DNA double-strand breaks in mammalian cells after exposure to intercalating agents

129

Biochimica et Biophysica Acta, 654 (1981) 129-134

Elsevier/North-HollandBiomedicalPress BBA 99875 DNA DOUBLE-STRAND BREAKS IN MAMMALIAN CELLS AFTER EXPOSURE TO INTERCALATING AGENTS WARREN E. ROSSa and MATTHEWSO. BRADLEY b From the a Departments of Medicine and Pharmacology, University of Florida, Gainesville, FL 32610 and the b Merck, Sharp and Dohme Laboratory, West Point, PA 19486 {U.S.A.)

(Received February 10th, 1981)

Key words: DNA damage; Double-strand break; Intercalating agent; Ellipticine; Actinomycin D; Adriamycin

Previous work has shown that exposing mouse LI210 cells to intercalating agents such as adriamycin, ellipticine and actinomycin D results in DNA single-strand breaks and DNA-protein erosslinks. To characterize further the interaction between these drugs and intracellular DNA we have employed a modification of the alkaline elution technique which allows the detection of DNA double-strand breaks. Ellipticine (1.25-5.0 ~tg/ml), adriamycin (0.5-3.0/lg/ml) and actinomyein D (1.5-3.0/tg/ml) all caused double-strand breaks in DNA from LI210 cells following a 1 h treatment. The number of double-strand breaks found per single strand break was highest for eUipticine, despite the fact that this is the least cytotoxie of the three drugs. By comparing the single and double strand break frequency caused by radiation to that caused by elliptieine, it appears that most if not all of the drug-induced single strand breaks observed actually represent double-strand breaks. We suggest that these doublestrand breaks may result from the action of an intracellular enzyme, perhaps topoisomerase, which breaks both strands in concert to relieve the topological strain caused by drug intercalation.

Introduction Intercalating agents are a generally cytotoxie group of drugs which share the property of reversible binding to DNA by intercalation between adjacent base pairs [5]. Recent work from this laboratory [11, 12] and others [8,14] indicates that exposure of mammalian cells to these drugs causes DNA strand scission and DNA-protein crosslinks. Since intercalating agents do not cause strand breaks in purified DNA [8], Ross et al. [12] have hypothesized that the lesions observed in cellular DNA may result from the action of intracellular enzymes responding to topological perturbations imposed by intercalation of drug into DNA. Further, they have suggested that these enzymes could be repair endonucleases or topoisomerases. The latter are enzymes which can alter the topological conformation of DNA by creating DNA nicks and then resealing them [4]. Recently a group of ATP-dependent topoisomerases have been de-

scribed which act by breaking and rejoining both DNA strands in concert [10]. Under appropriate experimental conditions, actual DNA double-strand breaks can be observed [6]. It was thus of interest to determine if treatment of mammalian cells with intercalating agents would result in DNA double-strand breaks. To examine this question we have used a recent modification of the DNA alkaline elution technique which allows simple and reproducible detection of DNA double-strand breaks [1 ].

Methods Mouse leukemia LI210 cells, grown in suspension in RPMI 1630 medium with 10% horse serum, were employed in all experiments. Details of tissue culture technique and the labelling of cells with radioactive thymidine have been published [7]. Drugs were obtained from the Developmental

0 005-278 7/81/0000-0000/$02.50 © Elsevier/North-HollandBiomedicalPress

130 K. The DNA was then eluted from the filter with tetrapropylanmaonium hydroxide, pH 12.1. Cells which contain [3H]DNA and had received 150 R prior to elution were included on each f'dter as internal standards. Total elution time was 15 h. A modification of the method described by Bradley et al. [1] was used to assay DNA double-strand breaks. Exactly 2.5 • l0 s cells were deposited on a polycarbonate filter (2/am pore size, Bio-Rad Laboratories) and lysed as in the single-strand break assay (see above). The native DNA was then eluted from the falter with 0.2% SDS in a buffer of tetrapropylammonium hydroxide at pH 9.6. Total duration of elution was 15 h.

