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.
Bridging Experimental Animal and Human Behavioral Toxicology Studies Deborah A. Cory-SIechta THE SCOPE AND AGENDA OF BEHAVIORAL TOXICOLOGY Behavioral toxicology can be generally conceptualized as that sci- entific discipline which strives to understand the mechanisms by which toxicants affect behavior. In this respect, it is similar to its counterpart, behavioral pharmacology, the goal of which is to delineate the mechanisms by which drugs modulate behavior. Since its inception, the scope of behavioral toxicology has expanded considerably, driven in part by the need to ascertain the role of existing environmental contaminants in producing functional impairment, as well as the need to develop procedures to preclude future introduction of neurotoxic chemicals. In this way, it differs from behavioral pharmacology which, instead, seeks to develop compounds or agents with specific behavioral actions for therapeutic purposes. As the above implies, behavioral toxicology actually has dual, overlapping agendas. One derives from the growing recognition of the need to screen for performance impairment prior to the introduction of new chemicals into the environment, as well as to provide infor- mation relating to risk assessment based on neurotoxic endpoints. The second agenda involves the more traditional role of behavioral toxicology as the scientific discipline defined above, whose goal is to understand both the behavioral and the biological mechanisms by which toxicants impact behavioral function. It is primarily the latter agenda to which the comments herein are addressed. 137
138 DEBORAH A. CORY-SLECHTA Despite recent advances in neurobiology and the obvious utility of in vitro approaches in the elucidation of biological substrates of behavior, it is difficult to conceive of any substitute for assessing the ultimate impact of a neurotoxicant on behavior, or of assessing behavioral mechanisms of toxicant action, other than in the whole organism. The links between molecular neurobiology or between neuropathological alterations and behavioral impairments are still obscure. Although the relationships between certain neurotransmitters and behavior have become increasingly evident, such as the role of dopamine in parkinsonism, other aspects of those relationships remain puzzling, e.g., the extensive dopaminergic depletions noted before any overt behavioral impairments appear. Thus, there are no substitute or alternative procedures for evaluating the functional impact of a toxi- cant. Put another way, behavioral toxicity cannot be reliably predicted from molecular events. STATE OF DEVELOPMENT Although the experimental capabilities for more precisely delin- eating behavioral and biological mechanisms of toxicant-induced performance impairment are generally at hand, the discipline remains largely at a characterization or descriptive stage of development. Much of its scientific literature attempts little more than to ascertain whether a particular toxicant alters a particular class of behavior, or to assess performance impairments produced by a toxicant across a range of behavioral endpoints; in some cases, only the barest approximation to a hypothesis may be invoked. Furthermore, little attempt may be made to rationalize the particular behavioral approach chosen, which may, instead, be based predominantly on available apparatus or technology in that laboratory. Nonetheless, owing more to the sheer magnitude of work with certain compounds, rather than to any systematic progression of studies within a laboratory, in certain areas these studies have begun to provide the prerequisite foundation from which more mechanistic approaches can now proceed. Studies of performance impairments induced by lead exposure provide one example. Lead may be considered a prototypical behavioral toxicant and undoubtedly has been the most extensively studied of such com- pounds, both at the experimental animal and at the human level. The permanent mental retardation, which in some cases was the residual effect of acute high-dose lead exposure in children, resulted in a subsequent focus of these studies on issues of learning deficits at lower lead exposure levels. Human studies of environmental lead exposure in children have almost invariably focused on age-appropriate IQ and other psychometric tests as their behavioral endpoint. The
