Proc. Natl. Acad. Sci. USA Vol. 75, No. 12, pp. 6149-6153, December 1978 Genetics Tumor promoter induces sister chromatid exchanges: Relevance to mechanisms of (12-0-tetradecanoylphorbol 13-acetate/bromodeoxyuridine labeling/recombination/mitotic segregation/recessive mutations) ANNE R. KINSELLA AND MIROSLAV RADMAN D6partement de Biologie Moleculaire, Universite Libre de Bruxelles, B 1640 Rhode-St-Genese, Belgium Communicated by J. D. Watson, August 31, 1978

ABSTRACT 12-0-Tetradecanoylphorbol 13-acetate (TPA), Six possible mechanisms for the segregation of a recessive a powerful tumor promoter, is shown to induce sister chromatid mutation from a heterozygous cell are presented in 1. exchanges (SCEs), whereas the nonpromoting derivative 4-0- Fig. (i) methyl-TPA does not. Inhibitors of tumor promotion-antipain, Chromosomal rearrangement or deletion and (ii) one-step leupeptin, and fluocinolone acetonide-inhibit formation of nondisjunction both lead to hemizygosity, while (iii) two-step such TPA-induced SCEs. TPA is a unique agent in its induction nondisjunction, (iv) aberrant mitotic segregation (in the absence of SCEs in the absence of DNA damage, chromosome aberra- of nondisjunction or mitotic recombination), (v) increase in tions, mutagenesis, or significant toxicity. Because TPA is ploidy plus chromosome loss, and (vi) mitotic recombination known to induce several gene functions, we speculate that it all lead to might also induce enzymes involved in . homozygosity. Only the events leading to homozy- Thus, the irreversible step in tumor promotion might be the re- gosity (iii-vi) are consistent with the observation that initiation sult of an aberrant mitotic segregation event leading to the ex- must precede promotion in order to produce an enhanced pression of carcinogen/mutagen-induced recessive genetic or carcinogenic effect. Furthermore, mechanisms i and ii are epigenetic chromosomal changes. made unlikely by the fact that chromosome loss or deletion is usually lethal to normal diploid cells (11). Carcinogenesis is usually a multistep process. Epidemiological We present here a cytological study of the effects of TPA on studies of human carcinomas indicate that several distinct Chinese hamster cells in culture to test the hypothesis of tumor heritable changes are necessary to change a normal cell into a promotion outlined in Fig. 1. This study was encouraged by the malignant cell (1). This stepwise process is best illustrated in the demonstration of promotion processes in cell culture (ref. 12 two-stage mouse skin carcinogenesis system, which is composed and references therein) and by the observation that griseofulvin, of "initiation" and "promotion" stages (2, 3). Initiators, such a known promoter of dimethylbenzanthracene-initiated mouse as radiation or chemical carcinogens, cause rapid irreversible skin tumors (13), induces high ploidy and subsequent chro- changes, mutagenic or epigenetic in origin, which result in a mosome loss in cell cultures (14), which would be consistent with "premalignant" state that is inherited by both daughters when mechanism v listed above. "initiated" cells divide. Promoters are neither significantly carcinogenic (3) or mutagenic (4, 5) when tested alone, but MATERIALS AND METHODS greatly increase tumor frequency and shorten the lag time for tumor appearance when added after an initiator. Thus, tumor V79-4 Chinese hamster lung fibroblasts were maintained