Genetic differences in susceptibility to chemically induced myelotoxicity and leukemia.

The Ah locus represents a complex "cluster" of genese controlling the induction of numerous drug-metabolizing enzyme "activities" by polycyclic aromatic compounds. Allelic differences at the Ah locus are reflected in the large differences in inducibility of cytochrome P1-450 and benzo[a]pyrene metabolism in numerous tissues when the mice receive the chemical daily in their diet. This experimental model system offers to the hematologist and clinical pharmacologist a means to study genetic differences in toxic chemical depression of the bone marrow, as well as a potential model to study aplastic anemia and leukemia explainable on a single-gene basis. The genetically "responsive" individual who is at increased risk for cancer caused by subcutaneous or topical or intratracheal polycyclic hydrocarbons is at decreased risk for toxicity of the bone marrow and leukemia caused by oral benzo[a]pyrene (when compared with the genetically "nonresponsive" individual receiving the same dose of the same xenobiotic). In other words, tissue sites in direct contact with the carcinogen develop cancer in responsive animals because of induced P1-450; tissues in distant sites of the body may develop malignancy in nonresponsive animals because more carcinogen reaches that tissue due to decreased P1-450 induction all over the body and therefore decreased detoxication. Not only the dct with the carcinogen develop cancer in responsive animals because of induced P1-450; tissues in distant sites of the body may develop malignancy in nonresponsive animals because more carcinogen reaches that tissue due to decreased P1-450 induction all over the body and therefore decreased detoxication. Not only the dct with the carcinogen develop cancer in responsive animals because of induced P1-450; tissues in distant sites of the body may develop malignancy in nonresponsive animals because more carcinogen reaches that tissue due to decreased P1-450 induction all over the body and therefore decreased detoxication. Not only the dose but the route of administration and the tissue in which the malignancy or toxicity develops are therefore very important in the interpretation of data from tumorigenesis or toxicity experiments involving P1-450 inducers such as polycyclic hydrocarbons. There exists sufficient evidence that heritable variation of the Ah locus occurs in man. Growing evidence indicates that persons with higher aryl hydrocarbon hydroxylase inducibility in their cultured mitogen-activated lymphocytes may have a statistically significantly increased risk for certain types of cancer and drug toxicity. It remains to be determined at the present time, however, whether this genotype can be used as a biochemical marker in the individual patient for predicting increased susceptibility to certain types of environmentally caused cancers or toxicity in man.


Introduction
To study the genetic control of drug metabolism is often called pharmacogenetics. In a single sentence, pharmacogenetics may be defined as the attempt to understand why the same dose of the same drug given to two different individuals (with the possible exception of identical twins) may cause widely varying responses. These responses include *Developmental Pharmacology Branch, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland 20205.
June 1981 therapeutic effects of a drug, e.g., anticoagulation or control of seizures, but also unwanted deleterious effects such as cancer or drug toxicity. The experimental system to be examined in detail in this chapter represents principally a genetic difference in receptor concentration; because of this defect, there are large genetic differences in the biotransformation and pharmacokinetics of certain drugs and other environmental pollutants, resulting in important differences in risk toward cancer, drug toxicity, mutation, and birth defects.
The general characteristics of the P-450-mediated monooxygenases and their coordinated enzymes 11 are first described. Secondly, the genetic differences in this model system in mice are examined. How these differences are associated with increased risk toward myelotoxicity and leukemia are then shown as examples. Numerous other conditions in mice associated with this genetic system are also listed. Lastly, current evidence for this genetic difference in man is briefly assessed.

Cytochrome P-450 Monooxygenases and Coordinated Enzymes
Many environmental pollutants and other foreign compounds are chemicals that are so hydrophobic they would remain in the body indefinitely were it not for the metabolism resulting in more polar derivatives. These drug-metabolizing enzyme systems, which are localized principally in the liver, are usually divided into two groups: phase I and phase II. During phase I metabolism, one or more polar groups (such as hydroxyl) are introduced into the hydrophobic parent molecule, thus allowing a handle, or position, for the phase II conjugating enzymes (such as UDP glucuronosyltransferase) to attack. The conjugated products are sufficiently polar, so that these detoxified chemicals are now excreted from the cell and from the body (1).