BEHAVIORAL TOXICOLOGY STUDIES 139 most recent of such studies have documented decrements in IQ and similar psychometric measures at blood lead concentrations as low as 10 ~g/dL (e.g., Bellinger et al., 1987; Fulton et al., 1987~. However, even with the separation of verbal and motor subscales, IQ tests represent extremently global measures of performance that encompass a variety of different behavioral functions, as well as the involvement of multiple sensory systems, many of which may be marginally affected or others of which may be more dramatically affected. Such global measures always present the possibility that particular subtle deficits may be obscured by the sheer multiplicity of concurrently measured behaviors or may be clouded by a reserve capacity of the organism. The specific nature of the IQ decrements in humans thus remains unresolved. Experimental animal studies can more readily address aspects of lead-induced changes in learning. Table 1 summarizes those studies that have assessed lead-induced changes in learning by using acqui- sition of a visual discrimination as a behavioral endpoint. The various studies are subdivided on the basis of both the type of visual cue utilized and the developmental period of lead exposure. Plus signs show those experiments reporting an impairment of visual discrimi- nation learning as a result of lead exposure, whereas minus signs accompany those that found no change. As indicated by the prepon- derance of plus signs in each column, two types of visual discrimination paradigms emerge as those more sensitive to lead exposures: discriminations based on differences in brightness and on size of visual cues. Shape-form discrimination shows little obvious impact of lead. A within-laboratory comparison provides further support for this across- study conclusion. Winneke et al. (1977) reported that lead-treated rats required more trials to acquire a size discrimination than did control rats, but were not impaired in the acquisition of a form cliscrim~ation. Although the effects noted with color-based dis0in~nation are suggestive, they are, at present, based on a restricted data set. The differential lead effects based upon visual cue emphasize the critical importance of the environmental context in modulating the behavioral effects of a toxicant such as lead. No generalized deficit in visual discrimination learning can be ascribed to lead; instead, such deficits depend upon environmental cues. Studies of visual discrimination learning following lead exposure can direct future efforts aimed at understanding the behavioral and biological mechanisms that might explain such differential effects. With respect to behavioral mechanisms, the possiblity of a generalized performance decrement for example, an increase in response bias or an alteration in motivation level obviously fails to accommodate the differential effects of visual cue. One explanation resides in the possibility of differential degrees of control exerted by the stimuli
140 o - o u o . ._ ._ u U. ·_I in u Cd - Cat ¢ Ed N ._ o it) U) .= O ~ 5 to >I" ~ ~ - Cal - o ~ O pt O +~ +~ _ _ Ct tt! C)= ~ _ _ - ._ a: + - - CIS Ct u ~ ~ ~ - - ~ - on - ~ - ~ ~ .~ u ~ ~ · a' += += + - - a; o it ~ - - - c'D == ~ (d a.~ ~ ~ y ~ m+ ~ - , a, ~ ~ - ~ ~ - ~ - ~ cn a~ a,, u ~ o ~ ~+ - oo o .~ o t~ - ~ - - - 3 Ct C~ o _ . ~ . - _ U. ~ ~, ~o ~ o.5 CO ~ ~ ~0 ._ ~n + P~ _ . ._ ~n 00 - o cn cn - Ct I - a, o . ~ O 41, ~ O 1_ O ~ CO .5:,
BEHAVIORAL TOXICOLOGY STUDIES 141 over behaviors, which consequently exhibit differential behavioral sensitivity to disruption. Numerous behavioral pharmacology stud- ies have shown that central nervous system (CNS) drugs may exhibit a much greater magnitude of effect on behavior that is under weak stimulus control (Laties, 1975), i.e., performances which exhibit rela- tively low overall accuracy levels or require an extensive number of trials to attain criterion performance. In such cases, the stimuli may be less salient, thus failing to generate strong control over performance and rendering it more easily disrupted by other factors. The fact that lead effects on visual discrimination learning depend upon the type of visual cue, as shown in Table 1, might also reflect differential sensory effects of lead on the visual system. In regard to such a possibility, Fox et al. (1982) have reported persistent decreases in both visual acuity and spatial resolution in 90-day old rats that had been exposed from birth to weaning to 200 ppm of lead acetate via the dam. Although these, as well as other visual system deficits resulting from lead exposure have been reported (Bushnell et al., 1977; Fox and Chu, 1988; Fox and Farber, 1988), their direct impact in mediating behavioral toxicity remains to be systematically investi- gated. Experimental animal studies also reveal that lead impairs learning based on the acquisition of spatial discrimination and can be catego- rized on the basis of both the route and the developmental period of exposure. Table 2 illustrates several additional issues of importance TABLE 2 Lead-Induced Changes in Spatial Discrimination Learning Developmental Period of Lead Exposure Exposure Route Oral Intraperitoneal Pre- or postnatal +Snowdon (1973)a -Brown et al. (1971)b +Bushnell and Bowman (1979a) +Klein et al. (1977) +Bushnell and Bowman (1979b) -Rosen et al. (1985) -Overmann (1977) +Levin and Bowman (1983) +Laughlin et al. (1983) Postweaning +Geist and Mattes (1979) -Brown et al. (1971) Adult +Avery and Cross (1974) -Snowdon (1973) +Lanthorn and Isaacson (1978) -Ogilvie (1978) +Ogilvie (1977) -Bullock et al. (1966) +Dietz et al. (1979) -Penzien et al. (1982) aEffect of lead reported. bAbsence of lead effect reported.