in promoters complete a process begun by initiators. Dulbecco's modification of Eagle's minimal essential medium These facts prompted us to consider the possibility that ini- supplemented with 10% fetal bovine serum in a humidified tiated cells contain specific recessive, autosomal, somatic incubator at 37°C in the presence of 5% CO2. For analysis of mutations and that promotion causes the expression of these sister chromatid exchange (SCE), exponentially growing cells mutations by inducing an aberrant mitotic segregation event. were exposed simultaneously to 30 ,M bromodeoxyuridine The initiated heterozygous cell is thus converted into a homo- (BrdUrd) and the tumor promoter TPA (diluted from a 10 zygous or hemizygous cell (see Fig. 1 and legend). This would mg/ml stock solution in acetone) for 28 hr (i.e., two genera- be consistent with the finding that the nonviral malignant (or tions). TPA and its derivative 4-O-methyl-TPA (4-O-Me-TPA) transformed) phenotype is usually suppressed on hybridization were obtained from the Consolidated Midland Chemical with nontransformed cells and that malignant segregants can Corporation, Brewster, NY. Similar incubations were carried reappear as a result of chromosome loss (7, 8). Although the out in the presence of inhibitors of tumor promotion such as term "recessive mutation" is used throughout this paper, our antipain and leupeptin or fluocinolone acetonide (FA). These hypothesis is compatible with any model of initiation based inhibitors were the gifts of W. Troll (New York University) and upon stable, recessive, genetic, or epigenetic chromosomal S. Yuspa (National Institutes of Health, Bethesda, MD), re- changes. spectively. Control cultures were treated with the same volume Aneuploidy and aberrant mitotic segregations have already of acetone as the TPA-treated cultures (10 ,ul per 10 ml of me- been postulated as a prerequisite for malignant transformation dium). Until chromosome fixation, all cultures were kept in the (1, 9). This speculation was applied specifically to tumor pro- dark. Twenty-six to 28 hr after TPA treatment, Colcemid (0.2 motion (6), and a preliminary study has shown that a nontoxic ,uM) was added to the cells for 2 hr, and the mitotic cells were dose of the tumor promoter 12-0-tetradecanoylphorbol 13- isolated by "mitotic shake-off." Metaphase preparations and acetate (TPA) induces the segregation of two recessive traits, differential staining were essentially as described by Perry and 6-thioguanine resistance and growth on agar, from doubly Wolff (15). All visible metaphases were analyzed for SCEs. heterozygous cell hybrids (10). Because promoting activity of TPA is correlated with skin ir- ritation and inflammation (3), the activity of the agents used The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "ad- Abbreviations: SCE, sister chromatid exchange; TPA, 12-0-tetra- vertisement" in accordance with 18 U. S. C. §1734 solely to indicate decanoylphorbol 13-acetate; 4-O-Me-TPA, 4-O-methyl-TPA; BrdUrd, this fact. 5-bromodeoxyuridine; FA, fluocinolone acetonide. 6149 6150 Genetics: Kinsella and Radman Proc. Nati. Acad. Sci. USA 75 (1978) A MUTAGENIC__| -|- 2n .- AGENT | somatic cell z cK to Induction Direct 0 of cellular mutagenic mutagenesis .2 processes

