Neurotoxicologyand Teratology,Vol. 12, pp. 175-181. PergamonPress plc. 1990. Printed in the U.S.A.
0892-0362/90 $3.00 + .00
Scientific and Regulatory Issues Relevant to Assessing Risk for Developmental Neurotoxicity: An Overview D. C O O P E R R E E S , .2 E L A I N E Z. F R A N C I S t A N D C A R O L E A. K I M M E L t 3
*Health and Environmental Review Division, Office of Toxic Substances Office of Pesticides and Toxic Substances and ~Reproductive and Developmental Toxicology Branch/HHAG Office of Health and Environmental Assessment, Office of Research and Development United States Environmental Protection Agency, Washington, DC 24060
REES, D. C., E. Z. FRANCIS AND C. A. KIMMEL. Scientific and regulatory issues relevant to assessing riskfor developmental neurotoxicity: An overview. NEUROTOXICOL TERATOL 12(3) 175-181, 1990.--The effects of chemical exposure on the developing nervous system have been documented in both humans and animals for a variety of agents. However, the comparability of these effects has not been carefully evaluated to determine the predictability of animal models to adverse effects in humans. A workshop sponsored by the U.S. Environmental Protection Agency (EPA) and the National Institute on Drug Abuse was held on April 11-13, 1989, to address the Qualitative and Quantitative Comparability of Human and Animal Developmental Neurotoxicity. Invited experts were asked to review the human and animal data on several agents that are known to cause developmental neurotoxicity in humans, including lead, methylmercury, selected abused agents, anticonvulsants, polychlorinated biphenyls (PCBs), ethanol and X-irradiation, and to make qualitative comparisons on a specific end point basis as well as on a functional category basis. In addition, they were asked to make quantitative comparisons when adequate dose-effect data were available. The data also were evaluated in the context of the proposed EPA developmental neurotoxicity testing battery to determine whether or not the battery would adequately detect the effects of each agent. Finally, four work groups were asked to reach consensus on issues relating to: 1) comparability of end points across species for developmental neurotoxicity; 2) testing methods in developmental neurotoxicity for use in human risk assessment; 3) weight-of-evidence and quantitative evaluation of data from developmental neurotoxicity studies; and 4) triggers for developmental neurotoxicity testing. Risk assessment
Developmental neurotoxicity
TSCA
FIFRA
LITERATURE in the field of behavioral teratology/developmental neurotoxicology extends back at least to the early 1960s when Werboff and Gottlieb (37) reviewed published work on the postnatal sequelae of developmental exposure to X-irradiation and psychoactive drugs [see historical overview by Butcher, (7)]. The potential for human developmental neurotoxicity to result from exposure to chemicals, however, generally was not recognized until the late 1960s and early 1970s [for review, see Vorhees, (35)]. This recognition was based primarily on data from exposures to methylmercury, lead, and alcohol. More recently, the developmental neurotoxic effects of abused substances such as heroin, methadone and cocaine also have increased our awareness of this problem. The risk of human developmental neurotoxicity resulting from the use and abuse of such agents was underscored at a New York Academy
NIDA
of Sciences Conference on prenatal abuse of licit and illicit drugs (13) where the developmental effects of alcohol, caffeine, cigarettes, opioids, marijuana and delta-9-tetrahydrocannabinol, phencyclidine, amphetamine and cocaine were discussed. Table 1 provides a list of generally recognized or suspected human developmental neurotoxicants. For some of these agents, severe neurological and psychological dysfunction may occur in the absence of structural malformations or may be observed more frequently than structural defects. For example, neurological dysfunction is more representative of prenatal and neonatal childhood lead poisoning than are structural malformations [see Davis et al., (8)]. In other cases, e.g., anticonvulsants, neurological dysfunction may be part of a syndrome of structural and functional defects [see Adams et al., (2)].
1The views in this paper represent those of the authors and do not necessarily reflect the views or policies of the U.S. Environmental Protection Agency. 2Present address: RJ Reynolds/Nabisco, Winston-Salem, NC 27102. 3Requests for reprints should be addressed to Carole A. Kimmel, Ph.D., U.S. Environmental Protection Agency (RD-689), 401 M Street, SW, Washington, DC 20460.
175
176
REES ET AL.