Therapeutics Program of the Division of Cancer Treatment, National Cancer Institute. Adriamycin was dissolved in distilled water, while actinomycin D was dissolved in ethanol and ellipticine in 0.01 N HC1. Cells were resuspended into fresh warm (37°C) medium at 6 • lO s cells/ml 1 h prior to drug treatment. Drug treatment was for 1 h at which time cells were washed twice and resuspended into fresh cold medium. Cells were radiated on ice using a 137Cs source (Mark I Irradiator, J.L. Sheppard and Assoc.) and kept cold until the time of elution. Rates of exposure were 225 R/min in the 0 - i 000 R range and 2 250 R/ rain for total doses above 1 000 R. The alkaline elution technique for assaying DNA single strand breaks has been described in detail elsewhere [7]. Cells containing 14C-labelled DNA were layered onto a poly(vinyl chloride) filter (pore size, 2/am) and lysed with a solution of 2% SDS, 20 mM Na2EDTA, 40 mM glycine and 0.5 mg/ml proteinase

Results Radiation produces both single and double-strand breaks in DNA and provides a useful calibration reference. Irradiating L1210 cells with doses of 150-900

O~

I0

06

08

°°1

I~

" ~ 150 R

Z U

Z

g

\ \\\

z o ~- 0 2

OE

~) \

-

~ ~I

0L .~..z~,7KR

(L 3~ ~-

LL

02

pH[Z I ~

I0

_

_

08

pH 96

~)600R

~900R 06

FRACTION 3 H - D N A

0.4 RETAINEO

3

6 HOURS

9

1~2

15

OF ELUTION

Fig. 1. Effect o f radiation on elution o f DNA from L1210 cells under denaturing (left) and non-denaturing conditions (Aght). L1210 cells on ice were radiated with 0 - 9 0 0 rads for the single-strand break assay or 0 - 9 000 rads for the double-strand break assay. Cells were lysed in the presence o f proteinase K and eluted at pH 12.1 (left) or pH 9.6 (right). Cells containing [ 3 H ] D N A were radiated with 150 rads on ice and included on the filter as an internal standard for the single strand break assay.

131 rads causes single strand breaks which can be observed by eluting the DNA under denaturing conditions (Fig. 1, left). In contrast, radiation doses approximately 10-fold higher are required to increase the elution rate of DNA under non-denaturing conditions (Fig. 1, right). This is consistent with the observation of others [9,15] that gamma radiation produces approximately one double-strand break per 10-20 single-strand breaks. The relationship between radiation dose and the fraction of DNA retained on the fdter after 13.5 h of elution was linear over a range of 0 - 9 000 rads (Fig. 2). The error bars provide an indication of the reproducibility of the technique. This laboratory and others [ 11] have previously shown that treating mammalian ceils with ellipticine

IO

u-08

ZO 7 0 c~ ~0~ g

, 04

z

_o

o2

Fig. 2. Relationship between radiation dose and retention of DNA on filter. L1210 cells were irradiated on ice, lysed and the DNA eluted from the filter at pH 9.6. The fraction of DNA remaining on the filter after 13.5 h of elution is plotted against the radiation dose. Each point represents the mean +S.E.

Oi

KRADS

IO

CONTROL

08

IO

Z

zO,~

~

08

o 06

--~

D 372

/.zg/ml

)o,.

I--

@: 0.4 U,-

u. ELLIPTICINE pHi2~

,o

OB 06 FRACTION 3H-DNA RETAINED

ELLpHPTI~C'NE

~ C O N T R O L

125y.g/ml

~

3 75~,Q / ml

5 tu., Ira'

0.2

04

~

6

9

~'2

~5

HOURS OF ELUTION

Fig. 3. Ellipticine causes single- and double-strand breaks in DNA from L1210 cells. Cells were treated for 1 h with various doses of ellipticine. After drug removal the cells were lysed and the DNA eluted at pH 12.1 (left) or 9.6 (right).