W4 Fi < \ >~~~~~~~~Oa % 0 M BM M 3 M l 0 E 0 0L Vo M@ Ml \ M \[E[a~~~~~ AD t a" [3 Oco Dilution of the wild-type protein 11Mao - x Expressed mutant phenotype FIG. 1. Fixation, segregation, and expression ofrecessive mutations in somatic cells as a model in the study of carcinogenesis. For sim- C plicity, only one homologous chromosome pair of a diploid somatic ds- cell is shown. The homozygous (+/+), wild-type constitution for one NNW- gene on this chromosome pair can be changed to the heterozygous A&- state (m/+) by exposure to mutagenic agents. Depending on the na- 19M ture of the mutagen, the mutation fixation either can occur directly 4w during DNA replication or may require induction of cellular muta- genic processes (6). Circles symbolize centromeres. Broken lines with .mr.,W. -W -.IRA open circles represent the newly synthesized chromatids. M, homo- 0 t zygous (/rm) or hemizygous (inO) mutant cells in which phenotype *k.*- will be expressed. See text for other comments. T v ol in this study was tested by spotting 10 ,l of 10 Ag/ml solutions 41 in acetone onto the ears of BALB/c mice. TPA spots caused 1 inflammation within 12 hr, while 4-0-Me-TPA and acetone V 40, alone did not. laid 0 RESULTS AddL--- 00 lb To test the hypothesis presented in Fig. 1, the effects of TPA Ah on ploidy, chromosome loss, chromosomal aberrations, and SCEs have been studied. Only an increase in SCE frequency was observed (Fig. 2 A, B, and C). In addition to the approxi- mately 2-fold increase in countable SCEs seen in the majority D e N_ of the cells,* up to 25% of the TPA-treated cells showed SCEs that were "too many to count" (more than 100 SCEs per me- ".~ ~~~~ taphase) (Fig. 2C). This "two-population effect" was also ob- served with two independent V79 clones derived from our cultures, thus eliminating the involvement of stable mixed cell populations (results not shown). TPA toxicity was seen to vary between batches, but in general 4e4 the maximum induction of SCEs was found in the absence of major toxicity as measured by cloning efficiency (Table 1) and growth curves (Fig. 3). The optimum SCE induction occurred at 1.6 ,gM (1 ,g/ml) TPA (Table 2); TPA at 5 ,gg/ml was lethal without causing further increase in SCEs (Table 2). In an attempt to correlate SCE induction with mouse skin FIG. 2. Effects of TPA on chromosomes from V79 cells differ- tumor promotion, 4-0-Me-TPA, a nonpromoting derivative entially stained for detection of SCEs. SCEs are shown in (16), was found to have no effect on SCEs (Fig. 4), although its metaphase chromosomes from untreated control cells (A) and cells treated with TPA at 1 ,tg/ml (B and C). A and B show 8 and 34 SCEs, * Because TPA induction of SCEs is independent of BrdUrd concen- respectively, while C is an example of metaphases that were classified tration (see Table 2), we assume that BrdUrd is not a requirement as having "too many SCEs to count" (see Table 2). (D) TPA-treated for SCE induction by TPA. metaphase with "banded" chromosomes discussed in the text. Genetics: Kinsella and Radman Proc. Natl. Acad. Sci. USA 75 (1978) 6151

