Biologic Markers in Reproductive Toxicology (1989)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

24 Introduction This section of the report examines many topics salient to the study of developmen- tal markers in persons more than 24 hours old. Rather than discuss in detail the development of the many systems and proc- esses of the human body, we focus the review on the development of the central nervous system. Then, using radiation as a paradigm, we examine the effects of timing and dose on central nervous system teratogenesis. The consequences of abnor- mal cell migration for host behavior are also examined. The critical topic of meth- odologic issues encountered in construct- ing inferences abut human effects from animal data is discussed next. Finally, neurodevelopmental outcomes that might be among the most sensitive markers of neurotoxicant-engendered dysfunction in humans are surveyed. Lead, the neuro- toxicant with the largest neurotoxicolog- ic data base, is used as a paradigm to ex- amine some of the methodologic issues in- evitably encountered in the pursuit of a valid collection of markers. Many discussions of biologic markers and development focus on using markers to ascertain effects of xenobiotic sub- stances on development. However, the in- fluence of growth and development on xeno- biotic disposition and effect is equally 265 important. Biologic markers that might be useful at one age might not be useful at other ages. Results of many assessments differ quan- titatively and qualitatively from one age group to another. That is particularly true of concentrations of substances in blood and pharmacokinetic (absorption, distribution, metabolism, and elimina- tion) characteristics in general. There- fore, interpretation of biologic markers of internal dose must take into account the developmental status of the subject. The pharmacodynamics (biochemical and physiologic effects of drugs and their mechanisms of action) of target-organ response and short- and long-term conse- quences of exposures also vary with age. From birth until 2 or 3 years of age, major developmental changes occur in the CNS and major changes occur in the regulation of the endocrine system during puberty. Thus, the end-organ effects in these sys- tems after exposure at different times of life are likely to be different, and the biologic markers used to assess the effects might be different. In neurodevelopment, the demonstration of sensitive, specific markers that clear- ly link exposure and outcome has been rare; many variables intervene between

266 toxicant exposure and demonstration of altered behavior. Demonstration of causal relationships between low-dose exposure and impairment has been particularly dif- ficult. Establishing causal relation- ships often depends on inferences drawn from the combined results of animal experi- ments and epidemiologic studies; neither might be sufficient in isolation. Animal studies raise the question of species dif- ferences; some human teratogens or neuro- toxicants might not cause measurable al- terations in commonly used species, and some animal teratogens have not been shown to be human teratogens. It is hard for epi- demiologic studies to establish causal links, owing to the complex exposure hlS- tories of the subjects to many compounds, inability to control exposure levels of the subjects, and the population-based statistical procedures. Relationships between phenotypic ef- fects of pollutants and behavioral altera- tions are increasingly apparent. The re- cent recognition that agents that affect physical development can be expressed at an early age in alterations of nervous system function—possibly as aberrant behavior-has given rise to a new disci- pline: behavioral teratology. Before beginning the discussion of neurodevelopmental markers themselves, we digress briefly to consider issues of developing and using biologic markers to assess the status of children, using as a paradigm the exposure to and effects of halogenated aromatic hydrocarbons. Our examples illustrate the need for careful integration of chemical, pharmacodynamic and pharmacokinetic, analytic, and cel- lular methods in epidemiologically sound studies to assess the multifaceted nature of health effects. BIOLOGIC MARKERS OF EXPOSURE: PHARMACOKINETIC CONSIDERATIONS Therapeutic agents usually are adminis- tered in known amounts for defined periods, but accurate quantitative his- tories of exposure to environmental chemi- cals are rare. Because polychlorinated biphenyls (PCBs) are found in most adipose NEURODE~ELOPMENTAL TOXICOLOGY tissue samples from randomly selected persons (Mes et al., 1982; Kreiss, 1985), it is difficult to design experiments to decide what degree of contamination con- stitutes significant risk. These consid- erations emphasize the value of intensive study of populations with massive expo- sures, very high body burdens, or evidence of chemically induced disease in