142 DEBORAH A. CORY-SLECHTA in understanding the behavioral toxicity of lead, in particular, which may apply equally to other toxicants: for example, the critical impor- tance of the kinetics of lead to its behavioral toxicity, the detrimental effects on spatial learning, and the consistency of the effect across a variety of different behavioral procedures. Lead-induced impairment of spatial discrimination learning has been observed following postnatal, postweaning, and adult exposures, in contrast to brightness discrimi- nation (Table 1) which appears most vulnerable in response to prena- tal exposures. Thus, different behavioral performances may exhibit quite different critical periods of exposure to a toxicant, or the critical exposure period for behavioral effects produced by a toxicant may be determined at least partly by the sensitivity of the behavioral procedure. Comparative changes in schedule-controlled behavior induced by lead exposure reveal additional aspects of its behavioral toxicity, as- pects that in turn may impact on, or even underlie, other lead-induced performance effects. The most extensively studied of such reinforcement schedules has been the fixed-interval (FI) schedule, in which the reward for responding is temporally based, with the contingency stipulating that the first response occurring after a specified interval of time elapses produces reinforcement. Figure 1 presents the dose-effect function that summarizes the various studies of lead-induced changes in FI schedule-controlled behavior. It plots a parameter of FI perfor- mance (as a percentage of the corresponding control data) in relation to treatment dose. In constructing this summary figure, the lead exposure dosage or concentration has been recalculated in milligrams per kilogram. Because not all experimenters used the same dependent variables, response rate was used where presented, but in other cases, the outcome was based on total number of reponses, median interresponse time (IRT), or group mean percentage of control reinforcements, all of which can be impacted upon by response rate changes. The data were plotted from the session or sessions in which peak effects occurred, with the exception of data from our studies (Cory-Slechta and Thompson, 1979; Cory-Slechta et al., 1983, 1985), which were restricted to results from the first 30 experimental sessions so as to be comparable to the number of sessions used in most other studies. Prenatal and oral preweaning exposure studies could not be included in this summary figure because it was not possible to ascertain the dose to which the developing animals were exposed. Figure 1 reveals an inverse U-shaped function relating dose of lead to performance on the FI schedule of reinforce- ment. That is, exposure to lower concentrations or doses of lead produces response rate or output increases on the FI schedule; as the dose or exposure to lead increases, however, the rates of responding
BEHAVIORAL TOXICOLOGY STUDIES 300 250 O ~ o 150 100 50 0 I 1 - 01 3 o 10 · Rat 0 Donkey ° Shop · Plgeon \ 8,6 \. 1 \ ~ l ~ \ o. ^\ ~ ..... \ \ · 8 ~ 9 1 I .1 1 10 LEAN:) DOSE (mg/kg) 143 1 00 1 000 FIGURE 1 Summary of studies investigating changes In fixed-interval performance (plotted as percent of the control group) as a function of lead dosage, taken from the session or sessions in which peak effects occurred. Data from studies involving prena- tal or lactational exposures could not be included because it was not possible to ascer- tain the dose to which the developing organisms were exposed. Different experimen- tal species are indicated by different symbols; numbers refer to different studies: (1) Rice et al. (1979); (2) Cory-Slechta et al. (1985); (3) Cory-Slechta (1989); (4) Cory-Slechta et al. (1983); (5) Cory-Slechta and Thompson (1979); (6) Van Gelder et al. (1973); (7) Barthalmus et al. (1977); (8) Angell and Weiss (1982); (9) Zenick et al. (1979); (10) Rice (1988). SOURCE: Cory-Slechta (1984). are depressed below control values. Plotting changes in FI response rate against the reported blood lead values in each study produces a similar function (Cory-Slechta, 1984~. The potential generality of the dose-effect function is evidenced by the similarity of the lead-induced changes in rates of responding that have been described in other temporally based reinforcement sched- ules. For example, Nation et al. (1983) reported a dose-effect func- tion for lead-induced changes in variable-interval (VI) performance comparable to that shown in Figure 1 for the FI schedule, with a daily dose of 1.0 mg/kg of lead increasing VI response rates, whereas both 5.0 and 10.0 mg/kg suppressed response rates. Rice and Gilbert (1985) reported no effect of exposure to a low level of lead (associated
144 DEBORAH A. CORY-SLECHTA with steady-state blood lead of 11-13 ,ug/dL) on response rate or on the mean IRT value of monkeys responding on another temporally based schedule of reinforcement, a differential reinforcement of low rate (DRL 10 s or DRL 30 s) schedule. However, on both the DRL 10- and the DRL 30-s schedules, they did report an increase in the num- ber of nonreinforced responses and a decline in the number of reinforcers received by lead-treated monkeys, an effect that would seemingly necessitate an increased overall rate of responding. This pattern of effects would be evident in the distribution of IRTs but might not have impacted the measured index, mean IRT, the value of which could be substantially influenced by a small number of very long IRTs. The response rate-increasing properties of low-level lead exposure derive from a decrease in the time between successive responses (i.e., interresponse times). In particular, the frequency of short interresponse times (less than 0.5 s) during the fixed interval is increased by lead exposure, such that successive responses occur more rapidly in lead- exposed organisms than in controls; no consistent changes in postreinforcement pause time are noted. Figure 2 shows the proportion of short IRTs of control rats (left panels) and of rats treated with 25 ppm of lead acetate (right panels) over the course of 40 experimental sessions on an FI 60-s schedule of food reinforcement in two separate replications (top panel, Cory-Slechta et al., 1985; bottom panel, Cory- Slechta, 1989~. As can be seen, the range of short IRTs exhibited by control and lead-exposed rats was actually quite comparable, but lead exposure yielded a shift of the distribution toward the upper extremes of the range (i.e., higher proportions) in both replications. Thus, control and lead-treated animals begin to respond at the same time during the fixed interval, but lead-exposed organisms then respond at much higher rates than control animals, engendering more responding per unit time. In contrast, schedules of reinforcement based on number of re- sponses (ratio based), rather than on temporal parameters, exhibit a different pattern of lead effects, another indication that its behavioral toxicity is dependent upon the environmental or behavioral context. Although high-level exposures to lead are reliably associated with decreases in response rate on ratio schedules, evidence for rate-enhancing effects at lower exposure levels is not compelling (Angell and Weiss, 1982; Barthalmus et al., 1977; Cory-Slechta, 1986; Padich and Zenick, 1977; Rice, 1988~. Although Angell and Weiss (1982) reported shorter median IRTs on an FR schedule in rats exposed to lead prenatally only, the IRTs were actually not significantly different from those of nonexposed controls.
