v~— THESIS This is to certify that the dissertation entitled THE BIOLOGICAL EFFECTS OF AFLATOXIN BI AND AFLATOXIN BI-DICHLORIDE IN HUMAN CELLS IN CULTURE presented by EILEEN MAHONEY-LEO has been accepted towards fulfillment of the requirements for / VIE GM degree in til/IR 7w Major professor ‘ Date Mfl ’21 [‘H‘Z’ MSUis an Affirmative Action/Equal Opportunity Institution 0-12771 MSU BETURNING MATERIALS: Place in book drop to LIBRARIES remove this checkout from .—_—. your record. FINES will be charged if book is returned after the date stamped below. THE BIOLOGICAL EFFECTS OF AFLATOXIN B AND AFLATOXIN Bl-DICHLORIDE IN HUMAN CELLS IN CULTURE 1 BY Eileen Mahoney-Leo A DISSERTATION submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Genetics Interdepartmental Program 1982 ABSTRACT THE BIOLOGICAL EFFECTS OF AFLATOXIN B AND AFLATOXIN Bl-DICHLORIDE IN HUMAN CELLS IN CULTURE 1 BY Eileen Mahoney-Leo Aflatoxin B1 is one of the most powerful hepato- carcinogens known, however it requires metabolic activation before it can exert its biological effects. A human cell- mediated system was used to study the cytotoxicity of AFB1 in human cells in culture. Cytotoxicity was observed in normal human fibroblasts at concentrations as low as luM. Aflatoxin Bl-dichloride (AFBl-Clz) is a: direct- acting carcinogen which is a model compound for the pro- posed ultimate metabolite, aflatoxin 31-2,3-oxide. The cytotoxic and mutagenic effects of AFBl-Cl2 were stu- died in repair-proficient human fibroblasts (NF) and in excision repair-deficient fibroblasts derived from xeroderma pigmentosunt (XP) patients. XP cells were more Eileen Mahoney—Leo sensitive than NF cells to the cytotoxic and mutagenic effects of AFBl-Clz. NF cells were able to recover from the potential cytotoxicity and potential m‘utagenicity of AFBl-Cl2 when held in a non-replicating state for up to 10 days following treatment. XP cells showed no ability to recover from the cytotoxic effects of AFBl-Clz. The loss of DNA-bound AFBleClZ residues was monitored in NF and XP cells held in the non-replicating state for up to eight days following treatment. It was observed that more than 60% of the total initially bound AF‘Bl-Cl2 residues were lost from the DNA of NP and XP cells by day 7. More than 60% of the total initially bound adducts were also observed to be lost from calf thymus DNA reacted E vitro with radioactive AFB -C12 and l incubated for eight days. HPLC analyses of DNA adducts revealed the formation of seven adducts in calf thymus DNA and in the DNA of NF and XP cells. The major product was lost from calf thymus DNA by day 8, and all adducts diminished in size over time. The three major products were identified as guanine adducts. The same DNA adducts were formed initially in NF DNA as were seen in calf thymus DNA. ACKNOWLEDGEMENTS I would like to thank my family and friends for their encouragement and support throughout this endeavor. I acknowledge Mr. Ted Van Noord for carrying out the calf thymus DNA experiment and for performing the HPLC analyses. ii TABLE Chapter LIST OF TABLES . . . . . . LIST OF FIGURES . . . . . LIST OF ABBREVIATIONS . . INTRODUCTION . . . . . . . LITERATURE REVIEW . . . . Aflatoxins . . . . . Metabolism of AFB1 . Toxicity of AFB . . Carcinogenicity of AF 1 . . Mutagenicity of AFB AFB -induced DNA adducts OF CONTENTS Evidence of repair of the cytotoxic and mutagenic DNA lesions induced by AFB Possible Aechanisms of mutagenesis by AFB AFB -C1 A5 a model compoun for AltiEate metabolite, AFB -2,3-oxide Somatic Cell Mutation Theory 0 Cancer Evidence in support of the somatic cell mutation theory . the a) The majority of carcinogens tested have proved to be mutagens b) Clonal nature of tumors . . c) DNA excision repair-deficient xeroderma pigmentosum patients 1) Characterization . . . . 2) Evidence that DNA excision repair protects cells from the potentially harmful effects of carcinogens . . . . . d) A second form of xeroderm pigmentosum disease, XP variants e) Bloom's syndrome . . . . . iii Page vii viii xi 17 18 19 22 24 24 25 25 25 29 32 33 Chapter Evidence for other mechanisms of carcinogenesis . . . . a) Teratocarcinomas b) Studies with plastic film . c) Arguments against Concluding Remarks . . . . . MATERIALS AND METHODS . . . . Chemicals . . . . . . . . . Preparation of Solutions . Phosphate buffered saline 6-thioguanine . . . . . Antibiotics . . . . . . Versene trypsin . . . . Standard saline citrate Kirby phenol . . . . Trichloroacetic acid Preparation of Medium Cell Cultures . . Storage of Cells . Culture Medium . . Spent Medium . . Selection Medium Testing of Serum . Testing for Mycoplasma Contaminatio the XP story n Benzo(a)pyrene (BP) Metabolism Assay Protein Assay . . . . . . . In Situ Cytotoxicity Assay . Replating Cytotoxicity Assay Assay of Biological Recovery from Potentially Cytotoxic Effects Cell—mediated Cytotoxicity Assay . Replating Mutagenicity Assay Assay of Biological Recovery from Potentially Mutagenic Effects Reconstruction Experiments . Isolation of DNA from Human Cells Purification of DNA Samples HPLC Analysis of DNA Adducts Studies with Calf Thymus DNA iv Page Chapter RESULTS 0 O O O O O O O O O O O O O O O O 0 Studies with AFB . . . . . . . . . . . . IdentificatioA of a human cell strain capable of metabolizing aflatoxin B The effect of cell number and exposure time on the extent of metabolism . . Cell-mediated cytotoxicity studies with AFB . . . . . . . . . . . . . The effect Ef varying the number of metabolizing cells on the cytotoxic response in target cells . . . . . . Optimal ratios of metabolizing cells to target cells . . . . . . . . . . AFB -induced cytotoxicity in Hs703T éells as targets . . . . . . . . . . Comparison of the cytotoxic response to AFB1 in three different target cells Use of AFBl-Cl as a model for AFB in studying the biological effeéts of AFB in human cells in culture . Studies with AFB -Cl . . . . . . . . . . Comparative sénsigivity to the cytotoxic effects of AFBl-Cl . . The relationship between level 8f DNA binding and survival . . . . . . . . Rate of recovery of confluent cultures from the potential cytotoxicity induced by AFB -Cl lesion(s) . . . Comparative sensifivigy to the mutagenic effects of AFBl-Cl . . Rate of recovery of confluent cgltures from the potentially mutagenic effects of AFB -C1 . . . . . . . . Rate of loss of cévalgntly bound AFB -Cl residues from cellular DNA Rate o 1055 of covalently bound AFB -Cl residues from calf thyAus DNA . . . . . . . . . . HPLC analysis of covalent y bound AFB -C12-induced DNA adducts in Aalf thymus DNA . . . . . . . . . HPLC analysis of covalently b und AFB -Cl -induced DNA adducts in Aumag cellular DNA . . . . . . . Page 71 71 71 74 78 82 83 85 85 9O 91 91 95 97 99 102 102 104 107 111 Chapter Page DISCUSSION . . . . . . . . . . . . . . . . . . . . 114 Cell-mediated activation of AFB1 . . . . . . 114 AFBl-induced DNA lesions are UV—like" . . . . . . . . . . . . . . . . 116 AFB -C12-induced DNA lesions Are UV-like" . . . . . . . . . . . . . . 117 Relationship between time for repair and the ultimate biological effects of the carcinogen . . . . . . . . . . . . 118 Lack of correlation between rate of loss of AFBl-Cl DNA adducts and recovery of celIs from the potential biological effects . . . . . . . . . . . . 120 Loss of the major DNA adduct induced by AFBl-Cl is non-enzymic . . . . . . . . 124 Isolation of Ehe major DNA adducts induced by AFBl-Cl . . . . . . . . . . . 126 Possible mechanisms of carcinogenesis by AFB and AFBl-Cl2 . . . . . . . . . . . 128 ConclusioA . . . . . . . . . . . . . . . . . 133 REFERENCES . . . . . . . . . . . . . . . . . . . . 134 vi LIST OF TABLES Table Page 1. HPLC gradients used . . . . . . . . . . . . . 69 2. The level of benzo(a)pyrene metabolism in several human cell strains . . . . . . . . . . . . . . . . 72 3. Level of BP metabolism in PC-3 cells following 24 and 48 hour exposures to BP . . . . . . . . . . . . . . 77 4. The induction of 6-thioguanine- resistant mutants by AFBl-Cl2 in NF and XP cells . . . . . . . . . . . . . 101 5. Levels of DNA binding in AFB -C12— treated human cells over time . . . . . . . 106 vii LI ST OF FIGURES Figure Page 1. Structures of AFB1 and related compounds . . . . . . . . . . . . . . . . . 5 2. Structures of AFBl-2,3-oxide and 2,3-dihydro-2,3-dihydroxy AFB 1 O 0 O O O O O 8 3. Structures of the major AFB DNA adducts and their hydrolyéis products . . . l3 4. Structures of AFB -dichloride and AFBl—2,3-oxide . . . . . . . . . . . . . . . 20 5. Serum test: growth of cells in thioguanine after seven days . . . . . . . . 47 6. Serum test: colony—forming ability in thioguanine . . . . . . . . . . . . . . 48 7. Serum test: growth of cells to high density . . . . . . . . . . . . . . . . . . 50 8. Determination of protein content in cellular samples . . . . . . . . . . . . 55 9. DNA absorption spectrum . . . . . . . . . . . 67 10. The relationship between cell number and protein content in PC-3 and normal cells . . . . . . . . . . . . . . . . 75 11. The relationship between number of cells assayed and the level of water soluble metabolites produced . . . . . . . . 76 viii Figure Page 12. Cell-mediated cytotoxicity of AFB in XP cells using two concentrafions of H3835T cells to activate AFBl . . . . . . 79 13. Cell-mediated cytotoxicity of AFB in XP cells using two concentrafions of Hs703T cells to activate AFBl . . . . . . 14. Comparison of the levels of cytotoxicity induced by AFB in XP cells in the presence of qual numbers of H5835T and Hs703T cells . . . . . . . . . . . . . . 81 80 15. The effect of varying the ratio of Hs703T cells to XP cells on the cytotoxicity induced by AFB1 in XP cells . . . . . . . . 84 16. Cytotoxicity of AFB in Hs703T cells following 24 and 48 hour exposure times 0 O O O O O O O O O O O I O O O O O O 86 17. Comparison of the levels of cytotoxicity induced by AFB in NF, XP and Hs703T target tells . . . . . . . . . . . . 87 18. Cytotoxicity of AFB -C12 in NF, XP, and Hs703T cells . . . . . . . . . . . . . . 89 19. Cytotoxicity of AFBl-Clz in NF and XP cells . . . . . . . . . . . . . . . . . . 92 20. Cytotoxicity of AFBl-Cl in three different XP strains 3nd normal cells . . . 94 21. Relationship between levels of cytotoxicity and DNA binding induced by AFBl-Cl2 in normal and XP cells . . . . . . . . . . . . . . . . 96 22. Recovery of confluent cells from the potential cytotoxicity of AFBl-Cl2 . . . . . 98 23. Induction of 6-thioguanine-resistant mutants by AFBl-Cl2 in normal and XP cells . . . . . . . . . . . . . . . . 100 ix Figure Page 24. Recovery of normal cells from the potential mutagenicity of AFBl-Cl 103 2..... 25. Loss of DNA-bound AFB —Cl residues from NF and XP cell; ovgr time held in confluence . . . . . . . . . . . . . . . 105 26. Loss of DNA-bound AFB -Cl2 residues from calf thymus DNA over time . . . . . . . 108 27. HPLC profiles of AFBl-Cl DNA adducts in calf thymus DNA on days 0, 2, and 8 after treatment . . . . . . . . . . . . . . 110 28. HPLC profile of AFBl-Cl DNA adducts in human cellular DNA on day 0 . . . . . . . 112 29. HPLC profiles of AFB -Cl DNA adducts in human cellular DNA following a 60 hour cesium chloride gradient . . . . . . . 129 AP BS BP BPDE cpm Ci DNA LIST OF ABBREVIATIONS aflatoxin B1 aflatoxin Bl-dichloride aflatoxin Gl aryl hydrocarbon hydroxylase apurinic/apyrimidinic Bloom's syndrome benzo(a)pyrene 7,8-diol-9,10-epoxide of benzo(a)pyrene centigrade counts per minute curie deoxyribonucleic acid disintegrations per minute gram high pressure liquid chromatography kilogram microgram : microliter : micromolar xi mg ml mm mM nm nM NF ppb RNA TG uv XP milligram milliliter millimeter millimolar molar nanometer nanomolar normal normal diploid human fibroblasts parts per billion ribonucleic acid thioguanine ultraviolet xeroderma pigmentosum xii INTRODUCTION It has been well documented that aflatoxins, and in particular aflatoxin B (AFBl), are very powerful l carcinogens in many mammalian species by several different routes of administration. For example, aflatoxin B1 induced liver carcinomas in male Fischer rats at dietary levels of l-lS ppb (Wogan gt 31.,1964). Although the liver is the major target organ for aflatoxin carcinogene- sis, other organs are susceptible, including the kidneys, stomach and colon,depending on the route of exposure (Wogan and Newberne,l967). Epidemiological studies show a positive correlation between the level of dietary intake of AFB1 and the incidence of human liver cancer in Asia and Africa (Alpert e; 31.,1974), suggesting that aflatoxins are hepatocarcinogens in man. The somatic cell mutation theory of the origin of cancer argues that carcinogens are mutagens, and therefore, that DNA is the target macromolecule for the majority of chemical carcinogens. A good deal of evidence supports this theory. It was therefore of interest to study the mutagenic potency of aflatoxin to determine whether it parallels its 2 carcinogenic potency. One modulating factor in mutagenesis and carcinogenesis is DNA repair. Cells are known to be capable of enzymically removing DNA-bound carcinogen residues, and it has been shown in human cells,that if DNA repair is carried out before the DNA is replicated, it acts to reduce the carcinogenic and mutagenic effects of these carcinogens (Maher gt; a_l_., 1979; Heflich gt 31.,1980; Yang gt a_1.,1980). We1therefore, chose to study the effect of DNA repair on the biological effects of aflatoxin B1- dichloride (AFBl-Clz), a reactive derivative of AFB1, in repair-proficient and repair-deficient human cells in culture. We carried out comparative studies on the loss of aflatoxin Bl-dichloride-induced adducts from the DNA of normal human cells and of repair-deficient fibroblasts derived from xeroderma pigmentosum patients and the corresponding recovery of these cells from the potentially mutagenic and/or cytotoxic effects of this agent, in an attempt to provide insight into the nature of the potentially mutagenic and cytotoxic lesion(s). Our purpose was to provide some insight into the mechanism(s) by which aflatoxin induces mutations and cytotoxicity in cells in culture, in order, possibly, to gain a better understanding of the mechanismIs) by which aflatoxin induces cancer _i__n vivo. LITERATURE REVIEW Aflatoxins Aflatoxins are a group of naturally occuring carci- nogens, produced as metabolic by—products of the mold Aspergillus flavus, which grows well on grains and nuts stored in damp conditions. They were first recognized as a problem contaminant in foodstuffs in 1960, when over 100,000 