the mother during gestation. Infants exposed in utero under such circumstances probably would manifest more readily detectable physical and biochemical findings and serve as mod- els for phenomena to be examined in off- spring of less-exposed populations. Many pharmacokinetic considerations are discussed in the previous part of the report. Here these are briefly discussed with regard to the specific situation of PCB exposures. Interpretation of concen- trations of a toxicant, or drug, depends on knowledge of the amount and timing of exposure relative to sampling of serum and route of administration. Highly sensitive and specific assays have been developed for various halogen- ated aromatic compounds (Albro et al., 1986~. Isomer-specific detection and quantification in the parts-per-trillion range are possible. The ability to deter- mine multiple compounds and their break- down products is critical, because expo- sures rarely involve pure substances; the blood of people reportedly exposed In Japan and Taiwan to PCB-contaminated rice oil contained products of thermal degradation, such as dibenzofurans (Chen et al., 1981; Kashimoto et al., 1981~. PCB isomers have different half-lives and different biologic effects with respect to enzyme induction (Safe et al., 1985), so interpretation of biologic markers of effects of these substances must account for chronic exposure and the presence of specific compounds at the time when the markers are examined. Although quantification of environmen- tal pollutants in biologic samples, in- c~uo~ng serum and tissues, is possible, pharmacokinetic interpretation of data remains difficult. Indeed, for such fat- soluble, extremely long-lived compounds, direct relationships among serum concen- tration (internal dose), concentrations ,

INTRODUCTION at receptor (biologically effective dose), and biologic effect are difficult to establish (Evans et al., 1986~. Furth- ermore, the distribution of lipophilic compounds can be altered by diet, disease states, and other drug or chemical exposure (Gibaldi and Perrier, 1982), as well as by developmental differences in metabol- ism, body size, and fat distribution from fetal life through adulthood. Continuous changes in composition and amounts of fat stores and altered metabol- ism in pregnant women and children would be expected to have profound effects on the availability of compounds like halo- genated aromatics to interact with recep- tors critical to their toxic effects and, consequently, affect the behavioral findings during childhood (Krauer et al., 1984; MacLeod and Radde, 1985~. Preg- nancy is associated with as much as a 25% increase in subcutaneous fat. Other phar- macokinetic changes during pregnancy include decreased and erratic absorption from the intestine, increased blood vol- ume, increased total body water, decreased protein binding due to a decrease in serum albumin concentration and an increase in endogenous displacing substances, in- creased glomerular filtration rate, and generally increased renal and hepatic clearance of many xenobiotics. Basing interpretation of body burden of a given lipid-soluble, slowly cleared compound on an isolated blood or serum concentration is extremely difficult. Similarly, fat stores in the developing fetus change mark- edly with gestational age. Body fat in- creases from less than 1% of body weight at 28 weeks to 15% at term. Total body bur- den in the newborn depends not only on the total maternal exposure, but on when that exposure took place and on the gestational age at birth. Mobilization of a compound from fat stores might increase its clearance via hepatic and renal mechanisms, but might also make more of it available at toxico- logically important sites. Sampling of fat from biopsy samples could provide a better index of total body stores, but this still might not reflect toxicologically relevant concentrations. Mobilization might be particularly high during lacta- 267 lion, increasing the clearance of lipid- soluble compounds from the mother, but also increasing delivery to a nursing in- fant. Complex pharmacokinetic mathemati- cal models could be helpful in assessing relevance of various tissues concentra- tions to body burden, quantification of exposure, and biologic effects. Such mod- els will have to incorporate additional complexity based on multifactorial deter- mination of outcomes. Establishing dose- response relationships in humans is further complicated by individual dif- ferences in biologic response to exposure (pharmacogenetics). BIOLOGIC MARKERS OF EFFECT: PHARMACODYNAMIC CONSIDERATIONS Toxic exposures that might affect the developing child can occur before or after birth. In assessing the relationship be- tween exposure and effects, one must study the mother, the neonate, and the child. Markers in the Mother Most studies of pharmacologic effects of pregnancy focus on pharmacokinetics. However, pharmacodynamics and suscepti- bility to toxicity