BEHAVIORAL TOXICOLOGY STUDIES 80 60 40 a, - V1 20 IL o of o o o 80 60 40 20 o 0 ppm $1 0 ppm #2 20 30 40 0 0 10 145 25 ppm #2 10 20 30 40 SESSIONS FIGURE 2 Proportion of short interresponse times (less than or equal to 0.5 s) of individual control rats (left panels) and rats exposed to 25 ppm of lead acetate (right panels) on a fixed-interval 60-s schedule of food reinforcement. Data are shown over the course of the first 40 experimental sessions. Top panels from Cory-Slechta et al. (1985); bottom panels, Cory-Slechta (1989~. As with the discrimination learning literature, comparative studies of lead-induced changes in schedules of reinforcement can guide mechanistically based experiments. Although an elevation in the rate of responding results from low-level lead exposure, the effect ap- pears to be restricted to temporally based schedules of reinforcement; no such effect is consistently noted when reinforcement delivery occurs under response-based contingencies. Microanalysis of FI performance reveals that if pause time can be construed as an index of timing behavior, it remains intact. However, once responding begins, re- sponse rates of lead-exposed rats greatly exceed those required by the reinforcement schedule, suggesting as one possibility, a decreased responsiveness of lead-exposed organisms to the feedback generated by their own behavior on the schedule. Higher lead exposure levels produce a generalized nonspecific suppression of responding, which has been noted on the fixed-interval, fixed-ratio, and variable-interval
146 DEBORAH A. CORY-SLECHTA schedules, raising the possibility of underlying motivational factors (e.g., a decline in reinforcer efficacy). Many scientists, including some behavioral scientists have diffi- culty in understanding the rationale for studying schedule-controlled behavior, both because of its less obvious correspondence to human behavior than, for example, conventional learning paradigms, and because of the ostensible difficulty in interpreting response rate changes. It should be emphasized, however, that in the human environment, as well as in the experimental laboratory, rewards or reinforcers for behavioral performances, including learning, occur under various re- inforcement schedules. The human environment obviously entails far greater complexity, with many reinforcement schedules and various reinforcers available concurrently. Nevertheless, human behavior occurs, and is consequated, under schedules of reinforcement. The study of simple schedules of reinforcement in the laboratory represents a simplified approach to such processes in order to provide a more molecular analysis of the variables controlling such performances. Furthermore, the lever-press response used in experimental studies is but an arbi- trarily chosen response. What if, instead, the particular response increased by lead exposure involved time out of seat or time off task in a classroom setting? Increased frequencies of such responses would have obvious detrimental consequences for children's scholastic per- formance. This raises a further issue, namely, that changes in response rate engendered by a toxicant such as lead might interact with, or even underlie, other behavioral deficits. Consider a situation, for example, in which lead-induced increases in rates of responding engender pre- mature responding to stimulus cues in a learning task. The resulting decreased level of accuracy might then be interpreted as a learning impairment. Furthermore, the acquisition of appropriate response patterns on schedules of reinforcement per se may represent learning deficits, as has already been described. As this discussion with respect to an intensively studied neurotoxicant such as lead indicates, much of our work remains at a descriptive stage. Nevertheless, the utilization by experimental animal studies of more sensitive and specific behavioral endpoints than have been incorporated into many of the human studies in this area, has led over the past five years or so to a striking correspondence between results in the two areas (shown in Figure 3) in the reported levels of lead in blood at which behavioral deficits are reported. The more recent studies in humans indicate performance impairments in children at levels as low as 10 ,ug/dL (Bellinger et al., 1987; Fulton et al., 1987~; studies in both rats and nonhuman primates document effects at