TABLE 1
TABLE 2
GENERALLY RECOGNIZED HUMAN DEVELOPMENTAL NEUROTOXICANTS
SECTION 4 CHEMICALS AND DEVELOPMENTAL NEUROTOXICITY TESTING
Ethanol Methylmercury Lead Heroin Methadone Cocaine Diphenylhydantoin PCBs X-Irradiation
The incidence of neurological disease in children has significant public health consequences. The Interagency Committee on Learning Disabilities has suggested that a reasonable estimate of the percentage of persons affected by learning disabilities is 5-10% (15). Although the etiology of these disabilities has not been clearly identified, a number of factors including exposure to drugs and environmental agents have been associated with these conditions. Together, the data indicate that neurological dysfunction following developmental exposure to several drugs and chemicals poses a significant health risk, and that the problems may occur in the absence of other more frank signs of toxicity. REGULATORY AUTHORITY
Despite evidence indicating the risk of developmental neurotoxicity following chemical or drug exposure, regulatory agencies have been slow to take action. Regulatory activity began in 1975, when Great Britain and Japan promulgated testing regulations for developmental neurotoxicity as part of their overall testing battery for reproductive and teratologic testing of pharmaceutical agents [reviewed by Kimmel, (16)]. However, these guidelines were very general and did not provide guidance on methods for evaluation of developmental neurotoxicity, data handling, etc. In 1983, the European Economic Community (10) proposed similar guidelines. In 1986 the World Health Organization (WHO) (39) proposed behavioral teratology testing guidelines that still were very general, but discussed some of the important considerations in testing and study design. Until recently, evidence that developmental exposure to chemicals can cause neurotoxicity has had little, if any, impact on regulatory actions taken by U.S. agencies [see review by Vorhees, (36)]. Within the last few years some Federal Agencies have demonstrated increased awareness of the potential for chemical insult to affect the developing nervous system. For example, the National Institute for Occupational Safety and Health (NIOSH), which is authorized to develop and establish recommendations for occupational and health standards, has recognized the potential for developmental neurotoxicity in formulating the Current Intelligence Bulletin for the glycol ethers (20). Under the Federal Food, Drug and Cosmetic Act (FFD&CA), the Food and Drug Administration (FDA) has the authority to regulate food additives, color additives, poisonous or deleterious substances in food, drugs, and biologics. The FDA traditionally has been supportive of research efforts in assessing developmental neurotoxicity [e.g., (4,18)] and has always had the authority to require testing for these types of effects. However, FDA has taken few actions to require testing in specific cases and has no requirement for routine developmental neurotoxicity evaluation. The National Institute on Drug Abuse (NIDA) is interested not only in the effects of abused agents but also in the development of
Reference No. Final Rules Diethylene glycol butyl ether Triethylene glycol monomethylether Isopropanol
(29) (31) (33)
Consent Order/Negotiated Testing Agreement Cyclohexanone 1,1,1-Trichloroethane
(27) (32)
Proposed Test Rule Cyclohexane
(28)
new pharmacotherapeutics to aid in the treatment of drug abuse. Funds provided in the Anti-Drug Abuse Act of 1988 support a major effort for establishing research and development programs aimed at developing drugs for the treatment of addictive disorders. Since these drugs will have to be tested for efficacy and potential toxicity first using animal models, and subsequently in clinical trials, NIDA recognizes that their potential for inducing developmental neurotoxicity must be considered. The U.S. Environmental Protection Agency (EPA) has taken the most systematic approach to evaluating data and requiring testing of agents for potential developmental neurotoxicity. The Agency's developmental toxicity risk assessment guidelines (30) identify functional deficits as well as death, structural abnormalities and altered growth as developmental effects of concern which should be considered when assessing risk in the developing organism. EPA has authority to require testing of agents for potential health effects, including developmental neurotoxicity, under two acts: the Federal Insecticide, Fungicide and Rodenticide Act (FIFRA), and the Toxic Substances Control Act (TSCA). Under FIFRA, all pesticides must be registered, which means that new pesticides must be tested according to specified standards and older pesticides must be reassessed to meet current standards. The standard for registration is that a pesticide does not cause an unreasonable risk to man or environment. Economic, social and environmental costs and benefits are also considered. Thus, the EPA must weigh the risks against the benefits of pesticide use. Currently, the only specific neurotoxic effect for which standards have been developed is the testing of organophosphates and their metabolites for delayed neuropathy in hens. An expanded approach to neurotoxicity testing has been proposed (34) and includes testing for developmental neurotoxicity. Under TSCA, the EPA regulates chemicals and has the authority to require adult and developmental neurotoxicity testing. Two sections of TSCA (Sections 4 and 5) are particularly relevant. Under Section 4, EPA has the authority, by rule, to require industry to test existing chemicals for possible adverse health and environmental effects. Testing responsibilities can be shared among all manufacturers and processors of a given chemical. Testing requirements are developed under two conditions: 1) when exposure may present an unreasonable risk of injury to health and the environment and there are insufficient data and experience upon which to reasonably predict or determine these effects [Section 4(a)(1)(A)], and 2) when the chemical is produced in "substantial" quantities which could result in substantial or significant human exposure or substantial environmental release
