0.5 MiCi/ml for the desired time interval. Then they were rinsed with 2 ml of phosphate-buffered saline, and left for 1--8 hr with 1 ml of 5% trichloroacetic acid. The cells were dissolved in 0.8 ml of 2% Na2CO3 in 0.1 M NaOH. A 0.6 ml aliquot of this solution was added to 10 ml of Triton X-100-toluene-Spectrofluor scintillator and neutralized with 0.2 ml of 50% trichloroacetic acid. The radioactivity of samples was measured in an Intertechnique scintillation counter. Autoradiography was performed with cells growing on round 15 mm glass cover slips. Carrier-free tritiated dThd was present during the incubations at 0.5 MCi/ml. Kodak NTB-2 nuclear track emulsion was used according to standard methods. Viability was determined by plating the cells on petri dishes and counting the colonies after 6-7 days. Streptovitacin A was a gift from Dr. B. K. Bhuyan of the Upjohn Co., Kalamazoo, Mich. RESULTS Inhibitors that stop cells reversibly in G1 A survey was made for compounds that reversibly arrest BHK cells in G1, in a manner similar to deprivation of serum or essential amino acids. Exponentially growing BHK cells were exposed to low concentrations of various compounds, and incorporation of [3H]dThd into trichloroacetic-acid-insoluble material was measured during the first and then the Table 1.
[3H]dThd incorporation after inhibition of exponential cells
[3H]dThd as % of control
MATERIALS AND METHODS Syrian hamster BHK 21/C13 and J1 (PyBHK) cells were obtained from the Imperial Cancer Research Fund stocks Abbreviations: BHK cells, baby hamster kidney cells; PyBHK, BHK cells transformed by polyoma virus; dThd, thymidine; PAN, puromycin aminonucleoside; SVA, streptovitacin A. * Present address: Sidney Farber Cancer Center, 35 Binney Street, Boston, Massachusetts 02115.
Control SVA Caffeine PAN 5-F-uracil*
0.075 450.0 12.0 250.0
Also inhibits de novo dTMP synthesis.
63 58 83 125
16 13 3 38
Cell Biology: Pardee and James 24
Caffeine o a SVA 24hr
Proc. Nat. Acad. Scd. USA 72 (1975)
10 20 Time in Hours
FIG. 1. Recovery of BHK cells from inhibition by SVA or caffeine. To BHK cells growing in Linbro wells in complete medium plus 0.1 pM thymidine for 3 days was added 0.075 pg of SVA/ml or 600 ug of caffeine/ml, for either 15 or 24 hr. Then the cells were rinsed and put into warm fresh medium (4% calf serum) plus 0.025 mM adenosine and [3H]thymidine (at zero time on the figure). Incorporation was measured at intervals thereafter.
second generation times. Compounds that permitted a large dThd incorporation initially but were more strongly inhibitory in the second cycle were examined further (Table 1). Streptovitacin A (SVA) and caffeine were reversible inhibitors of dThd uptake and did not diminish viability, making these compounds particularly suitable for further study. The inhibition appeared to be rapid when cells were in the G1 phase of the cycle (see below). By contrast, puromycin aminonucleoside (PAN) or 5-F-uracil were potent inhibitors during G1, but were not as readily reversible; the inhibited cells did not completely recover their ability to take up thymidine after 24-hr exposure. Position of the inhibition points of SVA and caffeine Several types of experiments were performed to localize the time in the cell cycle at which SVA or caffeine inhibited cell growth. An inhibitor was added to growing BHK cells for 15 or 24 hr, and then was replaced with complete medium. Most of the cells commenced dThd incorporation after approximately 10 hr (Fig. 1). Various experiments showed that cell division did not commence until later, some time between the 20th and 30th hr after release. These results are very similar to the kinetics of release from Go following serum starvation (Fig. 2, control). The synchrony seems. to be as good as that obtained by more prolonged deprivations of isoleucine or serum. However, it should be noted that more than 24-hr exposures to these drugs appeared to be more toxic than serum starvation, and increased the time of recovery to more than 10 hr. We conclude that the drugs appear to put cells into much the same state as the restriction point conditions studied previously (2). A second type of experiment was designed to determine whether the drugs could prevent escape of cells from Go. The cells were arrested in Go by serum starvation for 2 days. They were then released into complete medium with or without added drugs, for 9 hr. To see if the drugs had prevented progress from Go towards S, the cells were then put into complete medium plus [3H]dThd, and incorporation was measured at intervals. Drugs which are known to have no effect between Go and S give a curve like that of the control with complete medium (Fig. 2). Hydroxyurea is a drug of this sort; it is known to inhibit DNA synthesis, but not progress between Go and S.