Table 1. Cloning efficiency of V79 cells in BrdUrd, acetone, TPA, and the nonpromoting derivative 4-Q^-Me-TPA % of Clones per dish, untreated Additions to growth medium mean :1 SD control None 234.2 + 23.5 100 BrdUrd 172.0 + 9.3 73.3 BrdUrd + acetone 191.7 ± 7.04 81.8 BrdUrd + TPA (0.1 Mg/ml) 155.0 ± 12.5 66.2 BrdUrd + TPA (1.0~g/ml) 153.0 ± 14.1 65.3 BrdUrd + 4-O-Me-TPA (1.0 ,g/ml) 142.7 + 9.5 60.9 Cells were plated at a density of 500 per 60-mm plastic petri dish in 5 ml of medium. The mean number of clones per dish was deter- mined from four petri dishes. When added, BrdUrd was 30 gM. Ac- etone was added alone or as the TPA solvent at a concentration of 0.1%. Plating efficiency was 47%. Days toxicity was higher than that of TPA (Table 1). Furthermore, FIG. 3. Growth curves for V79 cells in the presence ofBrdUrd and two classes of inhibitors of TPA-induced tumor promotion were TPA. Growth was in modified minimal essential medium with 10% tested: the protease inhibitors antipain and leupeptin (ref. 17 fetal calf serum, without any addition (0); with 100MgM BrdUrd (0); and W. Troll, personal communication) and the steroidal with 100 AM BrdUrd and TPA at 1 ,g/ml (a); and TPA at 5 Mg/ml anti-inflammatory agent fluocinolone acetonide (FA), the most (A). Bars indicate standard deviations. potent inhibitor of mouse skin tumor promotion (18). Fifty to 80% inhibition of TPA-induced SCEs was seen in the presence toxicity (not shown; see also labeled metaphases in Table 2). of 1 mM concentrations of either antipain or leupeptin (Table These observations, together with the fact that TPA is known 2). Antipain was seen to inhibit the "too many to count" class to induce transiently several cellular changes and specific en- of TPA-treated metaphases (Fig. 5, Table 2). The simultaneous zymes (3, 19, 24), have led us to postulate that TPA-provoked presence of both inhibitors did not seem to enhance the inhib- SCE induction might result from the induction of cellular DNA itory effect (Table 2, entry II). FA (2.7 jM) gave 100% inhibi- recombination enzymes. tion of TPA SCE induction and at the doses usedt appears to The lack of SCE induction by the nonpromoting derivative be the most effective inhibitor. 4-O-Me-TPA (Fig. 4) and the inhibition of SCE induction by In addition to SCEs, the only noticeable effects of TPA on two known classes of inhibitors of tumor promotion (Table 2) V79 chromosomes were the appearance of "banded" chro- strongly implicate the role of SCE-type recombinogenic activity mosomes (Fig. 2D) (which were not seen in control metaphases in tumor promotion. The SCE-inducing effect of TPA can be and whose appearance was inhibited by antipain but not by FA) and an increase in unlabeled metaphases (Table 2), the latter indicating an inhibition of DNA replication. A transitory in- TPA I 4-0-Me-TPA 5 10 15 0 5 10 15 20 hibition of DNA synthesis by TPA has been seen both on mouse £~~~~~~~~~~ skin and in cell culture systems (for a review see ref. 19). 1cD 1.0 jug/mI £ 1.0,ug/mli 10

DISCUSSION X= 12.91 3.39 £££X=8.34+2.17& 9 AAAAA 5 The most striking chromosomal effect of TPA on V79 cells in culture is an increased frequency of SCEs (Fig. 2 B and C). The appearance of banded chromosomes (Fig. 2D) in TPA-treated 1c 1 * *- - --. .- ---m I *.------~ 1 I 10 cells may be tentatively interpreted either as being due to a partial, discontinuous incorporation of BrdUrd because of the X=7.88 +3.06 X=7.61 +3.37i 5 i ...... 