BEHAVIORAL TOXICOLOGY STUDIES Animal Data Rice et al., 1979 Rice et al., 1984 Rice, 1988 30 25 Cory-Slechta et al., 1983 20 15 Cory Slechta et al., 1985 Rice, 1985 10 5 To Human Data PbB (~9/dL) 30 Needleman et al., 1979 25 20 15 10 5 o FIGURE 3 Blood-lead levels at which behavioral effects are reported ]47 Shroeder et al., 1985 Shroeder and Hawk, 1986 Yule et al., 1981, 1987 Wigg et al., 1988 Bellinger et al., 1987 Fulton et al., 1987 comparable levels (Cory-Slechta et al., 1985; Rice, 1985~. Figure 3 further illustrates the decline in blood lead concentrations associated with performance effects in both humans and experimental animals over the past several years, as earlier studies paved the way for methodological improvements in subsequent efforts. BARRIERS TO ADVANCEMENT What has constrained advancement and kept much of the research in behavioral toxicology primarily at the level of characterization? Several factors probably play a role. One is that emerging scientific disciplines such as behavioral toxicology require reliable, systematic characterization studies as a sound base for further efforts. Proceeding to mechanistic-based studies in the absence of such information, or on the basis of unreliable information, would be premature and even counterproductive. Besides the relative youth of this area, another factor that may impose difficulties for a more mechanistically based science is the nonspecificity of most neurotoxicants. Many of the chemicals of interest have a diversity of biological effects, including CNS effects. This poses the distinct possibility that, even in the sim-
148 DEBORAH A. CORY-SLECHTA plest case, different behavioral effects of a particular toxicant may arise from different behavioral or biological substrates, i.e., from ef- fects on different neurotransmitter systems or pathological lesions in different brain regions. Thus, to fully delineate the gamut of behav- ioral and biological mechanisms of toxicity for any given neurotoxicant could require an extensive experimental commitment. A further impediment to more rapid progress in understanding the neurobiological substrates of performance impairment has been the relative lack of systematic experimentation aimed at directly defining the relationships between functional consequences and other neurotoxic effects resulting from exposure. Although many studies utililze a multidisciplinary approach, concurrently measuring various indices of behavioral outcome and changes in neurotransmitter levels in response to a neurotoxicant, for example, few studies undertake the types of definitive experiments required to determine the precise nature of such relationships, which then remain correlational in nature. Probably one of the primary factors constraining both the scope and the advancement of behavioral toxicology may be the prepon- derance of "apparatus-driven" research. In many cases, research questions are framed around the available behavioral apparatus within a labo- ratory, be it a radial arm maze or an open field, rather than upon a hypothesis formulated on the basis of the current scientific literature. This can often be noted in the introduction to published studies, which present only the barest approximation to a hypothesis and an obscure rationale, namely, this toxicant may affect that performance. This is likely also to be one of the factors contributing to the frequent shifts in the toxicant of current scientific interest, known colloquially as the poison of the month: if you cannot change the behavioral apparatus, and thus the behavioral question, you are left with changing the toxicant. The latter situation arises, no doubt at least in part, from limitations of equipment availability imposed by funding restrictions over the past several years. However, probably more importantly, it reflects m~umal or inadequate familiarity with the breadth of behavioral sciences in general and with contemporary state-of-the-art behavioral procedures. In addition, it may reflect a prevalent notion among many nonbehavioral scientists that anyone can conduct behavioral testing, a misconception apparently based on the ostensible simplicity of endpoints such as motor activity with which those outside the behavioral field tend to be most familiar. ACCELERATING THE PACE One notion that has been advanced to expedite progress in behavioral toxicology is the development of new behavioral tests. However, the
BEHAVIORAL TOXICOLOGY STUDIES 149 rationale for this argument is not compelling and may even be viewed as counterproductive. For instance, consider the multitude and vari- ety of procedures currently in use to examine learning. The number of new procedures that could be devised just to assess this particular function is almost limitless. However, it must be remembered that every newly developed procedure requires extensive behavioral in- vestigation to ascertain the variables controlling the performance, as well as pharmacological determinations of the comparative sensitivity of the procedure to other learning tasks. Might not a more productive and expeditious approach emphasize systematic and comparative studies of a toxicant's effects across existing, better-understood learning pro- cedures, in terms of both behavioral and pharmacological variables? A second approach to accelerating the pace would be to incorporate some of the more complex, state-of-the-art behavioral paradigms into behavioral toxicology. Consider again the