REGULATORY ISSUES: AN OVERVIEW
and there are insufficient data and experience upon which to reasonably predict or determine the effects [Section 4(a)(1)(B)]. If EPA determines that testing of a designated chemical is necessary they may publish a proposed and then final test rule that specifies the scope of testing required. If an agreement can be reached between the chemical companies involved and the EPA regarding a test program, the result may be a negotiated consent agreement which becomes legally enforceable. Developmental neurotoxicity testing has been recommended or required under Section 4 of TSCA for several agents as listed in Table 2. These few chemicals are only a sample representation of the number of chemicals for which EPA's Office of Toxic Substances (OTS), which carries out the mandates of TSCA, has required testing on any health or environmental effect. OTS has developed a list of criteria that should be considered in determining whether a chemical should be tested for developmental neurotoxicity (see further discussion under Work Group IV below). Under Section 5 of TSCA, EPA must evaluate the potential for adverse effects of new chemicals. The process of assessing new chemicals under Section 5 is often referred to as the Premanufacturing Notification (PMN) Process. Section 5 requires each manufacturer of a new chemical substance to submit a notice to EPA at least 90 days prior to beginning manufacture, processing or importing of the chemical. EPA reviews the information available on chemicals to determine whether any testing should be required prior to market entry, the time when regulation can be accomplished with the least cost. Section 5 provides authority to control the new chemicals, obtain additional information, and prohibit production entirely, as appropriate. Because there is no explicit requirement for the development of toxicity data prior to entry on the market, data often are not available. Consequently, preliminary assessment of potential human health effects must rely heavily upon structure-activity information. To date, no developmental neurotoxicity testing of new chemicals has been required prior to entry on the market. ISSUES AND APPROACHESTO DEVELOPMENTAL NEUROTOXICITYTESTING When considering approaches to assessing effects on the developing nervous system, a number of factors should be considered. The developing nervous system is sufficiently unique to warrant testing independent from adult neurotoxicity testing and general developmental toxicity testing. The dynamics of the developing nervous system may require different qualitative and quantitative considerations when assessing risk than would be the case for the mature nervous system. Although data from neurological testing in adults may provide useful information regarding the potential for effects in the developing nervous system, such information by itself may underestimate risk. Exposure of the developing nervous system may produce effects at lower doses than in the adult, and may result in long-term or permanent effects. For example, the neurotoxicity of lead is observed at lower dose levels and with longer-lasting effects in children than in adults [see review by Davis et al., (8)]. The effects on the developing nervous system also may be qualitatively different than those observed in adults. In some cases, for example, the effects on the offspring will result in permanent impairment, while those in the adult will be reversible. The long period of nervous system development with multiple "critical windows" of vulnerability provides the potential for the developing nervous system to be uniquely susceptible to chemical exposure. OTS developed the first systematic approach to developmental neurotoxicity testing by a regulatory office. This approach involved the development of a battery of tests chosen to evaluate the
177 TABLE 3 OTS DEVELOPMENTALNEUROTOXICITYBATTERY Gestation Day 0 Days 6-21 Day 6 Postnatal Days 1-21 Day 1 Day 4 Day 7 Day 13 Day 17 Day 21
Day 22 Day 45( - 2) Day 60( - 2)
Sperm and/or plug positive Body weights, OBS* Begin treatment Dams weighed, OBS Pups counted, weighed, OBS Litters culled to 4 males and 4 females, all pups weighed, OBS, tatoo All pups weighed, OBS Motor activity; all pups weighed, OBS Motor activity; all pups weighed, OBS Treatment discontinued at weaning: Motor activity; all pups weighed, OBS, neuropathology, brain weights Auditory startle response, test pups weighed, OBS Motor activity, body weights, OBS Motor activity, body weights, active avoidance, auditory startle response, neuropathology, brain weight, OBS
*OBS: Observations, including clinical signs and/or physical landmarks.