132 IO

08

061

I0

o

~05~g/mt

z

z 08

O5F~g/ml

z o a. 0

o

g

06

Z _(2

u-04

O2

ADR~AMYCIN pH 9 6 t5 i~q/

,o

'

o's

ds

'

0'4

02

~

'

~

FRACTION 3H-DNA RETAINED

A

b

,~

,~

HOURS OF ELUTION

Fig. 4. Adriamycin-induced DNA breaks. Cells were treated as in Fig. 3. Approximately the same range of single-strand break frequencies are observed as in Fig. 3 but fewer double-strand breaks are evident.

produces DNA single strand breaks. This effect is illustrated with L1210 cells in Fig. 3 (left). To determine if DNA double strand breaks also result from treating L1210 cells with these doses of ellipticine, the DNA was eluted under non-denaturing conditions (Fig. 3, fight). A dose-dependent increase in the elution of DNA from the f'dter was observed, indicating the presence of double-strand breaks. It is of interest to note that a given concentration of eUipticine produced a much higher frequency of double-strand breaks relative to single-strand breaks than was the case for radiation. Adriamycin and actinomycin D share with ellipticine the ability to intercalate between DNA base pairs and cause single-strand breaks [12]. It was thus of interest to determine whether these drugs might also cause DNA double-strand breaks. As shown in Figs. 4 and 5, both of these drugs produced both single (on the left) and double strand breaks (on the right) but

there were fewer double-strand breaks per singlestrand break than was the case for eUipticine.

Discussion Our data indicate that intercalating agents as a group cause DNA double-strand breaks in mouse L12!0 cells. Byfield et al. [2] observed a similar effect following treatment of rat sarcoma cells with adriamycin. However, these workers employed neutral sucrose gradient centrifugation to demonstrate double-strand breaks, their system was not calibrated and they studied no other intercalating drugs. The technique we have used to study double-strand break formation is an adaptation of the DNA alkaline elution technique. That the technique actually measures double-strand breaks is supported by three lines of evidence as originally described by Bradley et al, [1]. First, the DNA which elutes from the triter is double

133 1.0; CONTROL

0 8

06

I01~ ~l

.~ .

o ~ 0 8 1 -I

Z04

~

~

~

_

~

-- ~

~

-

~ CONTROL

- - - - - ~. , 5 ,u0/m,

LI.I z O.E ,5 ,q-

z I-

j 5 p.~/ml

,-,- o 2

LL

LL

0.4

02

ACTINOMYCIND pH121 ,o

'

o'8

ACTINOMYCIN O pH9,6

66

'

o'.4

'

FRACTION 3H-DNA RETAINED

~

~

$

,~

,'5

HOURS OF ELUTION

Fig. 5. Actinomycin D-induced DNA breaks. Cells were treated as in Fig. 3. The results are similar to those observed with adriamycin.

stranded by cesium chloride gradient centrifugation. Second, the method detects the double-strand cuts made by restriction endonucleases. Finally, the characteristic ratios of double-strand break frequencies produced by X-irradiation and bleomycin are observed using this technique. Correct interpretation of results obtained with this technique requires that the drug(s) not directly alter the DNA in such a way as to increase (or decrease) the dution rate independent of strand size. To confirm this we have added adriamytin (3/.tg/ml) directly to irradiated (5 000 R) and unirradiated cellular DNA isolated on the Filter just prior to elution; no effect on the rate of elution of the DNA was observed (data not shown). Intercalating agents create far more double strand breaks per single strand break than does radiation. In fact, in the case of ellipticine it could be argued that most if not all of the observed single strand breaks are actually double strand breaks which appear as the former under denaturing conditions. This inter-

pretation is based on the fact that radiation produces roughly 10-20 single-strand breaks for every doublestrand break [9,15]. Therefore, if elliptieine produced only double strand breaks, one would expect that a given concentration would cause 10-20 radequivalents of double strand breaks per rad-equivalent of single strand breaks. Such a relationship is in fact observed (see Figs. 1 and 3). Based on available evidence, it is unclear whether intercalator-induced single and double strand breaks are formed by a common mechanism. We consider there to be three possibilities to explain the coexistence of the two lesions. The first is that the double strand breaks represent closely spaced single strand nicks created by a single strand specific endonuelease. This hypothesis is suggested by an observation of Center [3]. Isolated nuclei from mouse fibroblasts were incubated with adriamycin and a single-strand specific endonuclease from Neurospora crassa. When these nuclei were then sedimented in neutral sucrose