turkeys died due to contamination of their food (Blount, 1961). At the same time, toxicity was seen in cattle(Loosmore,1961a),pugs (Loosmore,l96lb), and ducklings (Asplin,1961). The cause of the toxicity was traced to the common factor of Brazilian groundnut meal in the animal food. The toxic factor was extracted from the groundnut meal and discovered to be composed of metabolic by-products of the mold Aspergillus flavus (Sargeant,1961). From this mixture of products, four major fractions were isolated. They were named aflatoxin 81' B2, G1 and G based on whether they fluoresced green or blue in 2! ultraviolet light, and on their relative chromatographic mobilities (Nesbitt,l962). The relative toxicities of these 4 four components vary greatly. In human embryonic cells in 1 1 > G2 > B (Zuckerman,1968). This order can differ from one species to culture, the order of toxicity is B > G another, but in all species studied, AFBl (Figure l) is by far the most toxic of all (Garner and Martin 1979). Metabolism of AFB 1 Because it has been determined to be the most toxic, AFB1 has been the most widely studied of the aflatoxins. AFBl was first shown to require metabolic activation in 1971 when Garner gt. El. showed that AFB1 was toxic in Salmonella typhimurium TA 1530 and TA. 1531 only in the presence of rat liver microsomes and a NADPH-generat- ing system. In the absence of the activation system. no toxicity was observed. Garner, Miller, and Miller reported in 1972 that ,in addition to AFB AFGl and sterigma- ll tocystin, a product of Aspergillus versicolor, were also toxic to S; typhimurium TA 1530 and TA 1531 in the presence of an activation system, while several other metabolites were not. They also examined. the extent of AFBl-binding to commercial RNA in the presence of a rat liver microsome system. They suggested that the toxic metabolite of AFBl was the 2,3-oxide (Figure 1) based on their data that AFBl, AFGl, and sterigmatocystin were Aflotoxin B1 AflotoxindB1-2,3’ ox: e OCHS O OCH3 Aflatoxin G1 Sterigmotocystin Figure 1. Structures of AFB1 and related compounds 6 the three most active aflatoxin compounds in the microsome- mediated system, and based on structural data showing that all three had a 2,3 double bond in common, while the distal portion of these molecules differed. They cited as evi- dence the polycyclic aromatic hydrocarbons which, when activated by the mixed-function oxygenases of liver micro- somes produce epoxides, are highly reactive, toxic, and mutagenic in manunalian cells (Grover g1; 31., 1971; Huberman gt $1., 1971). They also observed that the factor which was toxic to the bacteria and the level of nucleic acid binding i_n_ Eli-£9 were dependent upon the metabolizing system. They, therefore, suggested that the toxic derivative and the derivative that reacted with nucleic acids were one and the same. Since AFB AFG 1' l’ and sterigmatocystin were known to be the most carcinogenic of the aflatoxin derivatives studied, they postulated that the toxic metabolites of these three compounds, which they proposed as the 2,3-oxides, might also be involved in the induction of liver tumors 13 vivo. All attempts to isolate the toxic derivative of AFB1 have been unsuccessful. The first biochemical evidence in support of the proposal that the 2,3-oxide was the toxic derivative was obtained by Swenson 23 a_1. in 1973. Swenson isolated an hydrolysis product of an AFB -RNA 1 adduct formed i9 vitro in the presence of rat liver microsomes. Mild acid hydrolysis of the adduct yielded a 7 product indistinguishable from synthetic 2,3-dihydro-2,3- dihydroxy-AFBl (Figure 2). These data provided evidence that the reactive species of AFB1 was the 2,3-oxide. This epoxide would be expected to have a strongly electrophilic carbon 2, at which to bind RNA or DNA, and would yield the 2,3-dihydro-2,3-dihydroxy-AFB1 on mild acid hydrolysis. In 1975 when Swenson gt gt. isolated the hydrolysis products of AFBl-RNA adducts isolated from rat liver i_n ttgg the major product obtained was identical to the 2,3-dihydro-2,3-dihydroxy-AFB1 obtained _i_r_1 _vi_t1_'2. This suggested that AFB1-2,3-oxide was responsible for a major portion of covalent binding of AFB1 to nucleic acids in rat liver 1 vivo, and that the 2,3-oxide was probably the ultimate carcinogenic metabolite of AFBl. Binding of this metabolite to rat liver DNA t2 vivo has since been correlated with the induction of AFBl-induced liver tumors (Swenson gt gt.,l977). Toxicity of AFB1 AFB1 has been shown to be toxic in a wide range of animal species including rats, pigs, ducklings, turkeys, chickens, and trout (Loosmore.1961a,b; Asplin,196l). In all cases the primary target organ is the liver (Garner and Martinpl979). Most of the damage is exhibited as liver cell necrosis. There are a wide range of factors influencing toxicity which include diet, age, and sex of the animal 0 OCH3 Aflotoxin 8172,3-oxide O OCH3 2,3-dihydro- 2,3- dihydroxy aflatoxin 81 Figure 2. Structures of AFB -2,3-oxide and 2,3-dihydro-2,3-dihydroxy AFBl 9 (Wogan,l968). Since AFBl requires metabolic activation in order to exert any biological effects on cells or animals, the level of metabolizing enzymes present in the tissues of different species largely determines the extent to which these species will be affected by AFB (Newberne,l973). l The liver has the highest concentration of the mixed function oxidase systems needed to metabolically activate AFBl to the reactive 2,3-oxide. Because it is the primary purpose of these enzyme systems to metabolically convert compounds that cannot be utilized by the body into forms that can be excreted from the body (usually water soluble), the liver is considered to be the primary site for "detoxi- fication." This is accomplished primarily by converting compounds into more polar (water soluble) and less toxic forms. Aflatoxins are highly toxic. Because they are lipid soluble,they partition into the endoplasmic reticulum of cells, which is where they are metabolized. During the detoxification process, compounds may be converted into forms which are actually more toxic than the parent com- pound. This is the process known as activation, and is essentially what happens in the case of AFB Thus, 10 besides being the major site for detoxification, the liver is often the target organ for activated species generated by this system. At the cellular level AFBl may induce toxicity by inhibiting DNA, RNA and protein synthesis (Wogan,l968). The 10 inhibition of protein synthesis is a result of AFB1 bind— ing at the polysomes (Garvican gt_ gl.,1973). AFBl has been shown to greatly inhibit hepatic RNA synthesis which may in part be responsible for the hepatic damage observed (Akao gt_ gt.,197l). AFB1 also inhibits rat liver mito- chondrial electron transport t2_ ZiEEQ (Doherty gt 91,4973) . In the presence of an activation system, AFB1 is toxic to several strains of Salmonella typhimurium and Escherichia coli (Garner and wright, 1973). AFBl has induced toxicity in V79 Chinese hamster cells (Langenbach _t gt.,l978) and in C3H10T} mouse embryo cells (Krahn and Heidelberger ,1977) when an activating system was present. Cultured epithelioid human tumor cells from (A549) lung are capable of metabolizing AFBl, and these cells exhibited sensitivity to the toxicity of AFB1 as measured by colony forming ability (Wang and CeruttL 1979). Carcinogenicity of AFB1 AFB1 is one of the most potent hepatocarcinogens known. It was first shown to be carcinogenic in 1961 when Lancaster reported that rats fed on peanut meal contaminat- ed with aflatoxins developed liver tumors. A linear dose response relationship was later observed by Newberne in 1965 when hepatomas were induced in rats fed varying doses ll of AFB1 in their diets. Liver tumors have been induced by AFBl in a series of laboratory animals, including mice (Weider t g" 1968), ducks (Adamson, 1973), and monkeys (Carnaghan _t 512.,1965). It induces liver tumors in the rat or trout at continuous dietary levels of l to 15 ppb (pg/kg) and in rats at a single sublethal dose of 7.65 mg/kg body weight (Carnaghan gt gt,1967). As indicated above, epidemiological evidence suggests that AFBl may be responsible for the high incidence of liver cancer in localized human populations. A typical report is that of Alpert gt gt" (1971), who showed a positive correlation between the estimated aflatoxin intake and the incidence of primary liver cancer among Ugandans. Mutagenicity of AFB1 The induction of mutations by AFBl has been studied in many organisms. In the presence of rat liver microsome fractions and NADPH, AFB is mutagenic in Salmonella l typhimurium (Ames gt gt.,1973). Mutations are induced in vegetative cultures of Neurogpora crassa in the absence of a metabolizing system, but, in resting conidia of N. crassa, mutations are induced only in the presence of a metabolizing system (Ongpl97l), suggesting that the vege- tative cultures are capable of metabolism of AFB1 while conidia are not. In Drosophila, AFBl was found to be mutagenic in males treated by injection (Lamb and Lilly, 12 1971) . Krahn and Heidelberger (1977) have shown that, in the presence of a rat liver activation system, AFBl induced 6-thioguanine-resistant mutants in V79 cells. Langenbach gt gt. (1978), have used a rat liver cell-mediated system to activate AFBl. In this system, AFBl induced ouabain- resistant mutants in V79 Chinese hamster cells. In both these systems, AFB induced lower frequencies of muta- 1 tions than did benzo(a)pyrene for the same survival level. AFB -induced DNA adducts 1 In 1977 Essigman _t gt first isolated the major DNA adduct formed by AFB1 12 ytttg. In the presence of rat liver microsomes, AFBl bound covalently to calf thymus DNA, and on mild acid hydrolysis, yielded a guanine adduct. By means of high pressure liquid chromatography (HPLC), this adduct was isolated and was then characterized by spectral and chemical means. The major adduct was identi- fied as 2,3-dihydro-2-(N7-guanyl)-3-hydroxyaflatoxin B1 (AFBl-N-7-guanine) (Figure 3). This DNA. adduct comprised 90% of the total carcinogen bound to DNA, and was formed by reaction of the C-2 atom of AFB with the N—7 atom of 1 guanine. At that time, Essigman noted that this N-7 adduct resembled that of the 7-alkyl-guanines in charge and propos- ed that apurinic sites might be formed spontaneously in 13 AFB1 -N-7-guonine (ring-opened form) Figure 3. Structures of the major AFB1 DNA adducts and their hydrolysis products 14 AFBl-modified DNA, as is seen in alkylated DNA. Later studies by Martin and Garner (1977) revealed that when calf thymus DNA was reacted with AFB1 in the presence of rat liver microsomes, there were four DNA adducts formed. All of these adducts co-chromatographed with adducts formed with poly dG in the same system. Of the four guanine peaks, there were two major products, both of which were formed through N-7 guanine substitution. D'Andrea and Haseltine (1978) showed that when AFB was reacted with §_._ coli 1 DNA in the presence of rat liver microsomes, it resulted in alkali-labile sites in the DNA. This data lent support to the earlier suggestion by Essigman that the N-7 adduct might spontaneously depurinate as a result of labilization of the N-glycosylic linkage at the N-7 atom of guanine. The report by D'Andrea and Haseltine also revealed the forma- tion of an adenine adduct in the system they used. When AFBl-DNA adducts were isolated by Autrup gt g_J_.. (1979) from cultured human bronchus exposed to AFBl, four distinct products were seen. The two major adducts observed were identical to the two major adducts previously reported to be formed 12 ELI-fl- In the human bronchus system, the two major adducts comprised 70% of the total bound carcinogen, and the ratio of the two products to one another was approximately 1. The second major guanine peak was tentatively identified as 2,3-dihydro-2-(N5-formyl- 2',5',6'-triamino-4'-oxo-N5-pyrimidyl)-3-hydroxyaf1atoxin 15 B1 (Figure 3). This adduct is formed from the N-7 guanine adduct as a result of fission of the imidazole ring of guanine. Therefore, it is a ring opened form of the N—7 guanine adduct. The adduct studies of Autrup gt a_1. (1979) in cultured human bronchus revealed that the AFBl- DNA adducts formed it: ELI-E2 were qualitatively similar, but differed quantitatively from the AFB -DNA adducts 1 found in the £2 vitro systems previously studied. Wang and Cerutti (1979) compared the loss of AFB1- DNA adducts from the DNA of epithelioid human tumor cells (A549). The kinetics of loss of AFB:L adducts was monitor- ed from the DNA of intact cells and from free DNA which had been isolated from A549 cells immediately after treatment and incubated in physiological conditions over five days. Their results showed a greater loss of AFBl-DNA adducts from whole cells than from DNA i_n m. They also showed that the percentage of the total adducts which the ring-opened N-7 guanine product comprised increased with time. They suggested that because of its persistence, this adduct might play an important role in the induction of mutations and cancers by AFBl. Hertzog g 9;. (1980) studied the kinetics of loss of DNA-bound AFBl in rat liver and _it 3729.3. Following intraperitoneal adminis- tration of AFB1 to male rats, the AFBl-DNA adducts were monitored over 48 hours in rat liver DNA i_n vivo, and 16 from rat liver DNA _i_g vitro, isolated from rat liver two hours after treatment of the rats with AFB Their 1’ studies showed that the half-life of the N—7 guanine adduct, both i_n Egg and i_n_ y_it_ro_, was approximately 24 hours. Based on these data, the authors concluded that a major portion of AFBl-DNA adducts lost i_n gtyg are lost non-enzymically. These findings are in contrast to those reported earlier by Wang and Cerutti (1979). The differences in these reports is probably attributable to differences in the pH of the E m incubation mix- tures. The studies of Hertzog gt gt. (1980) also showed that the ring-opened guanine derivative of AFB1 was a more persistent lesion than the N-7 guanine adduct i_n vivo and i_g vitro. A recent report by Groopman gt l. (1981) on in vitro reactions of AFB with DNA — -— —— 1 support the suggestion put forth by Hertzog, that a major portion of AFBl-DNA adducts lost 1 vivo are lost by chemical rather than by enzymic means. In this study AFBl was reacted with calf thymus DNA in the presence of a rat liver activation system. The loss of the major N-7 guanine adduct from DNA was monitored for up to 48 hours under different incubation conditions. They found that the kinetics of loss of the N-7 guanine adduct varied, depend- ing on the pH of the mixture and the level of DNA modifi- cation. However, under all conditions studied, spontaneous depurination was observed. 