might also differ. For example, susceptibility to tetracycline- induced hepatic injury increases marked- ly during gestation, owing to unknown mechanisms. Attempts to interpret data obtained from offspring of mothers exposed to poly- halogenated aromatics should include histories, physical examinations, and laboratory findings of the mothers. Again, people with massive exposure, such as those exposed to contaminated rice-cooking oil in Japan and Taiwan (Yusho and Yuchen) (Masuda et al., 1982; Lu and Wong, 1984~, or histories of maternal illness, should be studied. In humans, dermatologic find- ings-especially chloracne—have been the most consistent clinical feature among people with massive exposure to polyhalo- genated aromatics (Suskind, 1985~. His- tologic features of chloracne are not path- ognomonic of exposure, but the combination

268 of chloracne with measurement of exposure and of PCB concentrations in adipose tissue and plasma suggests chemical-induced disease. Other histologic findings in skin are not likely to be helpful in assess- ing the existence or extent of chemically mediated illness (Moses and Prioleau, 1985~. Other physical features and clin- ical laboratory findings associated with massive exposure include weight loss, porphyria, hepatic dysfunction, and peri- pheral neuropathies (Mocarelli et al., 1986; Kimbrough, 1987~. None is diagnos- tic of the PCB-induced disease. Target organs and toxic manifestations vary widely among species (Safe, 1986~. Recently, immunologic changes have been emphasized, particularly because thymic involution occurs in most animals suscep- tible to 2,3,7,8-tetrachlorodibenzo- p-dioxin (TCDD) toxicity. Subclinical changes in cell-mediated immunity have been reported in humans with possible TCDD exposure (Hoffman et al., 1986~; none of the subjects examined had other manifesta- tions of illness—including chloracne— and body burden of TCDD was not documented. Much work needs to be done in populations with clearly massive exposures, to verify whether any immunologic changes can serve as useful markers. No human data are avail- able about effects of TCDD exposure during pregnancy. Many effects of halogenated aromatics are mediated via a specific cytosolic re- ceptor (Roberts et al., 1985; Denison et al., 1986~. Induction of specific drug- metabolizing enzymes is the most widely recognized outcome of interaction of the compounds with receptors (Poland and Knut- son, 1982; Denison and Wilkinson, 1985; Denomme et al., 1986; Okey et al., 1986~. The toxicity of various halogenated aro- matics in different species is correlated with the affinities of various analogues for the receptor and with characteristics of the receptor in specific species and strains. Results of animal studies show that genetically determined receptor differences are correlated with induci- bility of specific cytochrome P-450 and with toxic outcomes of exposure to polycyc- lic aromatic hydrocarbons, including carcinogenesis and teratogenesis (Nebert NEURODE^LOPMENT~ TOMCOL~Y and Jensen, 1979~. Results of studies of human peripheral blood lymphocytes sug- gest that inducibility of arylhydrocarbon hydroxylase (AHH) is correlated with sus- ceptibility to lung cancer (Kellerman et al., 1973a,b; Kouri et al., 1984), but the mode of inheritance of inducibility in humans remains uncertain. The above considerations suggest that a variety of markers might be used to assess biologic effects of halogenated aromatic exposure. If susceptibility to adverse effects depends in part on the nature of the aryl hydrocarbon receptor (possibly under genetic control), biochemical and molecular analyses of differences in human receptors might be helpful in cor- relat~ng outcomes with exposures and in typing the population according to suscep- tibility. Such analyses could use human lymphocytes or skin cells. One strategy would be to include direct assessment of the presence and binding properties of the receptor, as well as molecular ap- proaches to the receptor gene and the whole AHH gene complex (Whitlock, 1986~. The major effect of most halogenated aromatic compounds is to induce a pattern of enzyme activity similar to that produced by 3-methylcholanthrene. Some PCB and polybrominated biphenyl (PBB) isomers, however, also induce enzymes typically induced by phenobarbital (Safe et al., 1985~. Enzyme induction in humans can be assessed with nontoxic, in vivo probes (e.g., caffeine for AHH), with examination of induction capacity in vitro in lympho- cytes (Kellerman et al., 1973a,b; Kouri et al., 1984), with direct measurement of enzymes in tissues (liver or skin biop- sies), and perhaps with molecular tech- niques, including quantification of mRNAs for specific forms of cytochrome P-450 (Jaiswal et al., 1985b). Enzyme activity can be influenced not only by genetic dif- ferences, but by a wide range of environ- mental exposures-such as cigarette smoke, diet (including consumption of charcoal- broiled foods or cruciferous vegetables), and medications (Okey et al., 1986~. Preg- nancy also affects activity of drug-metab- olizing enzymes (Krauer et al., 1984~. Interpretation of data obtained in the diverse human population requires