combined functioning of several integrated systems including sensory systems, neuromotor development, reactivity and/or habituation, and learning and memory. The developmental neurotoxicity testing guideline was designed to provide guidance for the generation of data on the potential functional and morphological hazards to the nervous system which may arise in offspring from exposure of the mother during pregnancy and lactation. This approach followed the recommendations of a variety of expert panels and committees [e.g., (17, 18, 38)]. The battery was developed specifically for use in testing the class of compounds known as the glycol ethers (29,31) under Section 4 of TSCA. The same battery has also been applied for the testing of isopropanol (33) and has been proposed for testing of cyclohexane (28). An outline of this battery is presented in Table 3. This guideline was subsequently used as a model for the testing of 1,1,l-trichlorethane (32). As a result of negotiations with industry and a consensus among participants of an agency-wide work group on developmental neurotoxicity, there was an effort made to expand the guidelines and to allow greater flexibility in the choice of assays to test particular functions. Some of the alternatives that were proposed are found in Table 4. Active avoidance had been specified in the diethylene glycol
TABLE 4 PROPOSED ALTERNATIVESTO DEVELOPMENTAL NEUROTOXICITYBATI'ERY 1) 2) 3) 4)
More flexible choice of tests for learning and memory Evaluation of learning and memory on Days 21-24 as well as on Day 60 Addition of neuropathology and brain weights on Day 4; more detailed protocol for neuropathology Glial fibrillary acidic protein (GFAP) assay, Days 4, 24 and 64
REES ET AL.
178
1) 2) 3) 4) 5) 6) 7) 8) 9)
TABLE 5
TABLE 6
CRITERIA FOR CHOOSING TESTS FOR COGNITIVE FUNCTION
UNCERTAINTY FACTORS AND THEIR USE IN THE RFD CALCULATION
Construct validity: Does the test unequivocally indicate an impairment of learning and memory? Well-characterized: Is there a large literature on parameters/ properties of task performance? Homologous with human function: Does the task predict a similar result in humans? Sensitivity: Does the task detect a wide range of compounds and at relatively low doses? Toxicology data base: Has the efficacy of the task been demonstrated with many neurotoxic compounds? Developmental profile: Does the task measure a form of cognition which develops postnatally? Time and cost effectiveness Number of subjects required Automation: Commercially available?
butyl ether (DGBE) test rule as a measure of learning and memory, based upon data indicating it to be a useful measure for this class of compounds (22). While active avoidance may be the appropriate test in some circumstances, other approaches also may be acceptable and guidance for selection of alternative tests for assessing learning, memory and/or performance has been outlined (Table 5). RISK ASSESSMENT FOR DEVELOPMENTAL NEUROTOXICITY
Data from a developmental neurotoxicity testing protocol should provide information that can be used in the risk assessment process. Risk assessment involves the evaluation of the toxic properties of a chemical and the implicit conditions of human exposure to determine the likelihood that exposed humans will be adversely affected and to estimate the extent of the effect in the population or levels below which exposure should not result in adverse effects. The National Research Council (21) has identified four major components in the risk assessment process. Hazard identification involves the evaluation of all available experimental animal and human data relevant to assessing potential toxicity. Hazard identification is only one component of risk assessment; it is intended to determine, as in the case of developmental neurotoxicity, the potential for such effects to occur in humans. Dose-response assessment defines the relationship between dose and effect with particular emphasis on high to low dose extrapolation and extrapolation from experimental animals to humans. It is recognized that there may be different dose response relationships for different effects. For noncancer and nonmutagenic end points (e.g., developmental neurotoxicity), the EPA uses one of two approaches in assessing risk, both of which are based on the lowest observable adverse effect level (LOAEL) or, preferably, the no observed adverse effect level (NOAEL). These include applying uncertainty factors to the NOAEL or LOAEL to establish a reference dose, or determining the ratio of the NOAEL to the human exposure estimate to give a margin of exposure (MOE). Unlike the approach taken for carcinogenic and mutagenic effects which assumes no threshold, these approaches assume the existence of a threshold. The justification offered for this assumption is that homeostatic, compensatory, and adaptive mechanisms must be overcome in order for noncancer and nonmutagenic effects to be observed. In the approach using uncertainty factors (3), various criteria are used as the basis for the magnitude of the total uncertainty