5s inhibition of DNA elongation in many chromosomal replicons 1 .4* mug I~mI0jgmmu... or as a visualization of TPA-induced structural changes in the * U. mrea@-@.. U chromatin (e.g., proteolytic removal of some components of the chromatin). o 0 ,g/mI Opg/mI Three observations implicate TPA as a unique inducer of 0 0 0 o0 - SCEs: other 10 00 X= 5.32+ 1.98 X 7.69 2.16 140 (i) SCE-inducing compounds cause damage to 00 0 DNA (20-22), whereas TPA does not (3, 23); (ii) the increase 0 0 0 0 0 0 00 0 in the SCE frequency even in the "too many to count" class 5 00000 00 0 -5 000000 0 0 0 occurs in the absence of a detectable increase in chromosome 000000 00 00 aberrations normally seen with other SCE-inducing agents; (iii) l1 0000000 00000 unlike other SCE-inducing agents, TPA at effective concen- un 0 u I D Wun trations is nontoxic (Fig. 3) and nonmutagenic for V79 cells (4, SCEs per metaphase 5). The increase in SCEs after treatment with TPA is not related FIG. 4. SCE induction in V79 cells treated with tumor promoter to cellular toxicity effects: 4-O-Me-TPA is at least as toxic as TPA and its nonpromoting derivative 4-O-Me-TPA. All visible me- TPA (Table 1), high doses of TPA are less effective SCE in- taphases with countable SCEs from two independent experiments ducers than were analyzed. Approximately 20% ofall cells in TPA-treated cultures nontoxic doses (Table 2), and FA potentiates TPA showed "too many to count" SCEs (Table 2), which were not seen in 4-O-Me-TPA treated culture. The difference in the spontaneous SCE t The doses were recommended by W. Troll for antiproteases and S. levels in the controls can be accounted for by the difference in BrdUrd Yuspa for FA, on the basis of their work with these agents in cell concentrations used: 30 MM and 100 MM in the TPA experiment and culture systems. in the 4-O-Me-TPA experiment, respectively. 6152 Genetics: Kinsella and Radman Proc. Natl. Acad. Sci. USA 75 (1978) Table 2. Inhibition of TPA-induced SCEs by inhibitors of tumor promotion % "Too many Unlabeled Meta- SCEs/ Induc- inhibition SCEs," meta- BrdUrd, phases metaphase tion of % total phases, Treatment AM analyzed + SD Range factor induction metaphases % total I. Antipain Control (acetone) 100 93 10.5 ± 4.0 4-20 1.0 - 0 0 TPA (1 tg/ml) 100 64 25.5 ± 9.1 10-58 2.4 25.6 1.0 TPA (1 ,g/ml) + 1 mM antipain 100 56 15.1 ± 5.0 7-29 1.4 70 5.3 2.0 TPA (5 Ag/ml) 100 38 17.5 ± 6.9 9-45 1.7 24.0 12.3 II. Antipain; leupeptin Control (acetone) 30 44 4.1 ± 1.3 2-7 1.0 - 0 0 TPA (1 ig/ml) 30 83 8.9 + 4.0 2-21 2.2 20.5 20.0 TPA + 1 mM antipain 30 51 6.2 + 2.9 2-14 1.5 56.7 0 7.8 TPA + 1 mM leupeptin 30 41 5.3 + 1.8 1-9 1.3 76.4 17.0 1.7 TPA + 1 mM antipain + 1 mM leupeptin 30 59 6.2 ± 1.9 3-12 1.5 56.7 6.8 6.8 III. Fluocinolone acetonide Control 30 66 4.3 ± 1.7 2-9 1.0 0* 0 TPA (1 ,g/ml) 30 134 7.9 ± 3.1 3-24 1.8 18.6* 17.9 FA (1 ,g/ml) 30 119 4.3 + 2.0 1-12 0.99 - 21.0* 20.1 TPA + FA 30 144 3.9 + 2.0 0-10 0.89 113 25.7* 18.7 Appropriate concentrations of the inhibitors were added simultaneously with TPA and the cultures were incubated for 27 hr. Antipain and leupeptin stock solutions were in water and TPA and FA in acetone; the controls were adjusted accordingly. The concentrations of inhibitors used were nontoxic and did not themselves induce SCEs. About 10 independent replicate experiments were performed, and all showed similar SCE-inducing effects of TPA. *No absolute distinction could be made between "too many to count" and banded chromosomes. related to the mitotic recombination mechanism in the hy- so far failed (26, 27), indicating that, if it occurs, mitotic re- pothesis of promotion in Fig. 1 only by assuming that SCE combination must be very rare. However, cytological evidence frequency can be considered as a cytological indication of shows that quadriradial chromosomes, although rare, represent cellular recombinogenic activity (20, 22, 25), particularly in the exchanges between both