Time in Hours
FIG. 2. Inhibition of recovery byBHK cells from serum starvation. BHK cells growing in Linbro wells for 3 days were put into medium with 0.5% serum + 0.1 mM hypoxanthine + 0.1 mM ornithine + 0.1 pM thymidine, and starved for 48 hr. They were then (at zero time) put into the following media: fresh medium (4% calf serum + 0.025 mM adenosine); fresh medium + 1 mM hydroxyurea; fresh medium + 600 Mg/ml of caffeine; fresh medium + 0.075 pg/ml of SVA; fresh medium + 100 ug/ml of 5-F-uracil. Then at 9 hr all cells were transferred to fresh medium (10% calf serum + adenosine + 1 uM thymidine and [3Hlthymidine). Samples were taken at intervals for measurement of 3H incorporation.
Complete inhibition of progress towards S is shown by the for cells left with low serum medium during the 9-hr interval; that is, the onset of DNA synthesis is delayed by 9-10 hr after the addition of serum. The drugs caffeine, PAN, and SVA all delayed progress towards S to an extent similar to that of low serum. However, there were small differences; the first two drugs permitted slightly earlier initiation than did low serum, whereas with SVA, dThd incorporation was slightly lower. 5-F-uracil inhibited dThd incorporation irreversibly. We conclude that, in contrast to the other drugs, SVA, caffeine, or PAN block at the same point as low serum or slightly later. A third type of experiment was designed to determine whether there is a time between GO and S beyond which the inhibitors become ineffective. Cells were released from serum starvation into complete medium plus [3H]dThd. Control cells were sampled at intervals to determine the time at which DNA synthesis commenced (Fig. 3). To other sets of cells, inhibitors were added at 2-hr intervals. The ability of these cells to incorporate [3H]dThd was measured using autoradiography, during more than one cycle (36 hr) after release from the low serum inhibition. Three of the drugs as well as low serum were effective inhibitors for approximately 5 hr following release from Go. Thereafter, more than half of the cells were capable of taking up dThd and so had passed the point of sensitivity to these drugs. This point for the average cell is approximately 6 hr prior to the initiation of DNA synthesis. With SVA the average cell became insensitive approximately 2 hr later than with the other drugs. Possibly this is because SVA is itself somewhat toxic and partly inhibits the protein synthesis required before and during DNA synthesis, or perhaps SVA is more immediately inhibitory than are the other compounds, which might require a delay before their inhibition is evident. In any event, all the drugs and also low serum (1) acted more strongly if they were added early in the Go to S period, as contrasted to addition later in this period. We conclude that these drugs are early G1-selective inhibitors. curve
Proc. Nat. Acad. Sci. USA 72 (1975)
Cell Biology: Pardee and James
ss/ 8 Time in
- Uptake dMhd //0.2%
FIG. 3. Determination of time of inhibition during the Go/S interval. BHK cells were grown for 3 days on glass cover slips, and then were put into 0.5% serum medium (+0.1 mM hypoxanthine and 0.1 mM ornithine) for 48 hr. They were then put into fresh medium (4% calf serum + 0.25 mM adenosine) + 0.5 MCi/ml carrier-free [3H]thymidine at zero time. One series (dThd uptake) was fixed at times shown on the abscissa, to determine the fraction of cells with labeled nuclei, by autoradiography. Five other sets were exposed at intervals, starting at times shown on the abscissa to medium without calf serum CS- (+ hypoxanthine + ornithine); 600 ,gg/ml of caffeine; 25,gg/ml of 5-F-uracil; 6,gg/ml of puromycin aminonucleoside; 0.75 gg/ml of SVA. After 36 hr all of these slips were fixed and the % labeled nuclei was determined. Symbols are as for Fig. 2.