case of learning, a much emphasized component of neurotoxicology and of neuroscience in general. Many of the more conventional techniques suffer from the limitation that once the organism has mastered the problem to be learned, only performance is being measured. For example, animals running a maze may learn to turn toward the appropriate side or color cue quickly. Similarly an organism may learn to lever-press only in the presence of a red light, and not in the presence of green, within only a few experimental sessions. This presents a particular problem in evaluating a toxicant whose effects have a delayed onset or which accumulates only slowly. It also makes the assessment of reversibility of toxicant-induced learning changes difficult. A more complex behavioral procedure known as repeated acquisition, or acquisition of response sequences, originated partly because of such a need (Borer, 1963~. In this particular task, the organism is required to learn a new sequence of responses of fixed length during each experimental session. A schematic of the apparatus configuration for this paradigm is illustrated in Figure 4. Initially, this paradigm engenders quite high error rates, but as the organism gains experience with the task, the error rate stabilizes from session to session and the organism learns each new sequence at a fairly constant rate. The procedure has several distinct advantages over other learning techniques. First, it allows the measurement of learning on a repeated basis, thus providing a stable baseline rate of learning across sessions from which perturbations can be assessed. In addition, the delineation of various classes and patterns of errors following chemical exposure can provide useful information about the type of learning deficit, facets which are relevant both to the issue of behavioral mechanisms of toxicity and to screening and risk assessment. Furthermore, the procedure is often run in conjunction with a per-
150 DEBORAH A. CORY-SLECHTA 120 100 o ~ 80 o C' o 60 lo C, ~ 40 C' 20 o a\ ~ \ ~ B \` \\ \~ ~ A _ l _ · Pet 0 Pigeon I l L I J ~ 0.1 1 10 Pb (mg/kg) 100 1 000 FIGURE 4 Schematic diagram of the intelligence panel used for the repeated acquisi- tion procedure. In the procedure, each three-light unit serves as a discriminative stimulus signaling the next correct response In the sequence of responses leading to reinforcement. The lights have no fixed relationship to the lever; the association be- tween lights and levers changes each session win the new correct sequence of responses. SOURCE: From Pollard et al. (1981). formance component in which the response sequence remains con- stant from experimental session to session. During the daily experi- mental session, the performance component alternates periodically (e.g., every 10 minutes) with the learning component (as defined above). Thus, both learning and performance are being concurrently assessed during an experimental session. This is especially advantageous be- cause the performance component allows the assessment of nonspecific chemical effects, as would be exemplified by changes that occur in both the learning and the performance components, whereas behavioral alterations observed exclusively in the learning component may more specifically reflect changes in learning. Representative performance of a nonhuman primate responding under such a schedule is shown in Figure 5, in which a separation of the effects of d-amphetamine on the learning and the performance components is evident. Finally, the
BEHAVIORAL TOXICOLOGY STUDIES 151 procedure has recently been further amended (Thompson et al., 1986) to include a memory component. This is accomplished by retesting the acquisition of the learning component sequence at various time intervals following the original learning. In spite of the emphasis on toxicant and chemical-induced alter- ations in learning, only two studies to date have utilized the repeated acquisition baseline to evaluate to~c~cant-~nduced lear~ung deficits. Paule and McMillan (1986) employed a variant of this procedure to track the time course of trimethylUn (TMT) effects on learning. By separating the various error components of repeated acquisition performance in their analysis, a differential time course of TMT on various classes of errors was noted. It showed that early responses in the sequences .,, MONKEY EV / / 8~ L / 8 / V' d Patina A · SPA 1 mg/kg L d-Amphotamine / / ~1 /1 (j L ant,'// / / FIGURE 5 Cumulative records illustrating the effects of the administration of d-am- phetamine on the performance of a monkey working on a multiple schedule which alternated repeated acquisiton or learning (L) and performance (P) components. Am- phetamine affected behavior primarily during the learning component, increasing the number of errors in a dose-related fashion, while leaving the performance component relatively intact. SOURCE: From Thompson and Moerschbacher (1979).