10H
10-Fold factor for variation in sensitivity among members of the human population 10A 10-Fold factor for extrapolation from animal data to humans 10S* 10-Fold factor when extrapolating from less than chronic results to a chronic exposure situation 10L 10-Fold factor when extrapolating from a LOAEL to a NOAEL MF (Modifying factor)
0892-0362/90 $3.00 + .00
Scientific and Regulatory Issues Relevant to Assessing Risk for Developmental Neurotoxicity: An Overview D. C O O P E R R E E S , .2 E L A I N E Z. F R A N C I S t A N D C A R O L E A. K I M M E L t 3
*Health and Environmental Review Division, Office of Toxic Substances Office of Pesticides and Toxic Substances and ~Reproductive and Developmental Toxicology Branch/HHAG Office of Health and Environmental Assessment, Office of Research and Development United States Environmental Protection Agency, Washington, DC 24060
REES, D. C., E. Z. FRANCIS AND C. A. KIMMEL. Scientific and regulatory issues relevant to assessing riskfor developmental neurotoxicity: An overview. NEUROTOXICOL TERATOL 12(3) 175-181, 1990.--The effects of chemical exposure on the developing nervous system have been documented in both humans and animals for a variety of agents. However, the comparability of these effects has not been carefully evaluated to determine the predictability of animal models to adverse effects in humans. A workshop sponsored by the U.S. Environmental Protection Agency (EPA) and the National Institute on Drug Abuse was held on April 11-13, 1989, to address the Qualitative and Quantitative Comparability of Human and Animal Developmental Neurotoxicity. Invited experts were asked to review the human and animal data on several agents that are known to cause developmental neurotoxicity in humans, including lead, methylmercury, selected abused agents, anticonvulsants, polychlorinated biphenyls (PCBs), ethanol and X-irradiation, and to make qualitative comparisons on a specific end point basis as well as on a functional category basis. In addition, they were asked to make quantitative comparisons when adequate dose-effect data were available. The data also were evaluated in the context of the proposed EPA developmental neurotoxicity testing battery to determine whether or not the battery would adequately detect the effects of each agent. Finally, four work groups were asked to reach consensus on issues relating to: 1) comparability of end points across species for developmental neurotoxicity; 2) testing methods in developmental neurotoxicity for use in human risk assessment; 3) weight-of-evidence and quantitative evaluation of data from developmental neurotoxicity studies; and 4) triggers for developmental neurotoxicity testing. Risk assessment
Developmental neurotoxicity
TSCA
FIFRA
LITERATURE in the field of behavioral teratology/developmental neurotoxicology extends back at least to the early 1960s when Werboff and Gottlieb (37) reviewed published work on the postnatal sequelae of developmental exposure to X-irradiation and psychoactive drugs [see historical overview by Butcher, (7)]. The potential for human developmental neurotoxicity to result from exposure to chemicals, however, generally was not recognized until the late 1960s and early 1970s [for review, see Vorhees, (35)]. This recognition was based primarily on data from exposures to methylmercury, lead, and alcohol. More recently, the developmental neurotoxic effects of abused substances such as heroin, methadone and cocaine also have increased our awareness of this problem. The risk of human developmental neurotoxicity resulting from the use and abuse of such agents was underscored at a New York Academy
NIDA
of Sciences Conference on prenatal abuse of licit and illicit drugs (13) where the developmental effects of alcohol, caffeine, cigarettes, opioids, marijuana and delta-9-tetrahydrocannabinol, phencyclidine, amphetamine and cocaine were discussed. Table 1 provides a list of generally recognized or suspected human developmental neurotoxicants. For some of these agents, severe neurological and psychological dysfunction may occur in the absence of structural malformations or may be observed more frequently than structural defects. For example, neurological dysfunction is more representative of prenatal and neonatal childhood lead poisoning than are structural malformations [see Davis et al., (8)]. In other cases, e.g., anticonvulsants, neurological dysfunction may be part of a syndrome of structural and functional defects [see Adams et al., (2)].
1The views in this paper represent those of the authors and do not necessarily reflect the views or policies of the U.S. Environmental Protection Agency. 2Present address: RJ Reynolds/Nabisco, Winston-Salem, NC 27102. 3Requests for reprints should be addressed to Carole A. Kimmel, Ph.D., U.S. Environmental Protection Agency (RD-689), 401 M Street, SW, Washington, DC 20460.
175
176
REES ET AL.