17 Evidence of repair of the cytotoxic and mutagenic DNA lesions induced by AFBl There is strong evidence for the involvement of DNA repair enzymes in removing AFBl-induced DNA damage in several systems. AFB1 induced toxicity in gt ggtt in the presence of a rat liver fraction (Garner and Wright, 1973). However, the degree of toxicity varied depending on the genetic markers carried. Mutants deficient in the excision repair pathway (uvr-) were more sensitive to AFBl than wild type strains, but were less sensitive than mutants deficient in recombination repair (recA). The double mutant (uvr-, recA) was the most sensitive. This suggests that both these repair pathways are involved in repairing AFBl-induced cytotoxic damage in & g_o_l_i_, but that excision repair appears to be more necessary than recombina- tion repair in surviving this type of damage. Similarly, in the presence of a rat liver system, AFBl has been shown to induce toxicity in strains of Salmonella typhimurium which have a deletion in the genes coding for repair of UV— light-induced damage (Garner gt g-I1972). In strains of Salmonella typhimurium without this deletion, no toxici- ty was observed. When AFB1 was activated by rat liver microsomes in the presence of human cells in culture, repair synthesis was induced in normal cells, but not in repair-deficient XP cells (Stich and Laishes, 1975; Sarasin g gt,1977). In the presence of rat liver microsomes, l8 AFB1 has also been shown to induce greater frequencies of mutations in strains of Salmonella typhimurium defi- cient in excision repair than in wild type strains (Garner and Wright, 1973). All this evidence implicates the involve- ment of DNA lesions in AFBl-induced toxicity and muta- tions. Possible mechanisms of mutagenesis by AFB1 There have been several mechanisms proposed to explain how AFBl-induced DNA lesions might induce mutations . One of these involves the major adduct which is formed as a 7of result of binding between the C-2 of AFB1 and N guanine. This proposed mechanism suggests that the 3- hydroxy group of AFB1 can hydrogen bond with the oxygen at C-6 of guanine and disrupt normal base pairing, resulting in a point mutation. However, when AFBl-Cl2 binds DNA, the 3-chloro group is retained and thus could not undergo similar hydrogen bonding. Since AFBl-Cl2 is as potent a mutagen as AFBl when normalized for amount bound to DNA, this mechanism seems unlikely (Swenson gt g_]_.., 1977). Another possible mechanism is that the large group introduced into DNA by AFB:L causes steric hindrance of the polymerase. This might result in a failure to replicate a section of DNA which would lead to a deletion mutation. It must be remembered that minor unidentified DNA adducts, rather than the major adducts, may be responsible 19 for producing the biological consequences of AFBl. For alkylating agents, it appears that minor adducts (for example, 0-6 alkyl guanine) are important in the induction of mutations in mammalian cells, rather than more predomi- nant lesions (e.g. ethylated phosphotriesters for ENU and 0-6 methyl guanine for MNU (Newbold e_t_:_ g__l.,1980; Peterson t alul979). AFBl-Clzas a model compound for the ultimate metabolite AFBl-2,3-oxide As stated above, attempts to isolate the reactive meta- bolite, AFBl-2,3-oxide, or to synthesize it ‘were unsuc- cessful, probably because of its highly reactive nature -Cl (Swenson gt_ gl_.,1973, 1975, 1977). AFB was 1 2 therefore synthesized by Swenson gt gt. (1977) as a model compound for the 2,3-oxide (Figure 4). They showed -Cl that AFB 2 is similar in chemical properties to l AFB1-2,3-oxide. It has an electrophilic carbon 2 and binds covalently to DNA and RNA at this position, as does AFBl-2,3-oxide. It reacts predominantly with guanine. Its hydrolysis products react only to a very small extent with nucleic acids. AFBl-Cl2 is a direct -acting mutagen and carcino- gen. It causes tumors at the site of injection in rats and induces skin papillomas in mice, whereas AFBl does not 2O O 0CH3 Aflotoxin 81-2,3-oxide Aflo'roxin B1— dichloride Figure 4. Structures of AFBI'dichloride and AFBl—2,3-oxide 21 (Swenson gt gt” 1977). It is mutagenic in Salmonella typhimurium, and when the mutagenicities of the dichlor- ide and AFBl were compared in Drosgphila melanogaster (Fahmy g a_1., 1978), it was observed that the two com- pounds induced the same mutational spectra, but with higher activity observed for the dichloride. This would be expect- ed since the reactive species would be generated more readily from the dichloride than from AFBl, which must first be metabolized. These studies did not include compari- sons of levels of DNA binding for the two carcinogens. We would expect that the greater mutagenic activity observed for AFB -Cl was the result of a greater level of DNA 1 2 binding for this compound in the target cells. The general toxicities induced in Drosophila by AFB1 and AFBl-Cl2 were virtually identical. That is, for males treated by injection, the LDso levels were reached at the same concentration for both chemicals. However, when the cytotoxicity of AFB1 and the dichloride was assayed in cells at various stages in the germ cell line, there were marked differences observed. In all cases, AFB -C12 induced greater cytotoxicity in these cells at 1 concentrations equal to AFBl. The difference in the cytotoxicities induced was the greatest for the mature sperm. The difference was less for metabolically active cells in the earlier stages of spermatogenesis. This suggests that the lower levels of toxicity by AFB1 seen 22 in these cells is the result of the extent of metabolic activation taking place. Somatic Cell Mutation Theory of Cancer Our interest in aflatoxin was based on the evidence for its nature as a powerful carcinogen. The somatic cell mutation theory of the origin of cancer argues that cancers are caused as a result of somatic cell mutations. We, therefore, chose to study the induction of mutations by aflatoxin Bl-dichloride in human cells in culture in an attempt to further our understanding of the mechanism(s) by which chemicals induce mutations, and possibly cancers, in human cells. The somatic cell mutation theory of the origin of cancer was first proposed by Boveri in 1914 in an attempt to explain the observation that the cancer cell phenotype appeared to be inherited, that is, daughter cells retained the transformed phenotype of their parent cancer cells. At the time that this theory was proposed, DNA was not yet understood to be the genetic material. In its present form, the somatic cell mutation theory states that cancer is a result of mutational events in somatic cells and, there- fore, that DNA is the target macromolecule for most chemical carcinogens and carcinogenic radiations. This 23 theory was unsupported for many years by a lack of corre- lation between carcinogenesis and mutagenesis. As a result, there were other theories put forth to explain the mecha- nism of chemical carcinogenesis. For example, Miller 'gt gt-(1963) proposed that protein was the target macromole- cule for chemical carcinogens when they found that some polycyclic aromatic hydrocarbons showed preferential binding tx> a particular protein fraction. A similar hypo- thesis was held by Heidelberger (1964). In an attempt to explain the heritable nature of the change resulting from loss of a protein from a cell, Pitot and Heidelberger (1963) proposed that reaction of a carcinogen with a growth-controlling protein resulted in inactivation of this protein, and they employed the Jacob-Monod model of gene repression to explain how a protein deletion could result in a heritable change in the cell. However, when Brookes and Lawley (1964) demonstrated a positive correlation between binding of polycyclic aromatic hydrocarbons to mouse skin DNA and their carcinogenic potency, it revived interest in the somatic cell mutation theory, with particu- lar emphasis on DNA as the target macromolecule. Further interest in this theory was generated as a result of the work of Miller and Miller (1968) who demonstrated the requirement for many chemical carcinogens to be metabolical- ly activated before they can exert their biological effects. Since the mutation assays used until this time did 24 not include means to metabolically activate the carcino- gens, the failure to find a correlation between mutagenici— ty and carcinogenicity may be related to metabolic acti- vation. This realization spurred development of activation systems suitable for use with mutational assays. Since this development correlations between mutagenicity and carcino- genicity as high as 90% have been found (McCann gt gt” 1975) . Evidence in support of the somatic cell mutation theory a) The majority of carcinogens tested have proved to be mutagens If cancers arise as a result of mutational events, then carcinogens should prove to be mutagens in one or more test system. Furthermore, there should be a correlation be- tween the carcinogenic potency and the mutagenic potency of carcinogens. In support of this, approximately 90% of the chemical carcinogens tested to date in the Ames' Salmonella assay have been shown to be mutagens (McCann gt gin 1975), although strong carcinogens have not always been found to be strong mutagens and strong mutagens are not always strong carcinogens. This is not surprising, since mutagenesis test systems utilize a different activation system than carcinogenesis test systems, and therefore, differences could be expected. Furthermore, carcinogenesis 25 tests involve animal studies, whereas mutagenesis tests usually involve bacteria, fungi or cell cultures. Consider- ing the enormous biological differences between these test systems, it is not surprising to find that the correlation between mutagenicity and carcinogenicity is not perfect. b) Clonal nature of tumors There is considerable evidence that human tumors arise as a clone from a single cell of origin. This evidence includes cytogenetic studies which show the same abnormal karyotype for all cells derived from a single tumor (Sandberg and Hossfield,l970) and studies on the level of glucose-6-phosphate dehydrogenase, which showed that cells derived from the same tumor all had the same levels of this enzyme (Linder and Gartler, 1965). c) DNA excision repair—deficient xeroderma pigmentosum patients 1. Characterization Perhaps the best evidence in support of the somatic cell mutation theory of cancer is the clinical and experi- mental data available on patients with xeroderma pigmen- tosum. 26 Xeroderma pigmentosum (XP) is a human syndrome, inherited as an autosomal recessive trait, which is characterized primarily by an increased sensitivity of the skin to sunlight, causing patients to develop abnormal pigmentation and skin tumors in sun-exposed areas at a very young age (Robbins gt gt, 1974). Mental retardation, areflexia, and other neurological disorders can also be associated with the disease (Robbins g g_l_., 1974). Affected individuals develop sunlight-induced skin cancer with a prevalence of approximately 100%-compared to an annual rate in the 0.8. of l to 2 per 103 for the general population (Setlow,l978). This frequency has been markedly reduced in patients protected from sunlight from an early age (Robbins gt £41974). In 1968 Cleaver observed that, compared to skin fibro— blasts derived from normal persons, fibroblasts derived from XP patients showed decreased ability to carry out excision repair of ultraviolet (UV) radiation-induced DNA lesions. It is now recognized that there are at least two different modes of excision repair operating in mammalian cells. One is designated nucleotide excision repair because the damaged base residue is excised within an oligonucleo— tide by two sequential single strand nicks. The first nick is introduced by an endonuclease and the second by an exonuclease. The correct nucleotides are inserted by a DNA polymerase, and the ends are sealed by a polynucleotide 27 ligase. (Grossman,l979). The second form of excision repair is called base excision repair. In this mode a modified base is released as a free base by a DNA glycosylase or leaves spontaneously, resulting in an apurinic/apyrimidinic (AP) site. The AP site is then removed in the same way as described above for nucleotide excision repair, with an AP-specific endonuclease and exonuclease responsible for cleaving the damaged oligonucleotide. A polymerase and ligase act to synthesize and join the new DNA segment as described. There is ewidence of a third method of removing damage from DNA, but it cannot be considered excision repair because: the DNA remains intact. This mode of repair results in the insertion ofa free purine residue (Chetsanga and Lindahl, 1979). Until 1980, excision repair of UV- induced DNA damage had been considered to be a classical example of nucleotide excision. However, Haseltine gt gt. (1980) found a dimer-—specific endonuclease in M. luteus with two enzymic activities. A pyrimidine dimes specific DNA glycosylase cleaves the glycosylic bond between the one thymine base and its sugar,leaving an AP site, but the thymine dimer is still intact. An AP endo- nuclease cleaves the DNA adjacent to the AP site, and exo- nuclease cleaves the oligonucleotide containing the AP site from the DNA. A. polymerase» and ligase act in the same manner as described for nucleotide excision. When Cleaver tested the XP cells for their ability to 28 carry on excision repair following exposure to UV light, he measured the extent of repair replication using 3H—BUdR. By use of buoyant density gradients, density-labeled, newly synthesized DNA was separated from DNA in which repair had resulted in insertion of 3H-BUdR in small patches. The XP cells proved to be very deficient in this process. These results suggested that the deficiency of XP cells in this DNA repair pathway was, at least in part, responsible for their increased incidence of sunlight-induced skin cancers. Cleaver further suggested that the mechanism by which this DNA repair defect resulted in tumors might be by causing enhanced frequencies of somatic mutations as a result of unrepaired DNA damage. XP strains that are deficient in excision repair have been subdivided into groups based on complementation analysis by means of cell fusion (Kraemer gt gt,,l975). If, upon cell fusion of two strains, the heterokaryon exhibited normal levels of repair synthesis, the cells were classified as being in separate complementation groups. If two strains showed low levels of repair synthesis upon fusion, they were categorized as being in the same complementation group. Seven complementation groups have been identified to date (Kraemer gt g;.,l975; Takebe, 1978). Therefore, mutations in at least seven different loci can give rise to the XP phenotype. It is not known whether these seven loci represent seven distinct cistrons. There 29 are different rates of excision repair observed among the different complementation groups. Although there are variations in the levels of exci— sion repair seen in the seven complementation groups of XP cells, all excision repair deficient XP strains appear to have a defect in the incision step. The introduction of T4 endonuclease into these XP cells restores normal repair replication (Tanaka gt _l_.. 