INTRODUCTION careful attention to experimental design- number of subjects and their clerical sta- tus, documentation of exposure, defini- tion of sensitivity and specificity of methods, and selection of control popu- lations. An ideal study would include a population with defined exposures (in- cluding large exposures), quantification of exposure in family members (especially fathers), determination of endogenous (perhaps genetic) and environmental vari- ables that might alter outcome, detailed medical and biochemical analyses of family members, and thorough study of pregnancy outcomes. Drug-metabolizing enzymes should be analyzed (particularly during pregnancy); the metabolism of hormones critical to pregnancy also would be of interest, because steroid hormone metab- olism might be altered by exposure to com- pounds such as TCDD (Okey et al., 1986~. Placental Markers Placental markers were discussed in the previous part of the report. Here, the markers relevant to PCB exposure are discussed in the context of the need to study population longitudinally. The placenta is obviously a critical organ in determining pregnancy outcomes; it is a potential target for environmental chemical toxicity and influences fetal exposure. No information is available from exposed people about placental tissue markers before delivery. Such techniques as chorionic villus biopsy (discussed previously) might prove useful in follow- ing pregnancies prospectively. The pla- centa early in gestation is different from the placenta at term; therefore, extrapo- lation of markers from different times in pregnancy should be undertaken cau- tiously. Similarly, vast differences among species in placental derivation and structure make cross-species extrapola- tion questionable. At-term placenta markers include quan- tification of halogenated aromatic com- pounds, determination of histopathologic changes, determination of enzyme induc- tion status, determination of presence and characteristics of aryl hydrocarbon receptor. Considerable data are available 269 on at-term placentas of women exposed to contaminated rice oil in Taiwan; AHH ac- tivity and 7-ethoxycoumarin-O-deethylase (7-ECOD) activity were markedly increased in placentas from such pregnancies (Won" et al., 1985a), even though exposures oc- curred 3-4 years before pregnancy. The induction by other factors, such as cigar- ette-smoking (Fujino et al., 1984), had to be ruled out. In the rice-oil-exposed population, the AHH induction was greater than expected, as determined by the meta- bolic profile of benzoapyrene in placen- tal homogenates. Cotinine and thiocyanate were used as markers to ensure that the subjects were not smokers and were not exposed to substantial passive smoking. Some methodologic issues in studying pla- cental tissue have been discussed (Won" etal., 1985a). In more recent studies, specific cyto- chrome anti-P-450 isozyme 6 rabbit anti- bodies were used to detect an increase in one P-450 protein in microsomes from the placentas of women exposed to contamin- ated rice oil (T.K. Wong et al., 1986~. Thus, direct measurement of P-450 activi- ties, use of monoclonal antibodies, and perhaps direct molecular probes are prom- ising markers of effects of exposure to enzyme inducers during late pregnancy. Neither maternal blood PCB concentra- tion nor clinical symptoms were correlated with AHH or 7-ECOD induction in the study (T.K. Wong et al., 1986~. The placenta of one of the nine mothers with clear ex- nosure and substantial clinical symptoms showed no induction. That is interesting, with respect to study of other environmen- tal variables or individual differences in susceptibility related to endogenous and perhaps genetic mechanisms. Aryl hy- drocarbon receptor has been measured in human placentas (Manchester et al., 1987), and correlations of induction with recep- tor characteristics will be of interest. Similarly, study of adducts of reactive metabolites—particularly of products of cigarette smoke-holds promise of quan- tifying exposures that lead to toxicolog- ically relevant tissue interactions. Placental enzyme induction by compounds that are metabolized to electrophilic toxicants (such as some components of