TABLE 1
TABLE 2
GENERALLY RECOGNIZED HUMAN DEVELOPMENTAL NEUROTOXICANTS
SECTION 4 CHEMICALS AND DEVELOPMENTAL NEUROTOXICITY TESTING
Ethanol Methylmercury Lead Heroin Methadone Cocaine Diphenylhydantoin PCBs X-Irradiation
The incidence of neurological disease in children has significant public health consequences. The Interagency Committee on Learning Disabilities has suggested that a reasonable estimate of the percentage of persons affected by learning disabilities is 5-10% (15). Although the etiology of these disabilities has not been clearly identified, a number of factors including exposure to drugs and environmental agents have been associated with these conditions. Together, the data indicate that neurological dysfunction following developmental exposure to several drugs and chemicals poses a significant health risk, and that the problems may occur in the absence of other more frank signs of toxicity. REGULATORY AUTHORITY
Despite evidence indicating the risk of developmental neurotoxicity following chemical or drug exposure, regulatory agencies have been slow to take action. Regulatory activity began in 1975, when Great Britain and Japan promulgated testing regulations for developmental neurotoxicity as part of their overall testing battery for reproductive and teratologic testing of pharmaceutical agents [reviewed by Kimmel, (16)]. However, these guidelines were very general and did not provide guidance on methods for evaluation of developmental neurotoxicity, data handling, etc. In 1983, the European Economic Community (10) proposed similar guidelines. In 1986 the World Health Organization (WHO) (39) proposed behavioral teratology testing guidelines that still were very general, but discussed some of the important considerations in testing and study design. Until recently, evidence that developmental exposure to chemicals can cause neurotoxicity has had little, if any, impact on regulatory actions taken by U.S. agencies [see review by Vorhees, (36)]. Within the last few years some Federal Agencies have demonstrated increased awareness of the potential for chemical insult to affect the developing nervous system. For example, the National Institute for Occupational Safety and Health (NIOSH), which is authorized to develop and establish recommendations for occupational and health standards, has recognized the potential for developmental neurotoxicity in formulating the Current Intelligence Bulletin for the glycol ethers (20). Under the Federal Food, Drug and Cosmetic Act (FFD&CA), the Food and Drug Administration (FDA) has the authority to regulate food additives, color additives, poisonous or deleterious substances in food, drugs, and biologics. The FDA traditionally has been supportive of research efforts in assessing developmental neurotoxicity [e.g., (4,18)] and has always had the authority to require testing for these types of effects. However, FDA has taken few actions to require testing in specific cases and has no requirement for routine developmental neurotoxicity evaluation. The National Institute on Drug Abuse (NIDA) is interested not only in the effects of abused agents but also in the development of
Reference No. Final Rules Diethylene glycol butyl ether Triethylene glycol monomethylether Isopropanol
(29) (31) (33)
Consent Order/Negotiated Testing Agreement Cyclohexanone 1,1,1-Trichloroethane
(27) (32)
Proposed Test Rule Cyclohexane
(28)
new pharmacotherapeutics to aid in the treatment of drug abuse. Funds provided in the Anti-Drug Abuse Act of 1988 support a major effort for establishing research and development programs aimed at developing drugs for the treatment of addictive disorders. Since these drugs will have to be tested for efficacy and potential toxicity first using animal models, and subsequently in clinical trials, NIDA recognizes that their potential for inducing developmental neurotoxicity must be considered. The U.S. Environmental Protection Agency (EPA) has taken the most systematic approach to evaluating data and requiring testing of agents for potential developmental neurotoxicity. The Agency's developmental toxicity risk assessment guidelines (30) identify functional deficits as well as death, structural abnormalities and altered growth as developmental effects of concern which should be considered when assessing risk in the developing organism. EPA has authority to require testing of agents for potential health effects, including developmental neurotoxicity, under two acts: the Federal Insecticide, Fungicide and Rodenticide Act (FIFRA), and the Toxic Substances Control Act (TSCA). Under FIFRA, all pesticides must be registered, which means that new pesticides must be tested according to specified standards and older pesticides must be reassessed to meet current standards. The standard for registration is that a pesticide does not cause an unreasonable risk to man or environment. Economic, social and environmental costs and benefits are also considered. Thus, the EPA must weigh the risks against the benefits of pesticide use. Currently, the only specific neurotoxic effect for which standards have been developed is the testing of organophosphates and their metabolites for delayed neuropathy in hens. An expanded approach to neurotoxicity testing has been proposed (34) and includes testing for developmental neurotoxicity. Under TSCA, the EPA regulates chemicals and has the authority to require adult and developmental neurotoxicity testing. Two sections of TSCA (Sections 4 and 5) are particularly relevant. Under Section 4, EPA has the authority, by rule, to require industry to test existing chemicals for possible adverse health and environmental effects. Testing responsibilities can be shared among all manufacturers and processors of a given chemical. Testing requirements are developed under two conditions: 1) when exposure may present an unreasonable risk of injury to health and the environment and there are insufficient data and experience upon which to reasonably predict or determine these effects [Section 4(a)(1)(A)], and 2) when the chemical is produced in "substantial" quantities which could result in substantial or significant human exposure or substantial environmental release