We have previously suggested (2) that BHK cells transformed with polyoma virus have lost their restriction point control. Such cells are not arrested in Go by serum starvation or amino acid deprivation. Thus, experiments were done to find out whether GI-selective inhibitors can block the growth of transformed cells. Growing cultures of BHK and PyBHK cells were exposed to these drugs for several days (Fig. 4). The BHK cells took up very little [3H]dThd. In contrast, the transformed cells continued to take up dThd rapidly during the entire experiment. Thus there is a dramatic difference in the response of the untransformed and transformed cells to the Gl-selective inhibitors. Another comparison was made by placing both BHK and PyBHK cells into media with drugs or low serum for 24 hr and then measuring [3H]dThd incorporation after the cells were put back into complete medium. Whereas the BHK cells showed a lag of about 10 hr before they commenced dThd incorporation (as in Fig. 1), the PyBHK cells immediately and rapidly started to take up dThd, indicating that if they were inhibited they were at least not in Go. Could the different growth responses of BHK and PyBHK reflect different degrees of inhibition of macromolecular synthesis by these drugs? SVA, for instance, is an inTable 2. Protein and DNA synthesis inhibition by various drugs
[35S] Methionine (6 hr)
[3H] Thymidine (48 hr)
None SVA SVA Caffeine Caffeine PAN 5-F-uracil
100 35 24 76 70 120 110
0.045 0.075 450.0 600.0 6.0 25.0
100 51 37 24 16 67 63
100 91 91 87 80
30 22 48 39 74 78
100 50 Time in Hours FIG. 4. Inhibition of BHK versus PyBHK cells. BHK and PyBHK cells were grown in Linbro wells for 2 days, to a loose network. The following inhibitors were added to four wells each: 0.07$ ,gg/ml of SVA; 600 gg/ml of caffeine; 6 gg/ml of puromycin amiinonucleoside; medium with 0.2% serum (CS) (+ hypoxanthine 0
ornithine). Every 24 hr 0.5,gCi/ml of [3H]thymidine was added to of wells and 24 hr later these cells were fixed with 5% trichloroacetic acid and counted for uptake of 3H. The cumulative incorporation is plotted versus time of sampling.
hibitor of protein synthesis (6); its relative ineffectiveness in inhibiting PyBHK cell growth could be due to its decreased ability to inhibit protein synthesis by these cells. But no difference in ability of SVA to inhibit [a5S]methionine incorporation was seen for the BHK and PyBHK cells. In contrast, after one cycle, dThd incorporation was more strongly inhibited with BHK cells (Table 2). The other drugs inhibited methionine incorporation into BHK cells less than into PyBHK cells. Thus, a selective inhibition of protein synthesis. does not seem to be responsible for the different sensitivity of dThd incorporation into the two lines. Protection of normal cells but not transformed cells by SVA or caffeine According to the experiment described above, after BHK cells are arrested in Go by SVA or caffeine, they should not be sensitive to drugs that are active only during the S phase of the cell cycle. Such drug-arrested cells were exposed to the toxic compounds hydroxyurea or cytosine arabinonucleoside, colcemid or carrier-free [3H]dThd (suicide decay). Their ability to take up [3H]dThd was little different from the activity of control cells that had not been exposed to the toxic agent (Fig. 5A and B). In contrast, the unprotected cuiltures showed a large decrease in their ability to take up thymidine after exposure to the toxic agent. In similar experiments, survival was determined as measured by plating (Table 3). Following exposure to the inhibitory agent (SVA or caffeine), the killing of BHK cells by toxic agents was greatly reduced. With polyoma-transformed BHK cells 90% of the cells were killed, even after. one of the protective agents was added (Table 3). The inability of SVA or caffeine to inhibit the growth of PyBHK cells is presumably the reason why cytosine arabinonucleoside or hydroxyurea kills these cells after pretreatment, in contrast to protection of BHK cells.
Cell Biology: Pardee and James
Proc. Nat. Acad. Sci. USA 72 (1975)
50 0 25 50 Time in Hours FIG. 5. Protection by SVA or caffeine against cytosine arabinoside or hydroxyurea. (A) To BHK cells growing in Linbro wells for 3 days were added: no drug; 0.075 gg/ml of SVA; or 3 mM caffeine for 24 hr. Then to half of each set was added 10 Ag/ml of cytosine arabinonucleoside (Ara C); the other half of each set was an unpoisoned control. Twenty-four hours later, the cells were rinsed and put into medium + 4% calf serum and adenosine + [3H]thymidine. Samples were taken at intervals for measurement of trichloroacetic-acid-insoluble radioactive material. (B) The same experiment was performed except that 1 mM hydroxyurea (OHU) was substituted for cytosine arabinonucleoside.