One of the most interesting of the phase I enzyme systems is a group of enzymes known collectively as the cytochrome P-450-mediated monooxygenases. * The genetic relationship between these inducible enzymes and cancer or toxicity has been reviewed recently (3). These membrane-bound enzyme systems are known to metabolize: polycyclic aromatic hydrocarbons such as benzo[a]pyrene (BP) (ubiquitous in city smog, cigarette smoke and charcoalcooked foods) and biphenyl; halogenated hydrocarbons such as polychlorinated and polybrominated biphenyls, insecticides, and ingredients in soaps and deodorants; strong mutagens such as N-methyl-N'-nitro-N-nitrosoguanidine and nitrosamines; aminoazo dyes and diazo compounds; N-acetylarylamines and nitrofurans; numerous aromatic amines, such as those found in hair dyes; nitro aromatics, and heterocyclics; wood terpenes; epoxides; carbamates; alkyl halides; safrole derivatives; certain fungal toxins and antibiotics; many of the chemo-*Cytochrome P-450 is defined as all forms of CO-binding hemoproteins associated with membrane-bound NADPHdependent monooxygenase activities. We define cytochrome P1-450 as all forms of CO-binding hemoprotein that increase in amount concomitantly with rises in induced AHH activity following polycyclic aromatic inducer treatment. In view of more than one such form of P,-450 (2), it is emphasized that this definition of PI-450 is simplistic. 12 therapeutic agents used to treat human cancer; most drugs; small chemicals such as benzene, thiocyanate, or ethanol; both endogenous and synthetic steroids; and other endogenous compounds such as biogenic amines, indoles, thyroxine, and fatty acids.
Evidence is growing that metabolism to reactive intermediates by cytochrome P-450-mediated monooxygenases is a prerequisite for mutagenesis, carcinogenesis, and toxicity caused by numerous drugs, polycyclic hydrocarbons, and other environmental pollutants. These reactive intermediates probably bind covalently to numerous cellular macromolecules. Most of this binding is probably random, but some may be nonrandom, i.e., specific binding dependent upon the chemical structures of the reactive intermediate and the cellular macromolecule. Among these various types of covalent binding, there probably exists a very small amount of important binding of the ultimate carcinogen to its critical subcellular target, thereby initiating tumorigenesis. Two examples of apparent specific binding include the binding of BP 7,8-diol-9,10epoxide to the 2N-amino of guanine (4) and of aflatoxin B1 2,3-oxide to the 7N of guanine (5).
The steady-state levels of these reactive electrophilic intermediates and, consequently, the rates at which they interact with the critical nucleophilic target are dependent upon a delicate balance between their generation and detoxication (Fig. 1). Changes in the balance between toxification and detoxication in any particular tissue of an individual may therefore affect his risk of tumorigenesis or toxicity.
The Ah Locus: Genetic Expression of Induced AHH Activity and Cytochrome P1-450 Induction The Ah locus is an experimental model system that has provided several good examples of a delicate balance between genetic and environmental factors in the etiology of cancer, drug toxicity, and birth defects (2). The Ah locus of the mouse regulates the induction (by polycyclic aromatic compounds such as 3-methylcholanthrene, BP, or 2,3,7,8-tetrachlorodibenzo-p-dioxin) of numerous drug-metabolizing enzyme "activities" associated with several new induced forms of cytochrome P1450. The induction of aryl hydrocarbon hydroxylase (AHH) activity and more than 20 other monooxygenase activities and associated P1-450 occurs in 3-methylcholanthrene-treated B6 (the inbred C57BL/6N mouse strain) and other genetically "responsive" inbred strains and is absent or always Environmental Health Perspectives much lower in 3-methylcholanthrene-treated D2 (the inbred DBA/2N mouse strain) and other genetically "nonresponsive" strains (at any given dose of inducer). Besides the liver, this genetic expression is seen in such tissues as lung, kidney, intestine, lymph nodes, skin, bone marrow, pigmented epithelium of the retina, brain, mammary gland, uterus, ovary, and testis. The genetic response is therefore called "systemic," or occurring throughout virtually all tissues of the animal. Responsiveness to aromatic hydrocarbons has been designated the Ah complex: Ahb is the dominant allele; Ahd is the recessive allele; the AhblAh(d heterozgote is phenotypically similar to the Ahb/Ahb mouse in terms of degree of responsiveness ( Fig. 2). Several studies indicate that the fundamental genetic difference is in the regulatory Ah gene, Ahb/Ahd X Ahb/Ahd Ahb/Ahd X Ahd/Ahd F2 Ahb/Ahb:Ahb/Ahd Ahb/Ahd:Ahd/Ahd Ahb/Ahd:Ahd/Ahd capable of binding to inducers such as 3-methylcholanthrene, BP, and 2,3,7,8-tetrachlorodibenzo-pdioxin. To our knowledge, only foreign chemicals bind to this receptor with high affinity (less than 1 nM). The B6 mouse appears to have at least 50 times more receptor (and/or increased affinity toward inducers of P1-450) than the Ah"'IAhV' mouse; translocation of the inducer-receptor complex into the nucleus has now been demonstrated in the phenotypically responsive heterozygote and homozygote (8). What happens in the nucleus is not yet known, but somehow the "message" (that these inducers of P1-450 exist in the cell's microenvironment) is received; the response is transcription of specific mRNA's, translation of these mRNA's into specific enzymes such as P1-450, and incorporation of P1-450 into cellular membranes. These induced enzymes may aid in detoxication or they may generate increased amounts ofreactive intermediates.