270 cigarette smoke) might protect the fetus from exposure (Manchester and Jacoby, 1984; Manchester et al., 1984~. High in- duction and placental metabolism might prevent compounds from crossing the pla- centa. Umbilical vein endothelial cells from placentas of women who smoke do not show enzyme induction in situ, but retain the capacity for in vitro induction (Man- chester and Jacoby, 1984; Manchester et al., 1984~. Similarly, decreased placen- tal enzyme induction appears to be corre- lated with an increased risk of birth de- fects. However, it is uncertain whether that is true for nonmetabolized inducers, such as PCBs or TCDD, in which case induc- tion of placental enzymes might not in- crease first-pass clearance of the com- pound. Studies have examined the roles of genetics and environmental variables on placental response to cigarette-smoke- mediated enzyme induction (Gottlieb and Manchester, 1986~. Studies of placentas from pregnancies that produce identical and fraternal twins show that heredity has a role in determining outcome, and position of the placentas in utero might contribute greatly to the extent of induc- tion. Little information is available to as- sess and predict interactions of markers of placental exposure, receptors, enzyme induction, function (including steroid hormone metabolism), anatomy (including pathologic changes), and pregnancy out- comes. An ideal approach integrates as- sessment of each of these factors; it is unlikely that any marker will adequately predict outcomes that are multifactorial. Intensive study of placental tissue— including potentially early -biopsies and at-term perfusion models-could lead to an integrated picture of critical determinants of outcome and their contributions. Markers in the Newborn The following discusses markers men- tioned in the previous part of the report specifically as they relate to assessing effects of PCB exposure. Newborns are unique pharmacokinetically and pharmacodynamically. Their fat con- NEURODEVELOPMENTAL TOXICOLOGY tent, total body and extracellular water, protein binding, and renal and hepatic clearance of compounds vary widely with gestational age and can be influenced by many prenatal factors (e.g., nutrition, disease, genetics, medication, and en- vironmental exposures) (Boreus, 1973~. Unique interaction of variables can lead to unusual susceptibility to toxic ef- fects that do not occur in other age groups, such as kernicterus resulting from in- creased red-cell turnover (Karp et al., 1985), decreased protein binding of bili- rubin (Karp et al., 1984), decreased hepatic capacity to metabolize bilirubin (Karp et al., 1984), and increased entry of bilirubin into the CNS because of im- maturity of the blood-brain barrier (Karp et al., 1984~. Newborns differ markedly from older subjects in organ sizes and physiologic roles (e.g., the fetal adrenal gland is a major drug-metabolizing organ). These considerations suggest that new- borns differ from children and adults in responses to a variety of exposures and require different interpretations of bio- logic markers—or different markers—of internal dose and effect. Many data are available on kinetic differences, but more work is needed on receptor maturation, response to enzyme inducers, and pharmaco- dynamic differences, including unique CNS responses of the immature brain; e.g., phenobarbital and antihistamines produce excessive excitation in infants and chil- dren and sedation in adults at comparable blood concentrations (MacLeod and Radde, 1985~. Physical examination of newborns ex- posed to halogenated aromatic compounds should be comprehensive and quantitative (including weight, height, head circum- ference, and such other measurements as internal and external canthal distances), and should be performed by persons un- aware of the nature of the exposure. Ap- propriate control subjects should be ex- amined simultaneously. Several features are possible sequelae of in utero exposure (Rogan, 1982), includ- ing low birthweight, conJunct~v~t~s with enlarged sebaceous glands in the eye- lids, natal teeth, pigmentary changes of gums and skin with deformed and pig-

INTRODUCTION mented nails, peculiar skin coloration, and chloracne. The findings of important dermatologic effects are interesting, given the similarity of sites of toxic effects in adults. Later necrologic devel- opment in children with these effects might also be abnormal (Harada, 1976), but such findings require further verification. The possible effects on immune function that appear in adults after TCDD exposure and effects in animals suggest developmen- tal immunologic abnormalities (Lubes et al., 1984; Silkworth et al., 1984) and need to be evaluated in newborns and children. Other than physical findings, few mark- ers of exposure have been documented. Cord blood concentrations of environmen- tal chemicals could be obtained and corre- lated with maternal and placental concen- trations. A great deal is known about developmental patterns for a variety of drug-metabolizing enzymes in humans and the effects of inducing substances, such as the effects of phenobarbital on glucuro- nyl transferase (Catz and Yaffe, 1968~. Probes, such as caffeine, would be worth using to assess the status of AHH-like cytochrome P-450 activity in infants with possible exposure to halogenated aromatics. Caffeine metabolism and clear- ance in the newborn are extremely limited and can be followed with a C-labeled