sister and nonsister homologous absence of DNA damage. All attempts to detect mitotic re- chromatids (25, 28). Increased frequency of quadriradial combination in mammalian cells by using genetic markers have chromosomes has been observed only in cells containing very high levels of either spontaneous (28) or induced (ref. 25; un- published data) SCEs. Hence, the assumption that SCEs reflect 15 cellular recombinational activity seems reasonable, although :: Control 1 ll of acetone no quantitative relationship between either the absolute fre- X 10.5 ± 4.0 10l a a n ='98 quency or the increase in frequency of SCEs and mitotic re- * fl.i combination frequency can be established. 5 *@ Bloom's syndrome, the striking human cancer-prone he- swassuessusessess a evidence our ... reditary disease, provides circumstantial favoring 4 8 12 16 20 24 28 36 hypothesis, because cells taken from these patients are char- acterized by very high levels of spontaneous SCEs and of qua- TPA (1 .,0 ,Ag/ml) + 1 mM antipain driradial chromosomes (28). Thus, we propose that Bloom's w 15 X =15.06 ± 5.01 syndrome may be caused by a mutation resulting in a promo- n= 56 a tion-constitutive phenotype. aCLI The fact that radiation and chemical mutagens not only are a initiators but also are carcinogenic at high doses would be

8 s 1s 2s consistent with our specific hypothesis of "two-stage" carci- m....mu...mm. nogenesis, because high doses of initiators are known to induce 4 8 12 16 20 24 28 both mutagenesis and SCEs in mammalian cells (20-22) and 20L mitotic recombination in Drosophila (29), Ustilago (30), and TPA (1.0,g/ml) yeast (31). The potential of initiators to serve as "solitary" X = 25.53 + 9.06 carcinogens at high doses may be masked by their toxic effects n = 64 on affected cells. In contrast, the negligible toxicity of the TPA 10 concentrations used to induce SCEs in vitro (Table 1, Fig. 3) and the low toxicity and proliferative effects of tumor pro- moters in vivo (3) can be expected to amplify greatly the effi-

: . of mitotic ...! .: ciency segregation by recombination, particularly 4 8 12 16 20 24 28 32 36 40 58 after multiple TPA applications. SCEs per metaphase I It is implicit that this hypothesis applies to late and irre- FIG. 5. Distribution of TPA-induced SCEs in the presence and versible step(s) in the tumor promotion process and does not absence of antipain. The presence of 1 mM antipain inhibited TPA- contradict observations on reversible effects of TPA such as induced SCEs, both countable and "too many to count," by ap- mimicry of transformation and inhibition of terminal differ- proximately 70%; 1 mM antipain alone had no effect on spontaneous for tumor such SCEs. Hatched bars, SCEs "too many to count"; empty bars, unla- entiation (3, 19). Other hypotheses promotion, beled metaphases. as derepression of mutant genes causing a stable switch in gene Genetics: Kinsella and Ra'dman Proc. Natl. Acad. Sci. USA 75 (1978) 6153 expression (3, 4, 19, 32) or induction of enzymes required for 13. Barich, L. L., Schwarz, J. & Barich, D. (1962) Cancer Res. 22, mutation fixation (33), are not necessarily in contradiction with 53-55. our hypothesis; several events might be,required to accomplish 14. Larizza, L., Simoni, G., Tredici, F. & DeCarli, L. (1974) Mutat. Res. 25, 123-130. tumor promotion (3), and these may reflect qualitatively dif- 15. Perry, P. & Wolff, S. (1974) Nature (London) 251, 156-158. ferent processes. 