Genetic Differences in Myelotoxicity
Large doses of oral BP (100 to 125 mg/kg/day) produce bone marrow toxicity in AhdlAh(I mice, whereas the Ahb/Ahb and Ahb/Ah d individuals are extremely resistant to oral BP-induced marrow presently no experimental evidence demonstrating unequivocally the subcellular location of a "critical target(s)" required for the initiation of drug toxicity or cancer or, for that matter, whether the "target" is nucleic acid or protein. Reproduced with permission from Dr. W. Junk Publishers. toxicity (9). Figure 4 illustrates the lethal effects of high doses of oral BP in Ahd/Ah(d mice. Concomitant oral phenobarbital treatment protects the AhdlAhd individual from oral BP toxicity, probably by inducing various drug-conjugating enzyme activities. Concomitant oral a-naphthoflavone treatment protects the Ahd/Ahd individual, presumably by inhibiting P1-450-mediated metabolism (in the myeloid precursor cells of the marrow) so that a decreased amount of toxic BP intermediates can be generated. These observations are supported by the markedly greater amount of radiolabeled BP (Fig. 5) which enters the marrow and which becomes metabolized and covalently bound in the marrow of the Ahd/Ahd mouse, compared with that of the AhblAhd mouse receiving the same diet.
14 Effect of Oral BP on Induced BP Metabolism Table 1 shows that daily doses of oral BP induced AHH activity in the Ahb/Ahd heterozygote more than 800-fold in bowel, approximately 3-fold in liver, and more than 16-fold in bone marrow. Daily doses of oral BP induced AHH activity in the AhdlAhd mouse about 50-fold in bowel and more than 7-fold in marrow, but a decrease in AHH activity was seen in liver. Further, the rate of increase in AHH activity as a function of days in mice receiving the BP diet was much slower in the nonresponsive Ahb/Ahd than in the responsive  BP metabolism in the bowel, liver, an both Ahb/Ahd and Ahd/Ahd mice. I oa-naphthoflavone treatment, on the otl not increase AHH activity in any of t] The data suggest than phenobarbital ] oral BP toxicity is caused by enzym June 1981 whereas a-naphthoflavone protection is caused by inhibition of BP metabolism.
Length of BP Exposure and Subsequent Appearance of Myelotoxicity Between 3 and 5 days of continuous oral BP (120 Barb, i.p. + BP mg/kg/day) was required to cause aplastic anemia * ._ ( Fig. 6): none died when exposed for only 2 days; 20% died when exposed for 3 days; about 83% died when exposed for 4 days; and 100% died when PhBarb, po. difference in radiosensitivity between normal (wiw) ver the 50-day and anemic (WIW) mice, however, has been reported iMarcel Dekker, (11) and is caused by differences in regeneration capability of erythropoietic tissue.] This type of latent effect (Fig. 6) is therefore Id marrow of distinctly different from that seen with chloram-Concomitant phenicol-induced aplastic anemia in man (12). To ier hand, did date, the calf is the only experimental animal in hese tissues. which aplastic anemia can be consistently produced protection of by a chemical after a latency period. In this case, ie induction, the agent is S-(1,2-dichlorovinyl)-L-cysteine (13). comprise some metabolites of BP, but more than 90% represents the inonmetabolized parent drug. Tissue samples were combined from groups of five or six mice (10). Reproduced with permission from Marcel Dekker, Inc.