152 DEBORAH A. CORY-SLECHTA were disrupted to a greater extent by TMT than were later stages of the sequences, suggesting an effect on learning, whereas the recall necessary for the longer sequences remained relatively intact. In an earlier study, Anger and Setzer (1979) reported that intramuscular carbaryl administration increased both session time and error rates of monkeys working on a four-response sequence repeated acquisition baseline. The acquisition of response sequences serves as but one example of a more sophisticated procedure for evaluating learning deficits. It emphasizes the point that increased awareness of advancements in behavioral sciences can accelerate the progression of behavioral toxicology toward its ultimate goal. One reason for the hesitancy on the part of many investigators to turn to such techniques may be the more extensive training required to produce stable baseline performances than are required when using simpler procedures. Given the potential of these techniques, however, future work designed to accelerate the training process would be of great benefit to behavioral pharmacology and toxicology. Another strategy useful to facilitate cross-species extrapolation, that is, to bridge experimental animal and human behavioral toxicology studies, is to strongly emphasize the use of behavioral paradigms that are directly applicable to both populations. Operant schedules of reinforcement exemplify one class of such baselines. Innumerable studies have documented the comparability of schedule-controlled performance in a wide variety of species, including humans (e.g., Dews and Wenger, 1977; Holland, 1958; Kelleher and Morse, 1969; Laties and Weiss, 1963; Richelle, 1969; Tews and Fischman, 1982), a point documented by Figure 6. Behavioral pharmacology studies have further shown the similarity across species of drug effects on schedule- controlled behavior. In fact, the study of schedule-controlled behavior was a strategy suggested by Bornschein et al. (1980) to enhance direct extrapolation of lead-induced behavioral impairments across species. Some of the more advanced complex techniques may be even more applicable. The repeated acquisition paradigm described above has been used to study learning in rodents (Schrott et al., 1980), pigeons (e.g., Thompson, 1980), and nonhuman primates (Moerschbacher and Thompson, 1980), as shown in Figure 5. The technique has also been utilized with human populations, including those with developmen- tal disabilities (Suessbrick, 1983) and, more recently, victims of Alzheimer's disease (Gershensen, Thompson, and Gisselquist, personal communication, 1988~. Figure 7 compares the decline in the number of errors on a five-link response sequence in normal elderly humans (top panel) to those with Alzheimer's disease (bottom panel). The greater number
BEHAVIORAL TOXICOLOGY STUDIES 153 , | 8 O-~Min R Gamin F! ~ 50-Min ~ / l i} [ // a// J Gamin Fl Gamin Fl Pet Lever Water Rat Lever Food Pigeon Key Food Pigeon Key Water ~ ~ ~~f ~ i' ~ 8 i ~ §5 '~ a i/ &( Pat Lear Food ~ | Rat Wh66' ~ O t Mob fir ~ fir ~ Chimpanzee Lever Food ~ ~ ~ &5 ~ am/ ~A a/ §/ ! I 15 min FIGURE 6 Generality of fixed-interval (PI) performance under various conditions. The ordinate shows the cumulative number of responses, whereas time is represented on the abscissa. The typical performance is characterized by little or no responding after reinforcement, followed by a gradually accelerating rate of responding. In all examples, a fixed-interval schedule of food or water presentation was in effect. The lower left frame shows performance under a 10-minute fixed-interval schedule. Each reinforcement delivery resets the recording pen to the baseline. The comparability of performance of the pigeon, rat, and chimpanzee, either pecking a key or pressing a lever, is evident. Likewise, during a 5-minute fixed-interval schedule (lower right frame), the comparability of performance of rats, pigeons, and cats pecking a key, pressing a lever, wheel-running, or pulling a knob is illustrated. SOURCE: From Kelleher and Morse (1969). of errors and slower decline in error rates in the patients with Alzheimer's are evident. Thus, although admittedly difficult to train, the response sequence paradigm shows comparable baseline performance across species and evidences sensitivity both to drugs and to neurodegenerative disease. Other learning and memory paradigms with direct cross- species applicability include procedures such as delayed alternation and delayed matching to sample. Obviously, the use of such behav- ioral tasks in the case of human evaluation may not be easily implemented