1977). Cell extracts from groups A, C, and D are capable of excising dimers from UV- irradiated naked DNA (Mortelmans gt a_ly 1976), but group A cells are unable to remove dimers from UV-damaged chroma- tin, suggesting that the defect may be in the recognition of the damage, or in a cofactor involved in the repair of the chromatin. 2. Evidence that DNA excision repairyprotects cells from the potentially harmful effects of carcinogens A large variety of physical and chemical agents induce DNA damage in human cells. The majority of agents studied appear to be handled by base excision or nucleotide exci- sion. Regan and Setlow (1974) have divided a series of carcinogens into two groups, according to the size of the excised oligonucleotides.One group causes large oligonucleo- tides to be excised resulting in long or "UV-like" repair. 30 This group included the bulky chemicals N-acetoxy-Z-acetyl- aminofluorene and 4-nitroquinoline—oxide. The other group of agents induces short or "X-ray-like" repair. There is evidence that excision repair in human cells can reduce the potentially cytotoxic and mutagenic effects of carcinogenic agents, including UV-light and chemicals. When excision repair-—proficient normal human fibroblasts and excision repair -deficient fibroblasts derived from several different XP patients were exposed to equal doses of UV light, the XP cells exhibited a higher frequency of induced mutations than did the normal cells (Maher gt gt,1976). More importantly, when a series of XP- derived skin fibroblasts, representing different complementation groups,was tested for their sensitivity to the mutagenic effects of UV light, cells with the lowest rate of excision repair (XPlZBE with less than 1% of normal) showed the greatest sensitivity to the mutagenic effects of UV. Other XP cells tested showed a positive correlation between the rate of excision repair and level of resistance to UV- induced mutagenicity (Maher gt, gt,,l979). XP cells have also been shown to be more sensitive than normal cells to the mutagenic effects of several polycyclic aromatic hydro- carbons and aromatic amides (Maher e_t gt, 1976 and Heflich gt _t.,unpublished). When UV-irradiated normal and XP cells were held in a non-dividing state, in order to allow time for excision repair, normal cells showed a 31 decrease in mutation frequency over time. XPlZBE cells showed no decrease, and XPZBE cells, with intermediate repair, showed an intermediate rate of decrease over time. Normal cells have also shown the ability to recover from the potentially mutagenic effects of some aromatic amides (Heflich gt EH 1980) and from BPDE (Yang gt a_1., 1980). XPlZBE cells did not. These data suggest that exci- sion repair processes act to reduce the mutagenic effects of UV light and these chemical carcinogens. Although the majority of investigators used UV radi- ation of wavelength 254 nm, Maher and McCormick and co- workers have observed similar effects with simulated sun- light (unpublished studies). Furthermore, Trosko gt gt. (1970) have shown that, when human cells in culture were ex- posed to two hours of midsummer sunlight in mid-afternoon, there was 0.07% thymine dimer formation. Thymine is the major lesion induced by 254 nm light. Therefore, studies using 254 nm are indicative of the kinds of biological effects occurring in the skin of persons exposed to sun- light. Takebe gt a_1. (1977) have shown a direct corre- lation between the extent of excision repair capacity and the age at which XP patients suffer the onset of skin can- cers. They studied 50 XP patients and found that all those who developed cancers before the age of eight were from the group which exhibited the lowest levels of excision repair. Only one patient from this group reached the age of 32 12 without developing skin tumors. In contrast, patients with intermediate levels of excision repair did not, on average, develop tumors until after the age of 17. These data suggest that susceptibility to sunlight-induced skin cancers is dependent on the excision repair capacity of the cells. The validity of this conclusion depends on the assumption that all patients had equal sunlight exposure. Taken together, these data suggest that the greatly elevated frequency of sunlight-induced skin carcinomas in XP patients is a consequence of elevated frequencies of sunlight-induced mutations in the target cells. The evidence suggests that. the increased frequencies of mutations and cancer are a direct result of the reduced ability of XP cells to repair the sunlight-induced DNA damage compared to normal cells. d) A second form of xeroderma pigmentosum disease, XP variants There is one group of XP patients, referred to as XP variants, who exhibit the same clinical manifestations as the classical XP patients, but who have approximately normal rates of excision repair following UV-irradiation (Cleaver,l972). These cells appear to have a defect in cellular processes involved in replicating DNA using a template containing unexcised lesions (Park and Cleaver, 1979). The results of Park and Cleaver indicated that in 33 the first few hours after irradiation, XP variant cells synthesize smaller fragments of DNA than do normal cells. However, the DNA increased in size at the same rate in normal and variant cells. They concluded that UV-damaged sites interrupt DNA replication in variant cells more often than in normal cells. When Maher ‘gt_ gt. (1976) investi- gated the frequencies of 8-azaguanine-resistant mutants induced by ultraviolet light, they found that higher frequencies of mutations were induced in XP variant-derived fibroblasts than in normal cells at equal doses of UV- light. When mutation frequencies per surviving cell induced by UV-light were plotted against the cytotoxic effect of the exposure, the XP variant cells exhibited a higher number of mutations per lethal event than did the normal cells or excision repair-deficient XP cells. These results suggest that XP variant cells employ an abnormally error- prone mechanism to deal with unexcised DNA lesions. The fact that cancer-prone XP variants have a defect which differs from that of classical XP patients, and yet which results in XP variant-derived cells being hypermutable by UV light is in support of the somatic cell mutation theory of the origin of cancer. e) Bloom's syndrome Another human syndrome which exhibits a high frequency of cancer is Bloom's syndrome (BS). These patients show a 34 marked increase in frequency' of leukemias and lymphomas over the normal population (Bloom, 1966). Fibroblasts derived from BS patients exhibit increased spontaneous levels of sister chromatid exchanges and chromosome instabi- lity (Chaganti gt gt., 1974), including breakage and rearrangements. It has recently been reported by Warren gt gt. (1981) that fibroblasts derived from two BS patients exhibited increased rates of spontaneous mutation to 6-thioguanine-resistance, when compared. to cells from normal human donors. In these studies, untreated skin fibroblasts from both normal and BS patients were assayed for background levels of HPRT- mutants after 15 days of growth in non-selective medium. These data also support the somatic mutation theory of the origin of cancer because they show a correlation between a high incidence of somatic cell mutations and the elevated frequency of cancers observ- ed in Bloom syndrome patients. The mechanism of the hyper- mutability in these cells is unknown. Evidence for other mechanisms of carcinogenesis a) Teratocarcinomas The most impressive evidence which argues against the somatic mutation theory of cancer is work on mouse terato- carcinomas, which are malignant epithelial tumors which can be induced to arise with high frequency when inbred mouse 35 embryos are transplated to an extra-uterine site. Mintz and Illmensee (1975) found that when cells from a mouse terato- carcinoma were injected into normal mouse blastulas, normal mice developed containing tumor-free tissues derived from the teratocarcinoma. Upon autopsy, analysis of tissue samples from these embryos revealed teratocarcinoma—derived cells in all of its tissues. This evidence challenges the concept implied by the somatic cell mutation theory that malignant transformation is irreversible. In contrast, it suggests that some tumor cells can be restored to normal behavior. These data argue in favor of an epigenetic mecha- nism of carcinogenesis. However, it is possible that terato- carcinomas arise in response to stimuli from the foreign environment of the kidney capsule (where mouse blastocyst cells gave rise to these tumors) and when such apparently induced tumor cells are returned to a normal environment, i.e. the blastula, they then receive normal stimuli and develop normally (Martin, 1980). b) Studies with plastic film When sheets of plastic film were inserted into the tissues of rats, tumors were induced (Alexander gt g_l., 1958). However, when perforations were made in the film, or when the film was ground and injected into the animals, no tumors were induced. It is difficult to explain these re- sults in terms of the somatic cell mutation theory, since, 36 if the intact film caused mutations, e.g. by introducing a mutagenic chemical, the same chemical should have been contained on the perforated or ground plastic and these should also have caused tumors. These data suggest that tumors were induced by interruption of cell-cell communi- cation in rat tissues as a result of the plastic film. One possible mechanism is that initiated cells are suppressed in some way by neighboring cells, and that when cell-cell communication is interrupted, these initiated cells are allowed to be expressed. c) Atguments against the XP story Although the fact that cells from persons with some cancer prone syndromes are deficient in excision repair and hypermutable by several carcinogens provides the best support for the somatic mutation theory of cancer, questions regarding the extrapolation of these data to the mechanism of cancer 12 ytyg have been raised. Cairns argued recently (1981) that if the increased incidence of skin tumors in XP patients is due to mutations induced. by exposure to sunlight, as assumed by the somatic cell mutation theory, and if XP cells are also hypermutable by other environmental carcinogens, these patients should show an increased risk of developing internal cancers as a result of exposure to environmental chemicals. In support 37 of this, Kraemer has reported at least thirteen XP patients with primary non-skin, non-eye neoplasms (1978). These include several neoplasms of the oral cavity (which may be the result of sunlight exposure), two primary brain tumors, a single testicular sarcoma, breast carcinoma, and a benign thyroid tumor. Cairns (1981) has argued that the incidence of internal cancers should be much higher than this based on the fact that several groups have shown that XP cells are deficient in the rate of excision repair of lesions induced by certain chemical carcinogens (Maher gt gt" 1976; Heflich E gtul980). However, in discussing the question, Cairns assumes that cells from XP patients are equally as deficient in rate of repair of such lesions as they are in removing UV-induced damage. The data does not support this view. For example, studies with polycyclic aromatic hydrocarbon derivatives (Maher gt gt.,1976) indicate that the differential in rate of repair of these lesions between normal and XP cells is three to six—fold. For repair of UV-induced DNA damage, the differential is ten to fifty-fold. Therefore, one would not expect environ- mental mutagens to be present at high enough levels to induce significant numbers of mutations even in the excision repair-deficient XP cells. Furthermore, there is no question but that exposure of the population to sunlight is greater than to chemical carcinogens in the environment because the annual incidence of new cases of sunlight- induced skin cancers in the general population is much 38 higher than of any other form of cancer (Braun, 1977). Nevertheless, the fact that skin cancer is common in patients given strong immuno—suppressive therapy (Kinlen, 1979) suggests that there are potentially malignant cells present in the cells of normal patients which are allowed to express only when their immune system is repressed. Dupuy and Lafforet (1964) have published evidence that XP patients exhibit a defect in cell-mediated immunity, determined by their inability to develop sensitization to dinitrochlorobenzene. This evidence suggests that the increased incidence of skin tumors observed in XP patients may, in part, be due to an immune deficiency. Thus, there may be more involved in the hypersensitivity of XP patients to developing skin tumors than merely increased sensitivity to somatic cell mutations caused by deficient rates of DNA repair. Taken together, these three examples suggest that the evidence available today, which argues in favor of the involvement of somatic cell mutations in the initiation of tumors, must always be tempered by that which cannot be explained by this theory. They remind one that the process of neoplastic transformation is very complex, and that in all likelihood there is more than one possible mechanism by which normal cells can become malignant. 