REGULATORY ISSUES: AN OVERVIEW
and there are insufficient data and experience upon which to reasonably predict or determine the effects [Section 4(a)(1)(B)]. If EPA determines that testing of a designated chemical is necessary they may publish a proposed and then final test rule that specifies the scope of testing required. If an agreement can be reached between the chemical companies involved and the EPA regarding a test program, the result may be a negotiated consent agreement which becomes legally enforceable. Developmental neurotoxicity testing has been recommended or required under Section 4 of TSCA for several agents as listed in Table 2. These few chemicals are only a sample representation of the number of chemicals for which EPA's Office of Toxic Substances (OTS), which carries out the mandates of TSCA, has required testing on any health or environmental effect. OTS has developed a list of criteria that should be considered in determining whether a chemical should be tested for developmental neurotoxicity (see further discussion under Work Group IV below). Under Section 5 of TSCA, EPA must evaluate the potential for adverse effects of new chemicals. The process of assessing new chemicals under Section 5 is often referred to as the Premanufacturing Notification (PMN) Process. Section 5 requires each manufacturer of a new chemical substance to submit a notice to EPA at least 90 days prior to beginning manufacture, processing or importing of the chemical. EPA reviews the information available on chemicals to determine whether any testing should be required prior to market entry, the time when regulation can be accomplished with the least cost. Section 5 provides authority to control the new chemicals, obtain additional information, and prohibit production entirely, as appropriate. Because there is no explicit requirement for the development of toxicity data prior to entry on the market, data often are not available. Consequently, preliminary assessment of potential human health effects must rely heavily upon structure-activity information. To date, no developmental neurotoxicity testing of new chemicals has been required prior to entry on the market. ISSUES AND APPROACHESTO DEVELOPMENTAL NEUROTOXICITYTESTING When considering approaches to assessing effects on the developing nervous system, a number of factors should be considered. The developing nervous system is sufficiently unique to warrant testing independent from adult neurotoxicity testing and general developmental toxicity testing. The dynamics of the developing nervous system may require different qualitative and quantitative considerations when assessing risk than would be the case for the mature nervous system. Although data from neurological testing in adults may provide useful information regarding the potential for effects in the developing nervous system, such information by itself may underestimate risk. Exposure of the developing nervous system may produce effects at lower doses than in the adult, and may result in long-term or permanent effects. For example, the neurotoxicity of lead is observed at lower dose levels and with longer-lasting effects in children than in adults [see review by Davis et al., (8)]. The effects on the developing nervous system also may be qualitatively different than those observed in adults. In some cases, for example, the effects on the offspring will result in permanent impairment, while those in the adult will be reversible. The long period of nervous system development with multiple "critical windows" of vulnerability provides the potential for the developing nervous system to be uniquely susceptible to chemical exposure. OTS developed the first systematic approach to developmental neurotoxicity testing by a regulatory office. This approach involved the development of a battery of tests chosen to evaluate the
177 TABLE 3 OTS DEVELOPMENTALNEUROTOXICITYBATTERY Gestation Day 0 Days 6-21 Day 6 Postnatal Days 1-21 Day 1 Day 4 Day 7 Day 13 Day 17 Day 21
Day 22 Day 45( - 2) Day 60( - 2)
Sperm and/or plug positive Body weights, OBS* Begin treatment Dams weighed, OBS Pups counted, weighed, OBS Litters culled to 4 males and 4 females, all pups weighed, OBS, tatoo All pups weighed, OBS Motor activity; all pups weighed, OBS Motor activity; all pups weighed, OBS Treatment discontinued at weaning: Motor activity; all pups weighed, OBS, neuropathology, brain weights Auditory startle response, test pups weighed, OBS Motor activity, body weights, OBS Motor activity, body weights, active avoidance, auditory startle response, neuropathology, brain weight, OBS
*OBS: Observations, including clinical signs and/or physical landmarks.