16. Hecker, E. (1978) in Carcinogenesis, eds. Slaga, T. J., Sivak, A. We thank S. Mousset and J. and C. Szpirer for stimulating discussions; & Boutwell, R. K. (Raven, New York), Vol. 2, pp. 11-48. Drs. J. Cairns, R. Peto, R. Thomas, and I. B. Weinstein for their criti- 17. Hozumi, M., Ogawa, M. Sugimura, T., Takeochi, T. & Umezawa, cism; Dr. P. Perry for advice on SCE methodology; Dr. S. Wolff for H. (1972) Cancer Res. 32, 1725-1728. his evaluation of the "too many SCEs to count" class of metaphases; 18. Viaje, A., Slaga, T. J., Wigler, M. & Weinstein, I. B. (1977) Cancer Mr. R. Legas for technical assistance; and the U.S.-Japan Cooperative Res. 37, 1530-1536. Cancer Research Program for the gift of limited amounts of antipro- 19. Weinstein, I. B., Wigler, M., Fisher, P. B., Sisskin, E. & Pietro- teases. A.K. holds a Royal Society research fellowship. This work was paolo, C. (1978) in Carcinogenesis, eds. Slaga, T. J., Sivak, A. & supported by Fonds Cancerologique of the Caisse d'Epargne et de Boutwell, R. K. (Raven, New York), Vol. 2, pp. 313-333. Retraite'and by Euratom-Universite Libre de Bruxelles research 20. Kato, H. (1977) Int. Rev. Cytol. 49,55-97. contract 224-76-I BIO B. 21. Perry, P. & Evans, H. J. (1975) Nature (London) 258, 121- 125. 1. Peto, R., Roe, F. S. C., Lee, P. N., Levy, L. & Clark, J. (1975) Br. 22. Wolff, S. (1977) Annu. Rev. Genet. 11, 183-201. J. Cancer 32,421-426. 23. Poirier, M. C., De Cicco, B. T. & Lieberman, M. N. (1975) Cancer 2. Berenblum, I. (1975) in Cancer 1, a Comprehensive Treatise, Res. 35, 1392-1397. ed. Becker, F. F. (Plenum, New York), pp. 323-344. 24. Verma, A. K. & Boutwell, R. K. (1977) Cancer Res. 37, 2196- 3. Boutwell, R. K. (1974) CRC Crit. Rev. Toxicol. 2,419-443. 3001. 4. Trosko, J. E., Chang, C., Yotti, L. P. & Chu, E. H. Y. (1977) 25. German, J. (1964) Science 144, 298-301. Cancer Res. 37, 188-193. 26. Rosenstraus, M. J. & Chasin, L. A. (1976) J. Cell Biol. 70, 5. Lankas, G. R., Jr., Baxter, C. S. & Christian, R. T. (977) Mutation 202a. Res. 45, 153-156. 27. Tarrant, G. M. & Holliday, R. (1978) Mol. Gen. Genet. 156, 6. Radman, M., Spadari, S. & Villani, G. (1978) J. Nat!. Cancer Inst., 273-279. in press. 28. Chaganti, R. S. K., Schonberg, S. & German, J. (1974) Proc. Nat!. 7. Stanbridge, E. J. (1976) Nature (London) 260, 17-20. Acad. Sci. USA 71, 4508-4512. 8. Jonasson, J., Povey, S. & Harris, H. (1977) J. Cell. Sca. 24,217- 29. Stern, K. (1936) Genetics 21, 625-730. 254. 30. Holliday, R. (1964) Genetics 50,323-335. 9. Ohno, S. (1974) in Chromosomes and Cancer, ed. German, J. 31. Roman, H. L. & Jacob, F. (1958) Cold Spring Harbor Symp. (Wiley & Sons, New York), pp. 77-94. Quant. Biol. 23, 155-160. 10. Kinsella, A. R., Mousset, S., Szpirer, C. & Radman, M. (1978) in 32. Cairns, J. (1977) in Origins of Human Cancer, eds. Hiatt, H., DNA Repair Mechanisms, eds. Hanawalt, P. C. & Friedberg, Watson, J. D. & Winsten, J. A. (Cold Spring Harbor Laboratory, E. C. (Academic, New York), in press. Cold Spring Harbor, NY), pp. 1813-1820. 11. Grote, S. J. & Revell, S. H. (1974) Curr. Top. Radiat. Res. 7, 33. Radman, M., Villani, G., Boiteux, S., Defais, M., Caillet-Fauquet, 303-309. P. & Spadasi, S. (1977) in Origins of Human Cancer, eds. Hiatt, 12. Kennedy, A. R., Mondal, S., Heidelberger, C. & Little, J. B. (1978) H., Watson, J. D. & Winsten, J. A. (Cold Spring Harbor Labo- Cancer Res. 38, 439-443. ratory, Cold Spring Harbor, NY), pp. 903-922.