zinski (9), PAN inhibits the appearance of cytoplasmic poly(A)-containing RNA in WI-38 cells, but not in simianvirus-40-transformed WI-38 cells. Caffeine could more specifically inhibit a process that produces a necessary product of gene action, or it could act via the cyclic nucleotides. We speculate that escape from the restriction point requires production of an activating substance that is under a control mechanism sensitive to a variety of conditions necessary for optimal growth. Our results do not favor a specific role of an individual amino acid or other metabolite as the controlling factor (10), a possibility that also seems unlikely owing to the variety of conditions that arrest cell proliferation at the re-
In this paper we demonstrate that metabolic inhibitorsstreptovitacin A, caffeine, 5-F-uracil, and puromycin ami-
nonucleoside-are capable of selectively blocking progress of BHK cells through the G1 phase of the cell cycle. The block appears to be at the restriction point where cells are stopped by a variety of conditions such as serum deprivation, amino-acid deprivation, high cyclic AMP, or high cell density (2). Synchronous cultures can be obtained by exposure to either SVA or caffeine for approximately one cell cycle interval, less than 24 hr. The synchrony obtained seems as good as that obtained by more prolonged deprivations of isoleucine or serum. The results support the idea that some specially sensitive metabolic event is required for progress through this stage of the cell cycle. The mechanism by which these diverse compounds (as well as various physiological conditions) arrest BHK and other normal cells at the restriction point is not clear and requires investigation. Since partial inhibition of protein synthesis by SVA stops BHK cells, control could depend on partial inhibition of protein synthesis. The mechanism could be similar to the stringent control of RNA synthesis when protein synthesis is inhibited in Escherichia coli (7). In animal cells the consequence of inhibition would be to shift the cells from proliferation to quiescence, rather than inhibition of RNA synthesis as in E. coli. It is quite possible, as suggested earlier by Kram et al. (8), that a pleiotypic regulator of cell growth could be produced. According to Cholon and Stud-
The Gi-selective inhibitory agents have quite different effects on BHK cells and transformed cells obtained after polyoma virus infection: the latter are far less inhibited than the former. This result is similar to the difference observed during serum deprivation of the two kinds of cells. This result was obtained earlier with PAN by Studzinski and Gierthy (4); they compared a "normal" mouse line (L cells) and a human tumor line (HeLa). Domon and Rauth (11) reported potentiation by caffeine of UV damage to L cells; in similar experiments with HeLa cells caffeine had less effect (12). Possibly the different susceptibilities of normal and transformed cells to these GI-selective drugs could be used to characterize transformation or even malignancy, much as the requirement for serum is utilized. Pre-treatment of BHK cells with SVA or caffeine for one generation (to put them into the Go state) protects the cells
Table 3. Preferential killing of PyBHK cells by drug combinations BHK cells Protective agent
2,000 10,000 4,000 2,000 8,000 7,000
None SVA Caffeine
0.075 mg/ml 600 Ag/ml
36,000 10,000 4,000
None SVA Caffeine
0.075,g/ml 600 ug/ml
24,000 12,000 4,000
The numbers are number of clone-forming cells per well.