39 Concluding remarkg Aflatoxin B1 is a very potent carcinogen in several mammalian species, and has been indicated in epidemiolo- gical studies as a possible hepatocarcinogen in man. The mechanisms responsible for cancer induction are not fully understood, but the theory which best correlates the facts known is the somatic cell mutation theory. We, therefore, attempted to study the induction of mutations by AFB1 in human cells, in an effort to further our understanding of the mechanism(s) by which chemicals induce mutations: and possibly cancers, in human cells. The availability of excision repair -deficient strains of human cells (XP- derived) provides a system in which the role of DNA repair in the induction of cytotoxicity and mutations in human cells can be studied. This system also provides a means by which the DNA adducts responsible for the biological effects of chemical carcinogens in human cells can be studied. In this study, we have attempted to study the and AFB -C1 in human 1 l 2 cells in culture, in particular, the cytotoxicity and biological effects of AFB mutagenicity of these carcinogens, the effect of DNA repair on these biological effects, and the nature of the DNA adducts responsible for these biological consequences. MATER I ALS AND METHODS Chemicals AFB1 was purchased from Calbiochem, Los Angeles, Ca. Unlabeled and generally tritiated AFB -Cl (0.78 l 2 Ci/mmole) were a gift from Dr. David Swenson, the Upjohn Company, Kalamazoo, MI. Calf thymus DNA, dGMP, dAMP, dTMP, 6-thioguanine, sodium—lauroylsarcosine, T1 and pancreatic RN ases, trichloroacetic acid, and bovine serum albumin were obtained from Sigma Chemical Co, St. Louis, Mo. Glass dis- tilled acetone, methanol, dimethylsulfoxide, chloroform, butanol, and hexane were purchased from Burdick and Jackson Laboratories, Muskegon, MI. Cesium chloride was purchased from Gallard-Schlesinger Chemical Manufacturing Corpor- ation, Carle Place, NY. Creosol was obtained from Matheson, Coleman and Bell, Northwood, OH. Phenol, Folincicolteau reagent, and hydroxyquinolene were obtained from Fischer Scientific Co, Pittsburgh, PA. Trypsin was obtained from Grand Island Biological Co, Grand Island, NY. Fetal calf serum was bought from KC Biologicals, Lenexa, KA. or from Grand Island Biological Co. Ham's F10 medium lacking hypo- xanthine, and Eagles' minimal essential medium (MEM) was 40 41 obtained from KC Biologicals. Penicillin, streptomycin sulfate, and gentamycin were obtained from Schering Corp., Kenilworth, NJ. Preparation of Solutions Phosphate buffered saline (PBS) consisted. of 8.0 (g NaCl, 0.2 9 KCl, 1.5 9 Na HPO 0.2 g KH P04 in one 2 4' 2 liter of distilled water. The solution was brought to a pH of 7.2 with l N HC1. Stock solutions of 6—thigguanine were prepared by dissolving 134 mg of 6-thioguanine in 2 m1 1 N NaOH. 98 ml of distilled water was added, and the solution. was then filter sterilized. This yielded an 8.0 mM Solution which was used at a 1:200 dilution for a final concentration of 40 uM. Antibiotics 5 g of streptomycin sulfate and 5x106 units of penicillin were dissolved in 166.7 ml distilled water and then filter sterilized. This stock solution was used at 1:250 dilution to yield a final concentration of 100 units penicillin/ml and 100 ug streptomycin/ml. Stock solutions of gentamycin were prepared by dissolving 1 g in 100 ml distilled water. The solution was then filter steri— lized and used at a 1:200 dilution to yield a final concen- tration of 50 ug/ml. 42 Versene trypsin was made with 8.0 9 NaCl, 0.4 9 KCl, 0.58 g NaHCO3, 0.50 g trypsin (1:250 dilution of stock solution obtained from Difco, Detroit, MI.), and 0.50 g versene ethylene diamine tetraacetic acid (EDTA) in one liter of distilled water. Standard saline citrate (1X SSC) (0.15 M NaCl, 0.015 M Na Citrate) consists of 8.8 9 NaCl, 4.4 g of Na Ci- 3 3 trate in one liter of distilled water. It was brought to a pH of 7.0 with l N HCl for these studies. Kirby phenol 11 ml distilled water was added to 100 g distilled phenol. 14 ml M-Creosol was then added and the mixture was heated at 60°C to allow solutions to mix thoroughly. 0.1 g 8-hydroxyquinoline was added and the solution was stored at -20°C. Trichloroacetic acid (TCA) To make a 100% solution of TCA: 453 9 TCA was dissolved in 125 ml distilled water. The volume was brought to 453 ml with distilled water. The solution was then refrigerated. To make 5% or 10%solutions the proper dilutions were made with distilled water. Preparation of Medium To make 20 liters of Eagles' MEM, 44 g NaHCO3 was added to two packets of powdered Eagles' medium and the 43 volume was brought to 20 liters with glass distilled water. Medium was then filter sterilized, and a contamination check was done by incubating several bottles from.Ieach batch for one week at 37cc. To make 20 liters of Ham's F-10, 24 g NaHCO and 3 0.012 g phenol red were added to two packets of powdered Ham's medium, and the volume was brought to 20 liters with distilled water. It was filter sterilized and a contami- nation check was run as described. Cell Cultures Stocks of normal diploid human fibroblasts derived from foreskins were established, cultured and stored until use in liquid nitrogen. Normal fibroblast cultures were prepared from foreskin material. Tissue samples received in test tubes containing serum—free medium supplemented with antibiotics were sliced into small fragments with a sterile scalpel. Individual fragments were then placed into 35-mm glass culture dishes and a few drops of serum were added. After 24 hours, the cells had attached to the culture dish and fresh serum-containing nedium was added. Cells derived from skin biopsies taken from XP patients XP 12BE (CRL1223), XPZBE (CRL 1166), XP4BE (CRL 1162), XP7BE (CRL1200) and cells from Lesch-Nyhan patients (CRL 1112) were obtained from the American Type Culture Collection, 44 Rockville, MD. Cells from patient XPZBI were obtained from Dr. Colin Arlett, Brighton, England. Cells derived from malignant human tumors which developed in various organs of the body and non-malignant human cells (see Table 2 in Results section) were obtained from the Naval Biosciences Laboratory, Naval Supply Center, Oakland, California. All cultures were tested for the presence of mycoplasma a number of times during the experiments, as described below, and determined to be free of contamination. Storage of Cells For the purpose of storing cells, a 1:10 mixture of dimethylsulfoxide (DMSO) and freezing medium (Ham's F-10 with 17% fetal calf serum, and 50 ug/ml gentamycin) was prepared. Cells were enzymically detached from the culture vessel with trypsin, suspended in a small volume of this freezing mixture and placed in freezing vials. The vials were transferred to a -80°C freezer in an insulated styrofoam carton, and left there for 1.5 to 2 hours, to allow for cells to undergo freezing at approximately 1°C per minute. Vials were then placed in a liquid nitrogen freezer and stored indefinitely. In order to obtain the greatest recovery and viability of cells when thawing cells, the cells were warmed in a 37°C water bath as quickly as possible and then pipetted directly into a culture flask which contained at least 15 ml of media which 45 had been pre-warmed and pre-equilibrated in a 37°C incubator, with 5% CO 99% humidity. 2' Culture Medium Cells were cultured in a humid atmosphere of 5% CO2 and air at 37°C. Normal cells were cultured in HAM's F-lO medium lacking hypoxanthine and supplemented with 10-15% fetal calf serum. XP and LN cells were supplemented with 15% fetal calf serum. Transformed cells derived from human tumors were cultured in Eagles' minimal essential medium, supplemented with 10% fetal calf serum. A11 culture medium contained antibiotics (100 units penicillin/m1 and 100 ug streptomycin/ml or 50 ug/ml gentamycin). Spent Medium The spent medium used in biological recovery experi- ments was medium which had been on exponentially growing stock cultures for 48 hours. Selection Medium For the selection of 6-thioguanine-resistant mutants, Ham's F-lO medium. supplemented with 10% fetal calf serum and 40 (M thioguanine, was used. 40 uM was chosen because it was shown to prevent normal cells from doubling (see 46 Figure 5) and to yield the frequency of resistant cells shown in Figure 6. Testing of Serum Because batches of fetal calf serum differ significant- ly, it was necessary to test each new lot for ability to support clonal growth and growth at high density. It was also necessary to test serum to insure that it was not so rich in purines that it competed with the thioguanine in the selection medium and permited non-resistant cells to survive. Testing serum for ability to support cloning involved plating untreated cells (normal and/or XP) into 60-mm- diameter dishes with medium containing the serum to be tested and dishes with medium containing control serum, whose ability to support cloning was already known. Cells were plated into these dishes at cloning density (100- 300/dish), fed, stained, and scored as described (see it gig cytotoxicity assay). The cloning efficiency of cells grown with the test serum was compared with that of the control serum and a determination was made as to whether the test serum is adequate to support cloning. In order to test serum for growth at high density, cells were plated into a series 60-mm-diameter dishes with 47 Eamon ©o>oumdm “nu. “Esuom ummu “now .o amp :0 woumHm mHHoo occ.mh mmz Esasoocm .mxmc cm>wm nouns wcficmsmofinu :fi maaoo uo nuaoum numb» Enumw .m wusmwm halo m2H2 3113:: so ,aaswnN 48 I00 ...I J 'I _ I TIIII I TUIIHII I I I IIIIII O.I I IrIIIII] PERCENT SURVIVAL CLONING ABILITY 0.0l I I [Hill] I I I I l .J l °'°°' IS 20 25 30 35 CONCENTRATION 6-THIOGUAUNINE (ILM) h J [Anni l IIIIIIII I IIIIIIII l IIIIIIII J Llllllj] o 0' 5 0 Figure 6. Serum test: colony-forming ability in thioguanine 49 medium containing test serum or control serum, at a density of 20,000/dish. Cells were allowed to grow and their densi- ty monitored over 14 days by means of an electronic counter (Coulter) (Figure 7). By comparing the density the cells reached in the test serum with the density they reached in the control serum, it was determined whether the new serum was adequate to support growth at high density. The ability of cells to clone well and to grow to high density are two factors which determine whether cultures are healthy. It is important to use healthy cultures in these experiments to insure good cloning efficiency for the most reliable re- sults. Before a new lot of serum could be used for mutation experiments, it had first to be tested to see whether growth of non-resistant cells was inhibited by thioguanine. Cells were plated into 60-mm—diameter dishes at a density of 40,000/dish. Thioguanine was added to the dishes at various concentrations (ranging from 0-40 uM) and cell numbers monitored for seven days. Figure 6 shows a typical curve obtained in this test. Cell numbers decreased with increasing thioguanine concentrations until they leveled off near 10 uM. If the serunI contained high levels of purines, these would compete with the thioguanine for incorporation into nucleic acids and allow cells to grow at higher thioguanine concentrations. This.would not be suita- ble to use for mutation experiments. 50 I3. IjIlIleI111T rIItW is NUMBER or CELLS (XIo‘E’) ES B11JIIIJ4IIIIII °I 3 S7 QII I315 DAYS Figure 7. Serum test: growth of cells to high density. (0) NFSL18; (D) NF812 51 Serum was also tested to determine whether it would yield a low enough frequency of thioguanine-resistant mutants from an untreated population to make it acceptable for selection of induced 6-thioguanine-resistant mutants (see Figure 6). Untreated cells were plated into dishes at different cell densities (ZOO-72,600) and thioguanine concentrations (0-40 pm), and were fed, stained, and scored as described. One population of treated cells, which were ready to be selected in an experiment using serum which had already been tested and was in use for mutation experiments, was also plated into dishes at a cell density of 72,600 (500/mm2) and thioguanine concentration of 40 pM. 'The number of induced. mutants obtained in the test serum and the control serum. were then compared as an indication of the ability of the new batch to support growth of newly induced mutants. Testing for Mycoplasma Contamination 50 pl 3H uridine (100 uCi/ml) and so [11 14c uracil (10 uCi/ml) were added to cell cultures in exponen- tial growth and the cells were allowed to incubate 4 hours or longer. The supernatant was poured off and the cells in each flask were washed twice with PBS. To each flask 5 ml of 0.1% sodium lauroylsarcosine (SLS) was added for 5 minutes. The SLS was poured off and added to 10 ml 20% TCA. Another 5 ml of 0.1% SLS was used to wash each flask and 52 this was also added to the TCA. This solution was left on ice for 30 minutes. Solutions were filtered individually, using Whatman GF/A paper,and the filters were washed with 10% TCA. Filters were dried, and placed into individual scintillation vials, and then 0.2 m1 of 0.2 M HCl was added to each. 10 ml of scintillation fluid was added and the samples were counted on a Beckman 9000 LS with dpm capabili- ty. If a ratio of 3H to 14C greater than 20:1 was seen, it was assumed that the cells were mycoplasma-free, since 14 only mycoplasma will incorporate the C-labeled uracil. Human cells will incorporate uridine, but not uracil. Benzo(a)pyrene (BP) Metabolism Assay In order to assay for BP metabolism, exponentially 2 T flasks with tri- growing cells were treated in 25 cm tiated benzo(a)pyrene (BP) in serum-containing medium, and allowed to incubate for 24 hours. The media was then taken off the cells, extracted twice with hexane, and the aqueous fraction was then separated and assayed for radioactivity by adding 1 ml to 10 ml aqueous scintillant and counting the radioactivity in the solution on a Hewlett Packard Tricarb Scintillation Spectrometer. Since many of the cellu— 1ar metabolites of BP are water soluble, while BP is not, the amount of radioactivity present in the water fraction was taken as a measure of the extent of BP metabolism that had taken place in the cells. 