combined functioning of several integrated systems including sensory systems, neuromotor development, reactivity and/or habituation, and learning and memory. The developmental neurotoxicity testing guideline was designed to provide guidance for the generation of data on the potential functional and morphological hazards to the nervous system which may arise in offspring from exposure of the mother during pregnancy and lactation. This approach followed the recommendations of a variety of expert panels and committees [e.g., (17, 18, 38)]. The battery was developed specifically for use in testing the class of compounds known as the glycol ethers (29,31) under Section 4 of TSCA. The same battery has also been applied for the testing of isopropanol (33) and has been proposed for testing of cyclohexane (28). An outline of this battery is presented in Table 3. This guideline was subsequently used as a model for the testing of 1,1,l-trichlorethane (32). As a result of negotiations with industry and a consensus among participants of an agency-wide work group on developmental neurotoxicity, there was an effort made to expand the guidelines and to allow greater flexibility in the choice of assays to test particular functions. Some of the alternatives that were proposed are found in Table 4. Active avoidance had been specified in the diethylene glycol
TABLE 4 PROPOSED ALTERNATIVESTO DEVELOPMENTAL NEUROTOXICITYBATI'ERY 1) 2) 3) 4)
More flexible choice of tests for learning and memory Evaluation of learning and memory on Days 21-24 as well as on Day 60 Addition of neuropathology and brain weights on Day 4; more detailed protocol for neuropathology Glial fibrillary acidic protein (GFAP) assay, Days 4, 24 and 64
REES ET AL.
178
1) 2) 3) 4) 5) 6) 7) 8) 9)
TABLE 5
TABLE 6
CRITERIA FOR CHOOSING TESTS FOR COGNITIVE FUNCTION
UNCERTAINTY FACTORS AND THEIR USE IN THE RFD CALCULATION
Construct validity: Does the test unequivocally indicate an impairment of learning and memory? Well-characterized: Is there a large literature on parameters/ properties of task performance? Homologous with human function: Does the task predict a similar result in humans? Sensitivity: Does the task detect a wide range of compounds and at relatively low doses? Toxicology data base: Has the efficacy of the task been demonstrated with many neurotoxic compounds? Developmental profile: Does the task measure a form of cognition which develops postnatally? Time and cost effectiveness Number of subjects required Automation: Commercially available?
butyl ether (DGBE) test rule as a measure of learning and memory, based upon data indicating it to be a useful measure for this class of compounds (22). While active avoidance may be the appropriate test in some circumstances, other approaches also may be acceptable and guidance for selection of alternative tests for assessing learning, memory and/or performance has been outlined (Table 5). RISK ASSESSMENT FOR DEVELOPMENTAL NEUROTOXICITY
Data from a developmental neurotoxicity testing protocol should provide information that can be used in the risk assessment process. Risk assessment involves the evaluation of the toxic properties of a chemical and the implicit conditions of human exposure to determine the likelihood that exposed humans will be adversely affected and to estimate the extent of the effect in the population or levels below which exposure should not result in adverse effects. The National Research Council (21) has identified four major components in the risk assessment process. Hazard identification involves the evaluation of all available experimental animal and human data relevant to assessing potential toxicity. Hazard identification is only one component of risk assessment; it is intended to determine, as in the case of developmental neurotoxicity, the potential for such effects to occur in humans. Dose-response assessment defines the relationship between dose and effect with particular emphasis on high to low dose extrapolation and extrapolation from experimental animals to humans. It is recognized that there may be different dose response relationships for different effects. For noncancer and nonmutagenic end points (e.g., developmental neurotoxicity), the EPA uses one of two approaches in assessing risk, both of which are based on the lowest observable adverse effect level (LOAEL) or, preferably, the no observed adverse effect level (NOAEL). These include applying uncertainty factors to the NOAEL or LOAEL to establish a reference dose, or determining the ratio of the NOAEL to the human exposure estimate to give a margin of exposure (MOE). Unlike the approach taken for carcinogenic and mutagenic effects which assumes no threshold, these approaches assume the existence of a threshold. The justification offered for this assumption is that homeostatic, compensatory, and adaptive mechanisms must be overcome in order for noncancer and nonmutagenic effects to be observed. In the approach using uncertainty factors (3), various criteria are used as the basis for the magnitude of the total uncertainty
10H
10-Fold factor for variation in sensitivity among members of the human population 10A 10-Fold factor for extrapolation from animal data to humans 10S* 10-Fold factor when extrapolating from less than chronic results to a chronic exposure situation 10L 10-Fold factor when extrapolating from a LOAEL to a NOAEL MF (Modifying factor)