134 gradients, DNA double strand breaks were demonstrated, suggesting that drug-induced local strand separation allowed endonucleolytic cleavage of both strands. A second explanation for the two lesions is that strand scission results from the action of a single enzyme which cuts both strands in concert. It may be argued that, since adriamycin and actinomycin D cause fewer double-strand breaks per observed singlestrand break than ellipticine, the two types of break can be mechanistically dissociated. However, both adriamycin and actinomycin D have bulky side groups while eUipticine does not. Such side groups might interfere with an enzyme-mediated nicking or closing mechanism of one of the strands, thereby resulting in half-formed or half re-sealed double strand breaks which would appear as single strand breaks. Finally, consideration must be given to the possibility that enzymes of both types may be operative. Further work is necessary to clarify this issue. DNA double-strand breaks are generally considered highly lethal lesions. We observe the highest frequency of double-strand breaks (relative to singlestrand breaks) following treatment with eUipticine. This is of interest, since this laboratory has previously shown that ellipticine causes far less cytotoxicity per single-strand break than either adriamycin or actinomycin D. In fact, the highest concentration of ellipticine used in this study, 5,0 ~g/ml, results in less than one log L1210 cell kill while both adrianlycin and actinomycin D at 3.0/.tg/rrd produce greater than 4 logs cell kill [13]. Given the observation that these drugs cause strand breakage in cellular DNA but not in purified DNA [8], and that the frequency of double-strand breaks does not correlate well with cytotoxicity, it is reasonable to postulate that the strand scission may represent an effort by the cell to relieve the topological perturbation caused by intercalation of drug between base pairs. An attractive hypothesis is that such a response is mediated by an enzyme which can break one or both strands in concert, thereby relieving torsional strain. The enzyme would then re-seal the strands so as to maintain structural integrity. Such an enzyme would be similar to, if not identical with, a DNA topoisomerase [4].

In summary, double strand breaks are caused by a variety of intercalating agents. These breaks do not appear to be directly cytotoxic and may in fact represent a repair process unique to the type of damage caused by intercalating agents. Further elucidation of the complex series of events which occur following intercalation of drug into nuclear DNA should improve our understanding of the basis for their antitumor effects.

Acknowledgements The authors wish to thank Myra Smith for her technical assistance and Suzanne Owens for the preparation of this manuscript. This work was supported by NIH Grant No. R01-CA-24586 and RCDA No. KO4-CA-00537 (to W.E.R.).

References 1 Bradley, M.O. and Kohn, K.W. (1979) Nucleic Acids Res. 7,793-804 2 Byfield, J.E., Lee, Y.C. and Tu, L. (1976) Int. J. Cancer 19,186-193 3 Center, M.S. (1979) Biochem. Biophys. Res. Commun. 84, (4) 1231-1238 4 Champoux, J.J. (1978) Annu. Rev. Biochem. 47,449479 5 Fuller, W. and Waring, M.J. (1964) Ber. Bungsenges. Phys. Chem. 68,805-808 6 GeUert, M., Mizuuchi, K., O'Dea, M.H., Itoh, T. and Tomizawa, J. (1977) Proc. Natl. Acad. Sci. USA 74, 47724776 7 Kohn, K.W., Erickson, L.C., Ewig, R.A.G. and Friedman, C.C. (1976) Biochemistry 15, 4629-4637 8 Lee, Y.C. and Byfield, J.E. (1976) J. Natl. Cancer Inst. 57,221-224 9 Lehmann, A.R. and Ormerod, M.G. (1970) Biochim. Biophys. Acta 217,268-277 10 Liu, L.F., Liu, C.C. and Alberts, B.M. (1979) Nature 281, 456-461 11 Ross, W.E., Glaubiger, D.L. and Kohn, K.W. (1978) Biochim. Biophys. Acta 519, 23-30 12 Ross, W.E., Glaubiger, D.L. and Kohn, K.W. (1979) Biochim. Biophys. Acta 562, 41-50 13 Ross, W.E., Zwelling, L.A. and Kohn, K.W, (1979) Int. J. Radiation OncologyBiol. Phys. 5, 1221-1224 14 Schwartz, H.S. (1975) Res. Commun. Chem. Path. Pharmacol. 10, 51-64 15 Veatch, W. and Okada, S. (1969) Biophys. J. 9,330-346