Size of Oral BP Dose and Onset of P1-450-Mediated Leukemogenesis
Although massive doses of 100 or 125 mg BP ingested/kg/day produce bone marrow toxicity and death in 100% of AhdlAh(I mice in less than 4 weeks, no responsive AhblAhb or AhblAh,( mouse develops aplastic anemia even when this dose is continued for 6 months (14). Because these are such large doses of BP, we wondered how small a dose of oral BP would still cause an effect associated with the Ah locus. Figure 7 shows the results of groups of 30 Ahd/AhdI or Ahb/Ahd mice which received estimated doses of 12 or 6 mg BP/kg/day. Differences in weight gain attributed to allelic differences at the Ah locus were detectable. To our surprise, however, the mice that became ill and began dying did not have hypoplastic or aplastic bone marrow but rather developed hematopoietic neoplasms, espe-16 cially of the lymph nodes, spleen and thymus. No increased incidence of leukemia or differences in weight gain between AhdlAhd and Ahb/Ah" mice were found at estimated doses of 1.2 mg of BP/kg/day in the diet for 240 days (data not illustrated). When a-naphthoflavone was added to the diet at a dose 20 times greater than that of BP (Fig. 7), the incidence of leukemia was prevented almost completely, and the general health of the Ah"'lAh(' mice remained as good as that of Ahb/Ah(' mice receiving 12 mg BP/kg/day. These data suggest that a-naphthoflavone-sensitive metabolism of BPpresumably cytochrome Pl-450-in the bone marrow of AhdlAh' individuals is responsible for producing the reticuloendothelial malignancies. AhdlAhd mice are more susceptible than Ahb/Ah('l mice to leukemia produced by percutaneously applied 3-methylcholanthrene (18). Obviously the presence or absence of murine leukemia virus expressed by the various inbred strains (19) will modify the response elicited by P1-450-mediated metabolism of polycyclic hydrocarbons under the control of the Ah locus.
Environmental Health Perspectives  aGroups of Ahb/AhdI and Ah(llAhd mice were placed on control, BP, BP plus phenobarbital, or BP plus x-naphthoflavone regimens for the indicated number of days (10). Microsomal fractions were prepared from the indicated tissues combined from groups of five or six mice, and AHH activity was determined. the occurrence or lack of occurrence of terminal aplastic anemia occurring several weeks later in Ah"lAh" mice (9). From top to bottom, groups of 30 mice each received oral BP 2, 3, 4, 5, 7, and 10 days, respectively, following which normal diet was reinstated. Deaths, histologically confirmed to be associated with hypoplastic bone marrow, occurred in 0/30, 6/30, 25/30, 30/30, 30/30, and 30/30, respectively. Following cessation of the oral BP and return to the regular diet, the mean time for the mouse to die from aplastic anemia was about 23, 16, and 3 days for the groups exposed to 5, 7, and 10 days of oral BP, respectively. All mice that were alive on day 37 of the experiment remained alive at 60 days, at which time the experiment was stopped. Reproduced with permission from Springer-Verlag.

June 1981
Protective Barrier by the Ah-Responsive Intestinal Epithelium BP treatment (30 F.g/ml of growth medium) is much more toxic to Ahb/Ahd marrow cells than Ahd/Add marrow cells in culture (unpublished data). When AhdlAhd mice having transplanted Ahb/Ahb marrow are given oral BP (100 mg/kg/day), their death rate is similar to sham-treated AhV'/AhO' mice with Ahd/Ahd marrow; Ahb/Ahb mice having transplanted AhdlAhd marrow are just as resistant to oral BP daily as sham-treated Ahb/Ahb mice with Ahb/Ahb marrow (20). We therefore conclude that the Ah-responsive intestine (and/or liver) is important in protecting the individual. If the target marrow cells are exposed directly to BP in culture, the cells having the higher levels of induced PI-450 are more prone to BP toxicity. Despite the genetic origin of the bone marrow, the mice having the Ah /Ahd intestine and liver are more prone to develop aplastic anemia following oral BP.