PyBHK cells % Killed
Cytosine arabinonucleoside 94 140,000 0 12,000 0 8,000 Hydroxyurea 92 130,000 33 (0)
99 90 94
800 3,000 700
99 88 88
Cell Biology: Pardee and James
against S-phase or M-phase specific agents, including hydroxyurea, cytosine arabinonucleoside, [3H]dThd decay, or colcemid. Protection occurs because the arrested cells do not enter the S or M period where they are sensitive to the lethal drugs. Bhuyan and Fraser (13) have reported that 0.2 ,ug/ml of SVA protects DON and L1210 cells against S-phase specific agents. Their unsynchronized cultures had lost 70% viability because 70% of the cells were in the S-phase at the time the toxic agent was added. As this paper shows, viability is better preserved if the cells are pre-treated with the protective drug, thereby allowing them to pass out of the S-phase into Go before the toxic drug is added. In addition, SVA probably protects S-phase cells by slowing DNA synthesis [see Liebermanetal. (14)]. Particularly interesting is the protective action of SVA and caffeine on BHK cells as compared with polyoma-transformed BHK cells. The superior protection of the "normal" cells suggests the possibility that pre-treatment with a compound such as SVA or caffeine might protect normal cells preferentially in vivo, and thereby allow a greater differential action of toxic chemotherapeutic drugs on tumor cells in the intact animal. In this regard Lieberman et al. (14) have demonstrated protection of intestinal crypt cells by treatment with protein synthesis inhibitors, including cycloheximide (an analogue of SVA), before adding several S-phase specific agents. Although we have chosen to work with PyBHK cells which represent an extreme case of loss of growth control (since DNA-virus-transformed lines vary greatly in their characteristics and the most highly modified lines are selected), the results may be relevant to growth control of cancer cells in vivo. However, we expect growth control to be differently modified for various tumors, since tumors grow at a wide range of rates. Only a very modest release from growth inhibition could result in the unchecked growth that characterizes tumors. Therefore, we expect only partial modification of the restriction point control for many tumors. This paper is dedicated to the memory of Gordon M. Tomkins. This investigation was supported by Public Health Service Research Grant no. CA-11595 from The National Cancer Institute.
Proc. Nat. Acad. Sci. USA 72 (1975) 1. Temin, H. M. (1971) "Stimulation by serum of multiplication of stationary chick cells," J. Cell. Physiol., 78, 161-170. 2. Pardee, A. B. (1974) "A restriction point for control of normal animal cell proliferation," Proc. Nat. Acad. Sci. USA 71, 1286-1290. 3. Walters, R. A. Gurley, L. R. & Tobey, R. A. (1974) "Effects of caffeine on radiation-induced phenomena associated with cell-cycle traverse of mammalian cells," Biophysical J. 14, 99-118. 4. Studzinski, G. P. & Gierthy, J. F. (1973) "Selective inhibition of cell cycle of cultured human diploid fibroblasts by aminonucleoside of puromycin," J. Cell. Physiol. 81, 71-84. 5. Clark, G. D. & Smith, C. (1973) "The response of normal and polyoma virus-transformed BHK/21 cells to exogenous purines," J. Cell. Physiol. 81, 125-132. 6. Felicetti, L., Colombo, B. & Baglioni, C. (1966) "Inhibition of protein synthesis in reticulocytes by antibiotics. II. The site of action of cycloheximide, streptovitacin A and pactamycin," Biochim. Biophys. Acta, 119, 120-129. 7. Gallant, J., Ehrlich, H., Hall, B. & Laffler, T. (1970) "Analysis of the RC function," Cold Spring Harbor Symp. Quant. Biol. 35,398-405. 8. Kram, R., Mamont, P. & Tomkins, G. M. (1973) "Pleiotypic control by adenosine 3':5'-cyclic monophosphate: A model for growth control in animal cells," Proc. Nat. Acad. Sci. USA, 70, 1432-1436. 9. Cholon, J. J. & Studzinski, G. P. (1974) "Metabolic differences between normal and neoplastic cells: Effects of aminonucleoside on cytoplasmic messenger RNA," Science 184, 260-261. 10. Holley, R. W. & Kiernan, J. W. (1974) "Control of the initiation of DNA synthesis in 3T3 cells: Low molecular weight nutrients," Proc. Nat. Acad. Sci. USA, 71,2942-2943. 11. Domon, M. & Rauth, A. M. (1969) "Effects of caffeine on ultraviolet-irradiated mouse L cells" Radiat. Res. 39, 207-221. 12. Wilkinson, R., Kiefer, J. & Nias, A. H. W. (1970) "Effects of post-treatment with caffeine on the sensitivity to ultraviolet light irradiation of two lines of HeLa cells," Mutat. Res. 10, 67-72. 13. Bhuyan, B. K. & Fraser, T. J. (1974) "Antagonism between DNA synthesis inhibitors and protein synthesis inhibitors in mammalian cell cultures," Cancer Res. 34, 778-782. 14. Lieberman, M. W., Verbin, R. S., Landay, M., Liang, H., Farber, E., Lee, T. & Starr, R. (1970) "A probable role for protein synthesis in intestinal epithelial cell damage induced in viwo by cytosine arabinoside, nitrogen mustard, or x-irradiation," Cancer Res. 30, 942-951.