53 Protein Assay In order to normalize the results of the many differ- ent cell samples assayed for metabolism of BP, a protein determination was made on each sample assayed. The specific activity for each sample was then calculated (water soluble metabolites produced per milligram of protein present in the cell sample). The protein assay used was that of Lowry modified by Oyama and Eagle (1956). Three solutions were used: 5 consisted of 209 Na CO 2 3’ tartrate in 1 liter distilled water; 2 consisted of 5 g 49 NaOH, 0.29 NaK CuSo4'5H20 in 1 liter distilled water; Q consisted of 50 parts 5 to 1 part _B_ prepared fresh daily. Folin- Ciocalteu phenol reagent was also used. After the medium was removed from the monolayer of cells for hexane extraction, the cells remaining in the flasks were lysed by incubating them in 3 ml Lowry solution _A for 15 minutes. A 0.5 ml aliquot of the resulting solution was added to 1 ml distilled water and the total volume was brought to 6.0 ml with solution 9. Phenol reagent (0.5 ml) was added and the mixture was incubated for an additional 30 minutes. The absorbance at 650 nm of each sample was read,using a Beckman spectrophotometer,and compared to those of protein standards run with each assay to determine the amount of protein present in each sample. The protein standards consisted of bovine serum albumin (BSA) samples at three different concentrations (20, 40, 60 ug/ml) which had been 54 incubated in 3 ml of solution _A for 15 minutes, after which the volume was brought to 6 ml with solution 9 and the phenol reagent was added as described. The absorbance at 650 nm was determined and the standard curve was constructed from which to determine the protein concentra- tion of the experimental samples. Figure 8 represents the results of protein assays determined for a series of cell strains which had been assayed for metabolism at the same time. The curve was constructed from BSA standards (illus- trated by closed circles). The protein concentrations of the experimental samples were determined by correlating the absorbance values obtained at 650 nm to the protein concen- tration, according to the standard curve. In Situ Cytotoxicity Assay Between 100 and 3000 cells were plated into 60-mm-dia- meter dishes (6—12 dishes per treatment dose) and allowed to attach for 15-18 hours. No more than 3000 cells were plated into this size dish because there appears to be an effect on cloning efficiency, such that cells plated out at higher density show an increased cloning ability than those plated at less than 3000. This tends to give an artificial- ly increased survival result when cells are plated at high- er density. There are also reports that at higher densities than these, toxins released from dead cells may reduce the 55 I I I B.I8 8.16*- _ @14- _. E c 8 9.12— _ co '— < 8.1 - _ LL] 0 <21 8 08F- m ' '7 or 8 in 0.06— _ < 8.04— — 2.02— _. B I I I I 2 28 4B 68 88 I BE PROTEIN CONCENTRATION 011g) Figure 8. Determination of protein content in cellular samples. (0) BSA control; (A) PC-3; (I) HsB35T; (Q) Hs703T 56 survival. The culture medium was then replaced with serum- free medium. (The particular medium used was determined by the cell type and is designated in the text.) The carcino- gen, dissolved in 100% glass distilled acetone dried over a molecular sieve, was introduced into the individual dishes by micropipette. The final concentration of acetone in medium was not more than 0.5%. After two hours, the medium containing carcinogen was removed and replaced with fresh culture medium. Cells were fed after one week and stained when clones were of macroscopic size (at approximately 2 weeks). At this time they were rinsed in saline, fixed in methanol, stained with 0.2% methylene blue, and counted. The cloning efficiency of the treated cells, divided by the cloning efficiency of the control cella.which received only solvent, determined the cytotoxicity of the compound and was expressed as a percent. Replating Cytotoxicity Assay Cells in exponential growth were trypsinized and plat- ed into 60-mm-diameter dishes. The desired densities (0.5x105 - 0.2x106) depended on the expected survival levels and corresponded to the cell densities at which cells were treated for mutation experiments (see below). Cells were allowed to attach for 15-18 hours. The medium was then changed and the carcinogen was administered in the 57 same way as has been described above for 1 situ experi- ments. After a treatment period of two hours, the medium was removed and the cells were rinsed with PBS, trypsiniz- ed, resuspendended in fresh culture medium and plated into 60-mm-diameter dishes (6-12 per treatment dose) at cloning densities (100-3000/dish). Colonies were fed, stained, and scored as already described. Assay of Biological Recovery from Potentialty Cytotoxic Effects Cells (”0.5x104) were seeded into 60-mm-diameter culture dishes, fed every 2-3 days with fresh culture medium and allowed to grow to confluence. Cells were fed with fresh medium on the day they reached confluence, then kept at confluence for 3 days without feeding before they were treated with carcinogen. (For example, cells which were fully confluent on Sunday were treated on Wednesday). Administration of carcinogen has already been described. Following 2 hour treatment, cells to be assayed for survi- val immediately (day 0) were trypsinized, resuspended, and plated into 60-mm-diameter dishes at cloning density. Those cultures to be assayed for survival at later times had the treatment medium replaced with the spent medium which had been removed from the dishes at the time they received serum-free medium. If they were maintained in confluence for more than 24 hours, they were fed daily with spent 58 medium. At the designated times, they were trypsinized, and plated out for assay of survival of colony-forming ability. Cell-mediated Cytotoxicity Assay Transformed human epithelial cells (Hs703T, H5835T, see Table 2) were grown to high density in 250 ml flasks (75 cm2) (Corning, NY), trypsinized, suspended in culture medium, and X-irradiated in suspension on ice with 3000 rads using a GE Maxitron 300 kvp X-ray unit (300 rads/mi- nute for 10 minutes). Irradiated cells were used immediate- ly or were frozen as described for later use. Cells were shown to retain their ability’ to attach to the plastic dishes following freezing and thawing by monitoring the numbers of cells which attached to plastic determined over several days post-irradiation. They were tested for their ability to metabolize carcinogens by comparing survival in target cells treated in the presence of metabolizing cells before and after X-raying. The cytotoxicity assay itself involved plating target cells (normal diploid or xeroderma pigmentosum-derived human fibroblasts (NF or XP)) into 60-mm-diameter dishes at desired densities (0.3-2.6x105/dish). After allowing 4 hours for the cells to attach, the transformed metabolizing cells were seeded on top of the target cell cultures at various densities (0.3-1.2x106), depending upon the 59 purpose of the experiment. (See Results section) Cells were then treated with AFBl which was dissolved in DMSO and administered in serum-containg medium because a treatment time of 24-48 hours was needed in order to allow time for the carcinogen to be metabolized into a reactive form. Keeping cells in serum-free medium for 24 hours was shown to have a deleterious effect on cloning ability. Following treatment, the medium was removed, and the cells were rinsed and trypsinized from the culture dishes. Because the transformed cells detached from the plastic more quickly than did the untransformed fibroblast target cells, following addition of trypsin to the dishes, nearly all the transformed cells could be seen to be floating in the trypsin while the untransformed cells were still attached. It was thus possible to separate the two different cell types from each other. When the majority of the epithelial cells had been removed, the target cells were trypsinized, counted with a hemacytometer, resuspended in fresh culture medium and plated into dishes at cloning density. They were then fed, and when the colonies had developed, were stained and scored as described. Replating Mutagenicity Assay Cells were grown asynchronously for three days, then trypsinized and plated into dishes at the desired densi- 6 ties. For each dose, a population of 0.33 - 1.3x10 cells 60 was plated into lSO-mm-diameter dishes. The number of dishes plated was determined to furnish a surviving popula- tion of at least 1.0x106 cells for each dose following treatment. Cells were allowed 15 to 18 hours to attach. Culture medium was replaced with serum free medium and the cells were exposed to carcinogen as described. Cells were allowed to replicate and undergo phenotypic expression in their dishes for approximately two cell divisions, after which each treatment population was trypsinized, pooled and replated at lower density as necessary (0.3-1.0x106/dish; 3-5 dishes per treatment) in order to maintain the cells in exponential growth during the expression period. The length of the expression period was adjusted for each determina- tion to allow the cells to undergo between 4 and 6 popula- tion doublings before selection of 6-thioguanine-resistant mutant cells. In order to monitor the number of population doubl- ings, it was necessary to estimate the percent survival for each treatment based on the dose administered and data from previous cytotoxicity experiments. From this we can esti- mate the number of cells which had been killed and the number of cells remaining alive. The increase in cell numbers during the expression period was monitored with an electronic cell counter. From the data, we estimated how many divisions the living cells had gone through, taking into account what fraction of the population is represented as 61 non-dividing dead cells. However, only after the results of the accompanying cytotoxicity experiments became available (approximately 14 days after treatment, which is usually several days after the cells have been selected) were the actual survival levels known and the number of cell divi- sions accurately determined in retrospect. From related studies carried out in this laboratory, it appears that at least four to six doublings are required before the cell is resistant to thioguanine (Maher gt gl., 1979). Cells are resistant to thioguanine when they do not have a functional hypoxanthine/guanine phosphoribosyl trans- ferase (HPRT) enzyme activity. This number of divisions is probably necessary because the cell must dilute out the residual number of HPRT enzyme molecules present in the cell before the mutational event occurred. At the end of the expression period, the cells were trypsinized, pooled, and plated into plastic lOO-mm-dishes (64 dishes per point) in selective medium at a density of 300-500 cells/cm2 (16,500-27,500 cells/dish). Normal diploid and XP-derived human fibroblasts have been demonstrated to be capable of transferring the phosphorylated nucleoside of 6-thioguanine (Corsaro and Midgeon, 1977). This process is referred to as metabolic cooperation. Passage of these nucleosides from HPRT+ cells to HPRT- cells results in cell death for the HPRT- mutant. Therefore, the cells had to be plated out at low enough density (300-500/mm2) for selection of 62 6-thioguanine-resistant mutants, in order to minimize the loss of mutants from metabolic cooperation. Thioguanine was present in the selective medium at a final concentration of 40 (AM in order to prevent non- mutant colonies from developing. Cells were refed with selective medium after seven days and stained after two 6 cells was weeks. A control population of 1.0-1.5x10 included with every experiment to correct for background mutation frequency. At the time of selection, cells from each treatment were also plated at cloning densities in non-selective mediuni in order to determine cloning efficiency at the time of selection. This number is needed in order to calculate the number of mutants induced from the number of mutant colonies observed because the cells plated in the selection dishes are plated at very low (nearly cloning) densities, and therefore corrections must be made for the cloning efficiencies of the cells. A cytotoxicity experiment (in some cases using the in situ technique, in others the replating technique) was carried out along with each mutation experiment to determine the percent survival. However, since the cell density at which cells were treated 1 situ was con- siderably lower than the density at which cells were treat- ed for induction of mutations, the cytotoxicity results could not be taken to represent exactly the amount of cell 63 killing which was happening in the mutation experiment. Therefore, in some cases, a replating cytotoxicity accompa- nied the mutation experiment. However, although these cells were treated at the same density as the cells in the mutation experiment, unlike the latter they were trypsiniz- ed directly after treatment, and therefore, they did not represent the same conditions under which the mutation experiments were carried out either. However, since it was not possible to carry out cytotoxicity assays under exactly the same conditions as the mutation assays, these two types of cytotoxicity assays were taken to represent the range of killing values produced. Fortunately, the variations between the results from the two types of cytotoxicity experiments was not very great, and therefore the cytotoxi— city experiments were assumed to give a reasonable approxi- mation of the actual cell killing. Assay of Biological Recovery from Potentially Mutagenic Effects Cells were seeded into 250 ml flasks (75 cmz), fed regularly with fresh culture medium, and allowed to grow to confluence. Cells were fed with fresh medium on the day they reached confluence, then kept at confluence for three days without refeeding (as described in the assay of biological recovery from cytotoxicity). Cells were given a two hour treatment with AFBl-Clz. Those cultures to be 64 assayed immediately for mutations (day 0) were trypsinized, pooled, resuspended and plated out at expression densities. Those cells to be assayed at later times (day 2, 4, 8) were fed with spent medium as described and assayed at the desig- nated times in the same way as day 0 samples. Determination of cell numbers, numbers of plates needed, population doubl- ing, selection densities etc. were made in the same way as in replating mutation experiments. Cytotoxicity experi- ments, and replating efficiency dishes were included for all experiments as described. Reconstruction Experiments As explained above, mutants can be lost as a result of metabolic c00peration. In order to determine the efficiency of recovery of induced mutations under the selection con- ditions used, and to detect the effect of metabolic coope- ration, a known number of HPRT- Lesch-Nyhan (LN) cells were seeded at the time of selection into a series of con- trol and experimental dishes. LN cells are derived from human patients with an inherited defect in HPRT activity. The control dishes contained only Iesch-Nyhan cells, while the experimental dishes contained cells from each treatment pOpulation and one untreated control population at the same density at which they have been seeded for selection. The number of Lesch-Nyhan clones observed for each treatment 65 group,divided by the number of Lesch—Nyhan clones seen in the control dishes, indicated the fraction of 6—thio- guanine-resistant mutants recovered. This number was a factor used in calculating the number of induced mutations from the number of mutant colonies observed. Reconstruction experiments accompanied all mutation experiments. The results indicated a: recovery of 75-100% of the LN colonies in our experiments. Isolation of DNA from Human Cells Cells were rinsed with PBS, then harvested in trypsin- versene, buffered at pH 7.0 with potassium phosphate. Cells were lysed in 0.5% sodium lauroylsarcosine and incubated at 37°C for 30 minutes. (At this point cells were frozen at -20°C for future analysis) Pancreatic RNase (25 ug/ml) and T RNase (20 units/ml) were added and the solution 1 was allowed to incubate for 30 minutes at 37°C. Pronase (0.1 ug/ml) was added and the solution was incubated for 30 minutes at 37°C. The solution was then extracted twice with a 24:23:1 mixture of Kirby' phenol, chloroform. and butanol. The two phases were separated by centrifugation at 12,000 x g for 10 minutes. The DNA in the aqueous phase was collected and dialyzed 3 times against 4 liters of SSC, pH 7.0. 