Importance of the Route of Administration
In sum, the picture which has begun to emerge from numerous studies is categorized in Table 2. When the carcinogen (or other toxic drug) is placed in relatively direct contact with the tissue being studied, the genetically responsive AhblAhb or  (16). Ten days later, the BP diet, prepared as described previously (14), was begun. In the case of x-naphthoflavone (ANF) (right), approximately 120 mg of a-naphthoflavone/kg/day was included with the 6 mg of BP/kg/day. Wasted animals were killed and studied when it was ju(dged that they probably would not live more than 1-2 days longer. We are grateful to Drs. Lawrence, Corash, Michael M. Orlando and Alan S. Rabson for their advice about performing autopsies and examining histological sections of lymph nodes, spleen, thymus, bone marrow, kidney and liver. Whole blood counts were not especially helpful in the diagnosis of hematopoietic tumors. Lymphocytic leukemias, apparent stem-cell leukemias, and reticulum-cell neoplasms were all scored as "leukemia," according to the classification and description by Murphy (17). At an estimated 12 mg of BP/kg/day (left), all Ah"lAh" mice died before 110 days on the diet; none of the starting 30 AhblAh"I had died by day 100, and three had died after 150 (lays. At an estimated 6 mg of BP/kg/day (center), 24 of the starting 30 Ah"/Ah" mice and two of the starting 30 Ahh/Ah" mice had died after 240 days on the diet. At an estimated 6 mg of BP/kg/day (right), 19 of the starting 30 not receiving a-naphthoflavone had died, and four of the starting 30 receiving a-naphthoflavone had died after 240 days on the diet. Reproduced with permission from Pergamon Press Ltd.
AhblAhd mouse is at increased risk for developing a tumor or toxicity in that tissue, compared with the nonresponsive AhdlAhd receiving the same dose of xenobiotic (Fig. 8). On the other hand, if the malignancy or toxicity is found at a site distant from the administered drug, the AhdlAhd mouse is at increased risk, compared with the Ahb/Ah" or Ahd/Ahd individual receiving the same dose of xenobiotic. In this latter case, we believe the data are explainable by the "first-pass effect," also termed "presystemic drug elimination" (31). Fundamentally, presystemic elimination reflects the metabolism and excretion of a drug before the drug reaches its site of action. How BP metabolism in the intestine can be induced 400to 800-fold by oral BP-yet not exhibit any apparent toxicity (9, 10)-is not clear; an increase in conjugating enzymes or mechanism of efficient excretion of toxic metabolites must be involved. It will be of interest to see if the AhdlAhd mouse is more susceptible than the Ahb/Ahd mouse to in utero fetal toxicity or 18 primordial oocyte depletion, when the polycyclic hydrocarbon is administered daily in the diet. The data summarized in this report demonstrate that Pl-450 induction represents a double-edged sword. Therefore, in all cancer and toxicity experiments, the dose and especially the route of administration and the tissue in which the malignancy or toxicity develops are all very important factors in the interpretation of the observations.

Evidence of the Ah Locus in the Human
Lindane (32)(33)(34), other insecticides (32,33), various anticancer chemotherapeutic agents (35), and chloramphenicol (36) have all been implicated in the cause of certain aplastic anemias in man. To prove that a drug or chemical is the direct cause of aplastic anemia has always been difficult in clinical medicine, and most cases remain categorized as Environmental Health Perspectives  idiopathic. Almost all of these agents mentioned peripheral lymphocytes have been cultured in the require P-450-mediated metabolism either for depresence of mitogens and an inducer of AHH toxication or for metabolic potentiation to attain the activity such as 3-methylcholanthrene, in order to desired pharmacological effect. Chloramphenicol assess the human Ah phenotype. In spite of the and p,p'-DDT toxicity are not associated with the shortcomings with this assay method reviewed in Ah locus (unpublished data). We suggest that ref. (37), a growing list of clinical disorders (Table genetic differences between inbred strains of mice-3) appears to be associated with the human Ah with respect to marrow toxicity caused by these locus. various agents known (or suspected) to cause There clearly exists sufficient evidence that heriaplastic anemia in man-might be developed suc-table variation of AHH inducibility occurs in man. cessfully as a useful laboratory animal experimental Experimental difficulties, however, make it imposmodel. Needless to say, such a model should help sible at this time to be certain of whether AHH define the etiologic mechanisms, and thereby a induction is controlled by a single genetic locus or better understanding about treatment and preven-by two or more loci (i.e., polygenic). Until one can tion, for certain human aplastic anemias.
increase the range of fold inducibility of AHH With the use of 20 to 40 cc of drawn blood, activity and/or decrease the magnitude of day-to-day Table 3. Human disorders that appear to be associated with the Ah locus.