probe for breath testing (to estimate oxidative demethylation) or with analysis of specif- ic urinary metabolites that reflect prob- able cytochrome P~-450 activity (Lambert et al., 1986; Campbell et al., 1987~. De- termination of aryl hydrocarbon receptor and P~-450 induction by molecular tech- niques in available cell types will add to a comprehensive picture of the effects of prenatal exposure on newborn drug- metabolizing capacity. Appropriate psy- chometric testing of newborns also should be performed for later comparison with results of followup studies. Markers in Children The extremely long half-lives of many compounds of interest warrant careful, long-term followup study of the conse- quences of exposure. Content of environ- mental chemicals in breast milk should 271 be documented if an infant is breast-fed. Concentrations might vary during a given breast-feeding and over an extended period as the maternal stores of lipid-soluble compounds are redistributed into breast milk. Accurate clinical descriptions of the skin with photographic records and his- topathologic examination are necessary. Routine growth characteristics and ap- propriate assessments of CNS development should be documented. Given the current focus on immunologic effects of halogen- ated aromatic compounds, knowledge of responses to routine childhood immuniza- tions and documentation of histories of infectious disease, allergy, and auto- immune disease will be helpful. The roles of environmental exposures and altered immunologic status (Marshall, 1986) are controversial, and verifiable markers and clinically accurate descriptions of outcomes are badly needed. The ultimate goal of studies of the long-term consequences of in utero expo- sure to environmental toxicants is to as- sess the effects of growth and development on handling of and response to the toxi- cants and the effects of the toxicants on the very processes of ontogeny. Standard estimates of half-lives of various halo- genated aromatic compounds are typically based on sparse data on adults. The devel- oping child changes dramatically in body composition, and the change likely has major effects on distribution and clear- ance of lipid-soluble compounds. Renal and hepatic clearance rates of most com- pounds generally are greater in children (until puberty) than in adults (MacLeod and Radde, 1985~. Kinetics of environmental contaminants depend on what compounds a subject is ex- posed to over a long period. For example, enzyme induction could result from persis- tence of the inducing substance or from an effect manifested long after exposure. The latter situation is exemplified by changes in steroid-hormone metabolism at puberty in rats that received one dose of phenobarbital during gestation (Gupta and Yaffe, 1981~. It is uncertain whether a similar phenomenon occurs in humans, but long-term effects caused by short-

272 exposures are possible; any effects noted in children who were exposed prenatally to halogenated aromatic compounds will have to be separated into effects of chron- ic exposures to compounds with very long half-lives and persistent effects of pre- vious acute exposures. A safe, noninvasive approach to assess- ing in viva phenotype with respect to cyto- chrome P' -450 induction would be extremely useful as a tool for comparing development in exposed and control populations. Oxida- tive demethylation of caffeine is mediated by cytochrome P~-450 and can be monitored with breath tests that use carbon-13 and measurements of urinary-metabolite ratios (Lambert et al., 1986; Campbell et al., 1987~. Such a procedure can be repeated on the same subjects. For example, metab- olism is high in children after the neo- natal period and later declines to values characteristic of adults; patterns of metabolism at the time of puberty differ between males and females (Lambert et al., 1986~. The relatively rapid P-450-medi- ated clearance of most compounds in prepu- NEURODEVELOPMENTAL TOXICOLOGY bertal children might make it more diffi- cult to assess the effects of enzyme inducers. Many children exhibit caffeine clearance as great as that in adults in whom it is maximal because it is induced by cigarette-smoking (Campbell et al., 1987~. Some of the most dramatic effects of exposure to environmental inducers might be noted at puberty, when drug-metab- olism rates normally decline to adult values. The simultaneous observation of symp- toms, monitoring of persistent specific chemicals, assessment of inducibility of AHH in lymphocytes, and continual moni- toring of P~-450 metabolism status will allow determination of the health implica- tions of exposures and induction. Growing understanding of individual differences in inducibility, perhaps related to dif- ferences in aryl hydrocarbon receptor characteristics, might make it possible to divide the population with respect to these variables and health outcomes, in- cluding long-range susceptibility to cancer.