66 Purification of DNA Samples A 4.5 ml aliquot of the dialyzed DNA solution was added to 5.46 g CsCl for each gradient, then centrifuged to equilibrium in a Beckman ultracentrafuge at 105,000 x g for 18 hours. The absorbance profile at 260 nm of each fraction of gradient was determined (Figure 9) and the DNA fractions were pooled. The DNA gradient was fractionated with a DENSI-FLOW fractionator, Searle, Fort Lee, NJ. In those cases where DNA samples were used for determining the level of carcinogen binding to DNA, the purified DNA solution was then dialyzed against 4 liters of SSC. For high pressure liquid chromatography (HPLC) studies, the DNA fractions were dialyzed once against SSC, then once against 0.1x SSC gfii 7.0 and finally once against 0.0015 M sodium citrate pH 7.0, in order to minimize the salt concentration present in the DNA solution. HPLC Analysis of DNA Adducts Cellular DNA samples were lyophilized to reduce the volume lO-fold before acid hydrolysis. The DNA samples were hydrolyzed under mild acid conditions (0.15 N HCl at 100°C for 15 minutes). After hydrolysis, all DNA samples (cellular and calf thymus DNA) were lyophilized and redis- solved in 0.5 m1 methanol. Potassium acetate (0.1M) was added to bring the pH to 5.5, and these samples (0.5-0.7 67 I I I I I I 8.8—f - g as— — I: <1 cm _ CK CD 00 ‘2 a4— - az— J .. I I l I I goo 22a 24a 26% 288 see WAVELENGTH (“I") Figure 9. DNA absorption spectrum 68 ml) were then subject to high pressure liquid chromatogra- phy, using a Spectra Physics (Sunnyvale, CA) SP8000 instru- ment equipped with a Waters Associates (Milford, MA) 18 column (3.9 x 300 mm). A 70 reverse phase II Bondapak C minute methanol/water gradient (concave see Table l) was applied to the column which was maintained at 50°C with a flow rate of l ml/minute. Fractions were collected at 5 minute intervals and mixed with 10 ml of aqueous counting scintillant (Amersham,Arlington Heights, IL.) and radio- activity was measured on a Beckman LS9000 Liquid Scintil- lation Spectrometer equipped with dpm calculation capacity. For all HPLC profiles of cellular and calf thymus DNA, the time (ie. day 0, 2, 8) reflects the time at which the DNA isolation procedure was started. This does not take into account the time required to isolate and purify the DNA, which was 24-72 hours. Therefore, the HPLC profiles reflect the DNA adducts present in the DNA following these processes. (See Discussion Section). Studies with Calf Thymus DNA For studies with calf thymus DNA, 5 uCi of 3H- AFB -Cl2 (2.5 ug) were reacted with 10 mg calf thymus 1 DNA in 7.5 ml 0.02 M potassium phosphate buffer, pH 6.7, and incubated at 37°C. Samples were taken at various times up to 8 days after treatment. The DNA was isolated, 69 Table 1. HPLC gradients used Time metfianol waEer Gradient l. 0.0 10 90 24.0 52 48 56.0 80 20 58.0 100 0 70.0 100 0 Gradient 2. 0.0 10 90 20.0 45 55 50.0 45 55 70.0 100 0 70 purified, and its specific activity determined as describ- ed. The optical standards used forcxrchromatographic com- parison with the experimental (cellular and calf thymus) DNA samples were synthesized by reacting 25 ug cold AFBl- Cl2 with 1 mg dGMP in 2 ml 0.02 M potassium phosphate buffer pH 6.7, or with 2 mg calf thymus DNA in 5 ml buffer for 30 minutes at 37°C. Unreacted compound was extracted three times with ethyl acetate. Standards made with calf thymus DNA were acid hydrolyzed as described and both dGMP and calf thymus standards were lyophilized and redissolved in buffer as described for DNA adducts. RESULTS STUDIES WITH AFBl Identification of a human cell strain capable of metabolizing aflatoxin B1 AFBl requires metabolic activation before it can exert its biological effects (Garner gt a_1.,l972). Since these cells are unable to metabolize AFBl, it was necessary to employ an activating system in order to study its effects in human fibroblasts in culture. Because benzo(a)pyrene (BP) and AFB are both metabolized by the l cytochrome P450 mixed function oxidase enzyme systems present in udcrosome fractions, it was reasoned that cells retaining the ability to metabolize BP would be likely to have AFB1 metabolism ability. Table 2 shows the results of screening eighteen differ- ent cell strains, including sixteen cell lines derived from human tumors, and two normal human strains for their abili- ty to metabolize BP into water soluble metabolites. The ability of Hs835T cells to produce such metabolites was 71 72 I maa wma. «.0 + ~.m mva vsa. wsocaoumoocwpm vaumumnumz Bowen: I owe mam. v.o + o.m moo «as. osooaoooo mesa same I msm mmv. v.v + s.m mssa vev. xooc mo neonaouwo naOEuopamm mmmm I coma wmv. e.o + s.oa .MNma owe. meooaotoo socoae wave wmvm vow. mamm com. I mom ssa. o.m + m.~a «awe omo. meooaoooo soooax ammoom m.ma mmmva ~m~.m neonaouwo um>aa amosmm I mesa wma. m + ~.e~ mesa «om. neonaoomo aooocsooeo momooo I mawm mam. w + w.s~ amsa mam. neonaounUOcocm vauwummumz Boosmm :aoDOum AmEV @E\©wuuo>coo om Camacho was» no mwaoeocnc Emu amu0u :Oaumauommo aaou usua>auom Damaowmm .mcamuuw aawo ewes: amuw>mm ea Emaaonmuwe ocmuxmanvoNcon mo am>wa one .N oaooe 73 v.o mm mmm. umMaQOAQam caxw anEuoz m.aommz . Nma mam. ucoaumm uoocwo oumumoum e.o H o.o sew mmo.a soda zooooe ooom muoo ~.a H ¢.~ «om «ms. ooaoooooa anatoz ocaosome I om vow. v.o + m.c vaa mos. meoummaQOaaw msam I no msm. v.c + m.o ow «mm. neocoamz Bmamm: m.o mm mma. neonaouMUOauocu mdb I ova «mm. w.o + m.a mva moo.a nEooummOHQam omoaem I mma vva. w.a + v.~ sma msm. nEOaaw Bmmwmm I mmv wmv. v.0 + ~.m amv wam. mnmuocmm mo mEOCaoumo vaumuwmuoz wmsmm Camacho AmEV mE\pouuo>coo mm eaou0um mama mo moaoeocmc Emu amuou ceaumauomoo aaoo usua>auom Dawaommm Ap.ucoov .m wanna 74 assayed as a positive control with each series of cells screened to insure consistency in the test system. The results indicated that six of the strains tested (all of which were transformed cell lines) exhibited a consistently greater level of metabolism than the other twelve strains tested. The data were expressed in terms of activity per mg of protein as determined by the Lowry protein assay. This assay was used to facilitate comparison of different cell lines rather than cell numbers because the amount of protein was found to vary from one cell type to another. For example, the data in Figure 10 indicated that PC-3 cells have a higher cellular protein content than do NF811. The effect of cell number and exposure time on the extent of metabolism Aryl hydrocarbon hydroxylase (AHH) activity was measur- ed as a function of cell number and time of exposure to BP. Figure 11 illustrates that, in PC-3 cells treated with BP, the level of water soluble metabolites produced increased linearly with the number of cells treated. The data in Table 3 indicate that increasing the exposure period from 24 to 48 hrs for PC-3 cells treated with BP resulted in a significant increase in the level of metabolic conversion of BP. Langenbach et gt. (1979) have shown that primary 75 NUMBER OF CELLS (XIO'S) BL I I I I B O. I 8.2 8.3 B . 4 TOTAL PROTEIN 0113) Figure 10. The relationship between cell number and protein content in PC-3 and normal cells. (A) NF; (0) PC-3 76 5 I I I 5- / '- "5‘ 0° 0 . :2 4- 9/ a t. o ‘0. U 3— 7 _ - LL 0 E 0 G1 5 2" “ Z: C)c> C>0 I 00 _ o I J J 2 3 4 I 5 6 CPMIon’Z) Figure 11. The relationship between number of cells assayed and the level of water soluble metabolites produced. PC-3 cells 77 Table 3. Level of BP metabolism in PC-3 cells following 24 and 48 hour exposures to BP Specific activity: nanomoles of BP converted/mg. protein 24 hr. exposure 0.254 0.286 0.368 0.269 average = 0.294 t .051 48 hr. exposure 0.568 0.771 0.856 0.856 average = 0.763 + .136 78 rat kidney cultures exhibit a higher level of AFBl meta- bolism than do rat skin fibroblasts, suggesting cell specificity in carcinogen metabolism. We, therefore, chose to use the transformed human liver cell line (Hs703T) in our cell-mediated system for activation of AFB1. We also chose to test the transformed human kidney line (Hs835T) because it exhibited properties which made it easier to culture and it also exhibited high levels of BP metabolism. Cell-mediated gytotoxicity studies with AFB1 Figures 12 and 13 show the cytotoxic response induced by AFBl in XP target cells in the presence of H3835T (Figure 12) or Hs703T (Figure 13) metabolizing cells. In the presence of either of these metabolizing systems, concentrations of AFB1 as low as 1 (III were cytotoxic to the XP target cells. XP cells showed no cytotoxic response in the absence of an activation system. These data present evidence of metabolic conversion of AFB1 to a toxic metabolites in this cell-mediated system. Figure 14 compares the cytotoxic response of XP cells in the presence of the two different metabolizing cell lines, viz. Hs703T and Hs835T when seeded at the same density. The results showed there was more killing in the presence of Hs703T cells than in the presence of Hs835T cells, suggesting that metabolic conversion of AFB1 was more efficient in Hs703T cells than in Hs835T cells. 79 IBB PERCENT SURVIVAL CLONING ABILITY A A A ,9 J I I I a 12 2a so 49 so DOSE AFB1 (pH) Figure 12. Cell-mediated cytotoxicity of AFB in XP cells using two concentrations6 of Hs 5T gells to activate AFBl. ( A) 0.6x10 ; (A) 1.2x10 PERCENT SURVIVAL CLONING ABILITY 180 12 8O J J I Io 2o 3o 4o so DOSE AFB,L (pM) Figure 13. Cell-mediated cytotoxicity of AFB in XP cells using two concentrations 0% Hs703T cells to activate AFB . (A) 0.4x10 ; (A) 0.6x10 6 1 81 ‘99 I I I I PERCENT SURVIVAL CLONING ABILITY m I I I 8 I8 28 38 48 S8 DOSE AI-‘Bl (pM) Figure 14. Comparison of the levels of cytotoxicity induced by AFB in XP cells in the presence of equal numbers of Hs835T and Hs703T cells. (A) H5835T; (I) Hs703T 82 The effect of varying the number of metabolizing cells on the cytotoxic response in target cells Figure 12 shows the effect that varying the number of metabolizing cells present during AFBl treatment had on the cytotoxicity observed in target XP cells. When XP cells were treated with AFBl in the presence of two different ratios of H3835T cells to XP cells, the cytotoxic response was greater when the metabolizing cells were present at the higher ratio. The exposure time in both cases was 24 hours. The same effect was observed when the cells were treated in the presence of two different ratios of Hs703T cells to XP cells, as shown in Figure 13. This is probably because more of the AFB1 was converted into the reactive metabolite when more metabolizing cells were present. We also noticed a distinct leveling off in both of the XP survival curves obtained with AFBl in the presence of the H5835T cells. This is probably caused by saturation of metabolizing enzymes in the cells at that concentration, or by some sort of feedback inhibition. It should be noted that this bend in the survival curve was not as evident when Hs703T cells were used as the source of metabolism. This suggests that the enzymes in the Hs703T cells were not saturated by the concentrations of AFB1 administered. Further evidence that the Hs703T cells have a higher capacity for metabolism of AFB than 1 83 do H3835T cells can be seen in Figure 14. In the presence of the same number of metabolizing cells, XP cells showed more cell killing in the presence of Hs703T cells than with Hs835T cells. A concentration of 20 DM resulted in 70% survival with H3835T cells compared to 25% survival with Hs703T cells. This is further evidence that the Hs703T cells have a higher capacity for metabolism of AFBl. This is consistent with i_r_I_ vivo data which has shown that rat liver carries out greater levels of AFB1 metabolism than does kidney (Wogan, 1968). Based on these results, Hs703T cells were determined to be a more desirable system with which to study the effect of AFB1 in human fibroblasts. Optimal ratios of metabolizing cells to target cells To further optimize the system, we wished to determine the ratio of Hs703T:XP cells at which the maximum conver- sion of AFB1 into cytotoxic metabolites was observed. It would be desirable to use cells at this ratio to ensure that consistent results were obtained. Figure 15 shows the effect on the survival in XP cells that varying the ratio of Hs703T cells to XP cells had while the concentration was kept constant at 10 pH. The results showed that cytotoxici- ty increased with increasing ratios until it leveled off at a ratio of Hs703T cells to XP cells of around 5:1. It 84 I88 PERCENT SURVIVAL CLONING ABILITY ”a I I I I 8 55 I8 ‘HS 28 125 RATIO HS783T CELLS:XP CELLS Figure 15. The effect of varying the ratio of Hs703T cells to XP cells on the cytotoxicity induced by AFBl in XP cells. 85 appears that this ratio allows for the maximum level of metabolic conversion possible in this system at this particular concentration of AFB1. AFBl-induced gytotoxicity in Hs703T cells as targets Hs703T cells, themselves, were treated. with varying concentrations of AFBl, allowed to incubate for 24 or 48 hours, and then assayed for survival. Figure 16 shows the results obtained. AFBl induced cytotoxicity in Hs703T cells after both incubation times, however the: level. of cytotoxicity was much greater after a 48 hour exposure than after 24 hours. For a 1.25 uM concentration, the survival after 24 hours was 60% and after 48 hours, survival was 1%. This may represent an induction of the enzymes involved, since a two-fold difference would be expected if this was simply an additive effect caused by the longer exposure time. Comparison of the gytotoxic response to AFBl in three different target cells The cytotoxic responses of NF, XP, and Hs703T target cells to AFB:L are compared in Figure 17. It is interest- ing to note that, while XP cells proved more sensitive than NF cells, Hs703T cells appeared to be more sensitive than XP cells. These data suggested that either Hs703T cells 86 ‘88 I I I I 1 I I I I I q "I " I >- I- -I .n— H . - _I H — —-I m < (25 19 I— '1 g i I 3 I. I U h . u 2‘ 0 " > ._ _- H > g *2- : ' Lu L .. U I. .. a LAJ '- u O. I— .- I9 I l I J I J I 14 I I 8 5 I8 CONCENTRATION AFBl cpm Figure 16. Cytotoxicity of AFB1 in Hs703T cells following 24 and 48 hour exposure times. (I) 24 hours; (0) 48 hours 87 PERCENT SURVIVAL CLONING ABILITY ‘8 I I I I I I J 8 I8 28 3B 48 CONCENTRATION AFBl cpm Figure 17. Comparison of the levels of cytotoxicity induced by AFB in NF, XP and Hs703T target cells. (0) NE; (A) xp; (CI) Hs703T 88 were deficient in the repair of AFBl-induced cytotoxic lesions, or that the Hs703T were receiving a higher effec- tive concentration of AFBl. The latter is a more likely explanation since metabolisnI is taking place within the target cell itself. The reactive metabolite of AFBl, the 2,3-epoxide, is very unstable, and is believed to have a half—life of only a few seconds in water. Therefore, it would Iufl: be expected to reach the target cells as readily as it would reach the metabolizing cells themselves. Since AFBl metabolism is probably taking place in the nuclear membrane of the Hs703T cells, more of the metabolites produced at a given concentration will reach the DNA in Hs703T cells than in XPs, where the highly unstable metabo- lite must be transported from Hs703T cells to XP cells and then into the nucleus. Thus, one would expect the actual amount of the reactive metabolite reaching the XP cell DNA to be much lower than that reaching Hs703T DNA. One way in which to test this hypothesis was to expose XP, NF and Hs703T cells to a carcinogen which does not require metabolic activation, and assay for survival. We, therefore, exposed all three cell strains to AFB -C1 1 2' the direct-acting analog of AFBl—2,3-oxide. The cells were treated 1 situ with the same carcinogen solution all at the same time. The results (Figure 18) showed that XP cells were more sensitive to the direct-acting carcino- gen, AFBl-Cl than were NF and Hs703T cells, which 2' 89 I88 PERCENT SURVIVAL CLONING ABILITY m I I I I 8 5 I8 IS 28 25 CONCENTRATION AF81 -CLZ CHM) Figure 18. Cytotoxicity of AFBl-Cl in NF, XP, and Hs703T cells. (0) NF; (“2 xp; (I) Hs703T 90 appeared to be equally sensitive. These results suggest that Hs703T cells have normal DNA repair and that the greater sensitivity that Hs703T cells showed for AFBl was probably the result of a higher effective concentration. Use of AFBl-Cl2 as a model for AFBl in studying the biological effects of AFB1 in human cells in culture In order to investigate the role of excision repair on the cytotoxic and mutagenic effects of AFB1 in human cells, it was necessary to use repair-proficient cells. However, it became evident from these preliminary studies that if cytotoxicity and mutations in normal cells were to be induced with the cell-mediated system it would require very high concentrations of AFBl. In fact, with these cells, we saw no toxicity for the concentrations we administered. This is probably because during the length of exposure time needed for metabolism (24 or 48 hours) normal cells are continually carrying out excision repair of the potentially cytotoxic DNA lesions induced by AFBl. Because of the high concentrations of AFB1 required for normal cells, the need to treat large number of cells for the binding and adduct studies planned, and the high cost of radioactive AFBl, it became apparent. that ‘these studies would be very difficult and very expensive. Al- though Hs703T cells required much lower concentrations than 91 normal cells before cytotoxcity was induced, they could not provide a system in which to study the biological effects of AFB1 because the recovery experiments proposed (see below) required that cells be held in a non-dividing state by means of contact inhibition. Hs703T cells, like most transformed cell lines, are not contact inhibited. As an alternative, we chose to use AFBl-Clz, which has been synthesized as a model compound for AFBl-2,3- oxide, the probable reactive metabolite of AFB1. AFBl- C12 is a direct-acting carcinogen, which binds to DNA and induces mutations in Salmonella typhimurium in the absence of any activating system, and produces sarcomas at the site of injection in rats. As expected for AFBl-2,3- oxide, it has an electrophilic carbon 2, and the major DNA adducts are formed between the carbon 2 and nucleophilic sites in DNA. (Swenson gt gt'l975) -Cl STUDIES WITH AFB1 2 Comparative sensitivity to the cytotoxic effects of AFB -Cl l 2 Figure 19 shows the cytotoxic response of normal and XP cells to AFBl-Clz. For these studies, NF and XP cells were seeded and treated l situ at the same time with the same carcinogen solutions to insure uniformity of 92 DOSE AFBI-Clg (nIVI) ' I2 Is >_IOO F I I t: .. _—J_ 2 _ oso — g _ 6 _I _. .. 040 I _I s - ‘ - ,— "" A g: . k D :20- - 2 ‘ A LIJ O 0: LIJ 0. A I0 I L 'l I J J Figure 19. Cytotoxicity of AFB -C1 in NF and XP cells. Cells were treated it situ. (0) NF; (A) XP 93 concentration administered. The results showed that normal diploid human fibroblasts were less sensitive than XPlZBE cells to the cytotoxic effects of AFBl-Clz. These data suggest that normal cells are capable of excising more of the potentially cytotoxic lesions induced by AFB -Cl 1 2 than are XP cells. To test this hypothesis, we compared the sensitivity of a series of XP-derived cells to the toxicity induced by AFBl-Clz. Since each of the XP cell strains we examined had been shown to possess a different level of excision repair capacity for UV-induced DNA. (Robbins ‘gt ,gt.,l973), a correlation between their rate of Texcision repair of UV damage and their resistance to AFBl-Clz- induced toxicity would support the hypotheses that excision repair of the potentially cytotoxic lesion by NF cells was 1'C12' induced toxicity. It would also indicate that one or more responsible for their greater resistance to AFB steps were common to the repair of both kinds of lesions. In this study, normal cells and a series of XP cells derived from different complementation groups were exposed to varying concentrations of AFBl—Cl2 and assayed for survival 1 situ. The results are shown :hI Figure 20. XPlZBE cells, with less than 2% the normal rate of excision repair, showed the greatest sensitivity to AFBl-Cl2 of all cells tested. XP7BE cells from Group D, with an inter- mediate excision repair rate,were lesssensitive than XP12BE 94 199 I 1 I I I 1 r l '— l r O l I l PERCENT SURVIVAL CLONING ABILITY @ I I l I I 1 8 6 12 I8 CONCENTRATION AFB1-CLZ Figure 20. Cytotoxicity of AFB -C1 in three different XP strains and normal cells. (I) XP12BE; A XP7BE; (O) XP4BE; Iv) NF 95 cells, but more sensitive than XP4BE cells, which have a normal rate of excision repair, but have abnormal replica- tion past DNA lesions. All three of the XP strains tested showed a greater sensitivity to AFBl-Cl2 than NF cells. It appears from these data that the degree to which a cell strain is capable of repair of UV-induced lesions cor- relates with the ability of cells to survive AFB —Cl 1 2 treatment. The relationshtp between level of DNA binding and survival If normal cells are capable of excising potentially cytotoxic DNA lesions induced by AFBl-Cl2 more rapidly than are XP cells, we would expect that, if both sets of cells received equal initial binding levels, normal cells would exhibit a higher survival than XP cells because they would have been able to excise some of these lesions before the potentially lethal effect was fixed in the cell. To test this hypothesis, a series of confluent NF and XP cul- tures were treated with varying concentrations of tritiated AFBl-Cl2 and were immediately assayed for percent sur- vival and for the number of AFBl-Cl2 residues bound to DNA. Figure 21 shows the relationship between survival and the number of AFBl-Cl2 residues initially’ bound. to DNA in NF and XP cells. These data showed that for equal levels of initial DNA binding, NF cells showed a greater survival. 96 l8 I I fi 1 I8;- ._ I r PERCENT SURVIVAL CLONING ABILITY I I I I 2J3 8 8.5 l AI='B,_-CLz RESIDUES PER IIa6 NUCLEOTIDES Figure 21. Relationship between levels of cytotoxicity and DNA binding induced by AFBl-Cl2 in normal and XP cells. (0) NF; (A) XP 97 Although binding is measured immediately after treatment, survival is dependent on events in the cell cycle and is determined at some time later than initial carcinogen binding. The data suggest that the reason NF cells exhibit a higher survival than XP cells when their initial DNA bind- ing levels are equal is because, before the critical time for cell killing is reached, NF cells can excise a portion of the lesions, thus lowering the toxicity, while XP cells cannot. Rate of recovery of confluent cultures from the potential cytotoxicity induced by AFBl-Cl2 lesion(s) If normal cells are more resistant than XPlZBE cells to the killing effects of AFB -C12 because of their 1 ability to repair the cytotoxic lesion more rapidly than XP cells, they should be able to recover from the potentially cytotoxic damage of AFBl-Cl2 if they are given time to carry out the repair processes needed before the critical time in the cell cycle, which determines cell killing, is reached. To test this hypothesis, a series of NF and XP cells, which were in the Go state after being grown to confluence, were treated with various concentrations of carcinogen and assayed for survival at various times over a 10 day period following treatment. The results are shown in Figures 22a and 22b. Cultures of normal fibroblasts reveal- ed the ability to recover to a survival of up to 85% under 98 IOOOL I '9 ... >_ 600 _ L: _ :I “-2 4o< _ L9 g o 2 F ‘ O o’ a. g 20“ I I I I I I I I I I" ; I00 I I I I I I I I I I O: IE : 8 60A A A A. .. A a E 40; A - E F A A A 1 O. 20- A . AXPIZEE A IO *- . b - A A_. - L I I I I I I I I I O 2 4 6 8 IO DAYS HELD AT CONFLUENCE Figure 22. Recovery of confluent cells from the potential cytotoxicity of AFBl-Clz. 99 these conditions. Cells treated with a concentration of carcinogen which gave an initial survival of 7 60% exhibit- ed 85% survival by 4 days, while cells with an initial ~ survival of 40% required 8 days to reach 85% survival. ~ Cells with an initial survival of 35% showed a survival of only 70% after 10 days at confluence (Figure 22a). In contrast, XP cells exposed to three different concen- trations of carcinogen showed no significant change in survival over seven days (Figure 22b). These data strongly suggest that normal cells are capable of excising the potentially cytotoxic lesion(s) while XP cells are not. Comparative sensitivity to the mutagenic effects of AFB -Cl 1 2 In order to investigate the mutagenic effcts of AFBl-Cl2 in human cells in culture, normal and XPlZBE cells were treated with varying concentrations of AFB1- C12 and assayed for the frequency of induction of 6-thio- guanine resistant mutants. XPlZBE cells showed a markedly greater sensitivity to the mutagenic effects of AFB1- Cl than did NF cells (Figure 23 and Table 4). These data 2 suggest that it is the ability of NF cells to excise the mutagenic lesion(s) induced by AFBl—Cl2 that is re- sponsible for the greater resistance of NF cells to the mutagenic effects of AFBl-Clz. lOO Isa T711 flTj ITTT III] I _ I I I I I g h ‘ an E} - I a} IzsI— .— 0 F- .1 LIJ ’ “I _I -‘ - m _. < - A. Ioo — '— > I- :- L0 I- .. 9. r 4 5 7SI- - o- F- . -I m — u- E I— u- < I. .. e- sot— ‘- D I- -I Z a I 9'1 U h d Q ~ - E; 25.1 ‘— Z I- .- H . .. I) . . d 8 JIJJIJJIIJJJJIJIIII- 8 S 18 IS 28 CONCENTRATION AFBl-CLZ (nM) Figure 23. Induction of 6-thioguanine-resistant mutants by AFB -C12 in normal and XP cells. (OI NF; (A) xp 101 .OCauMamwu sh meow mum: mucweaummxo suaoax0uouso * mm mm.o vso. ws.a v aa .om as am.o aa. ws.a ma mm .va mm No.0 oa. os.a v av o.m c vo.o Na. ws.a o ooa 0 «m2 ma ~m.o sv. ws.a Na om ~.o oa ~m.o am. ws.a m so o.v va mm.o mv. ws.a ma om o.m s.v mm.o sm. ws.a m sea o m2 av o.a cm. wo.a m vm o.m mm mm.o om. oo.a w sv m.m ma o.a vw. oo.a v sm s.m o o.a sa. oo.a o ooa 0 «m2 cva o.a mmo. ws.a ma sm o.m maa o.a mmo. os.a am am m.a o o.a smo. ms.a o ooa o mmmamx A caxv mucousE no A caxv weacoaoo am>a>usm *znv socowoobu suo>oooO mo seawaoauum memaQ ucmuwawou ucoouom NaUI mm< cawuum coaumunz mocmaoamwm meaumammm maaoo we annamauo coaumuucoocoo aamu maawo ax can mz ca maUIammv so mucouse ucmumawoqucacwsmOaEUIm no cOauoopca one .v wanna 102 Rate of recovery of confluent cultures from the potentially mutagenic effects of AFB -C12 1 At the same times that survival was assayed in the above biological recovery experiments, normal cells were also assayed for induced nmtation frequencies, in order to determine whether they were capable of removing the poten- tially mutagenic DNA lesions induced by AFBl-Cl The 2. results of these studies are shown in Figure 24. The normal cells appeared to recover from the potentially mutagenic effects of AFBl-Clz, when held. in a: nondividing state. These data, together with the differential mutagenicity observed for NF and XP cells, suggest that normal cells are capable of excising the potentially mutagenic DNA lesion induced by AFB -Cl . l 2 Rate of loss of covalently bound AFBl-Cl2 residues from cellular DNA In the course of the biological recovery experiments described above, we also monitored the rate of loss of the covalently bound radioactive AFB -Cl2 residues from 1 human cell DNA by determining its specific activity at different times after treatment. At the same time that cells were harvested for assaying of survival and mutation frequency, a portion of cells (25-150 x106) were harvest- ed for isolation of cellular DNA, and the specific activity 103 .maOIamme mo suaoacwmmuss awaucouod on» scum maaoo amEuoc mo suw>ooom .vm mesmam wozwauuzoo h< Gum: w>