154 50 40 30 10 Contrd #1-Total Baseline O ~~ ,.~ ~ ~ 0 10 20 30 TRIAL 60 co 50 o fir 40 LU LL 0 30 20 he 10 30 - Alz #1-Baseline, 5 Unks 40 on Q en O \ O 30 ~ 20 _ ~ ~ O \ O 20 10 _ ~ ~d o DEBORAH A. CORY-SLECHTA Contrd #2-Total Baseline O ~ 0 10 20 30 TRIAL Alz #2-Baseline, 5 Unks _~\ 10 l~:44~ ol ' 0 10 20 0 10 20 TRIAL TRIAL FIGURE 7 Performance of normal elderly humans (top panels) and Alzheimer's pa- tients (Alz; bottom panels) working on a five-link repeated acquisition paradigm. Total number of errors over successive trials of a session are presented. The rapid decline in errors during initial trials of the session and the low subsequent errors over subsequent trials of normals are in stark contrast to the slower initial decline in errors and the higher sustained rate of errors over the remainder of the session in the Alzheimer's patients. 1988). SOURCE: From Gershensen, Thompson, and Gisselquist (personal communication, when large sample populations are involved. However, one initial alternative is to study more intensively a smaller proportion of the more highly exposed individuals within the population. Although multidisciplinary studies pave the way for establishing biological mechanisms, most stop short of experimentally evaluating and defining the nature of the connection between various param- eters of neurotoxicity. For example, a toxicant may induce changes in some aspect of behavior, as well as decrease the levels of some neurotransmitter. It is common to hypothesize a causal relationship between the two, but far less frequent to test such a hypothesis di- rectly. More systematic research exploring the relationships between behavioral toxicity and other parameters of neurotoxicity would expedite
BEHAVIORAL TOXICOLOGY STUDIES 155 our understanding of the biological bases of toxicant-induced behav- ioral effects. Finally, it should be remembered that both experimental animal and human behavioral toxicology studies have common goals: to understand the mechanisms by which toxicants affect behavior and to implement procedures to screen for neurotoxic properties of chemicals. A better base of communication and interfacing between human and experimental animal research would accelerate progress on both fronts, allowing each to benefit from advances made by the other. REFERENCES Angell, N. F. and B. Weiss. 1982. Operant behavior of rats exposed to lead before or after weaning. Toxicol. Appl. Pharmacol. 63:62-71. Anger, W. K., and J. V. Setzer. 1979. Effects of oral and intramuscular carbaryl administration on repeated chain acquisition in monkeys. J. Toxicol. Environ. Health 5:793-808. Avery, D. D., and H. A. Cross. 1974. The effects of tetraethyl lead on behavior in the rat. Pharmacol. Biochem. Behav. 2:473~79. Barthalmus, G. T., J. D. Leander, D. E. McMillan, P. Mushak, and M.R. Krigman. 1977. Chronic effects of lead on schedule-controlled pigeon behavior. Toxicol. Appl. Pharmacol. 42:271-284. Bellinger, D., A. Leviton, C. Waternaux, H. Needleman, and M. Rabinowitz. 1987. Longitudinal analyses of prenatal and postnatal lead exposure and early cognitive development. N. Engl. J. Med. 316:1037-1043. Boren, J. J. 1963. Repeated acquisition of new behavioral chains. Am. Psychol. 17:421. Bornschein, R., D. Pearson and L. Reiter. 1980. Behavioral effect of moderate lead exposure in children and animal models: Part 2, Animal studies. (CRC Critical Revews in Toxicology 7:101-152. Brady, K., Y. Herrera, and H. Zenick. 1975. Influence of parental lead exposure on subsequent learning ability of offspring. Pharmacol. Biochem. Behav. 3:561-565. Brown, D. R. 1975. Neonatal lead exposure in the rat: Decreased learning as a function of age and blood lead concentration. Toxicol. Appl. Pharmacol. 32:628- 637. Brown, S., N. Dragann and W. H. Vogel. 1971. Effects of lead acetate on learning and memory in rats. Arch. Environ. Health 22:370-372. Bullock, J. D., R. L. J. Wey, J. A. Zaia, I. Zarembok, and H. A.Schroeder. 1966. Effect of tetraethyl lead on learning and memory in the rat. Arch. Environ. Health 13:21- 22. Bushnell, P. J., and R. E. Bowman. 1979a. Reversal learning deficits in young monkeys exposed to lead. Pharmacol. Biochem. Behav. 10:733-742. Bushnell, P. J., and R. E. Bowman. 1979b. Persistence of impaired reversal learning in young monkeys exposed to low levels of dietary lead. J. Toxicol. Environ. Health 5:1015-1023. Bushnell, P. J. R. E. Bowman, J. R. Allen, and R. J. Marler. 1977. Scotopic vision deficits in young monkeys exposed to lead. Science 196:333-335. Carson, T. L. 1976. Auditory and Visual Discrimination Learning in Sheep Prenatally and Postnatally Exposed to Lead. Ph.D. thesis, Iowa State University, Ames.
156 DEBORAH A. CORY-SLECHTA Carson, T. L., G. A. Van Gelder, G. C. Karas, and W. B. Buck. 1974. Slowed learning in lambs prenatally exposed to lead. Arch. Environ. Health 29:154-156. Cory-Slechta, D. A. 1984. The behavioral toxicity of lead: Problems and perspectives. Pp. 211-255 in Advances in Behavioral Pharmacology, Vol. 4, T. Thompson, P. B. Dews, and J. E. Barrett, eds. New York: Academic Press. Cory-Slechta, D. A. 1986. Effects of lead exposure beyond weaning on fixed ratio performance. Neurobehav. Toxicol. Teratol. 8:237-244. Cory-Slechta, D. A. 1989. The lessons of lead for behavioural toxicology. In Lead Exposure and Child Development: An International Assessment. M. A. Smith, L. D. Grant, and A. I. Sors, eds. Dordrecht: Kluwer Academic Publishers. Cory-Slechta, D. A. and T. Thompson. 1979. Behavioral toxicity of chronic postweaning lead exposure in the rat. Toxicol. Appl. Pharmacol. 47:151-159. Cory-Slechta, D. A., B. Weiss, and C. Cox. 1983. Delayed behavioral toxicity of lead with increasing exposure concentration. Toxicol. Appl. Pharmacol. 71:342-352. Cory-Slechta, D. A., B. Weiss, and C. Cox. 1985. Performance and exposure indices of rats exposed to low concentrations of lead. Toxicol. Appl. Pharmacol. 78:291-299. Dews, P. B., and G. Wenger. 1974. Rate dependency of the behavioral effects of amphetamine. Pp. 167-227 in Advances in Behavioral Pharmacology, Vol. 1, T. Thompson and P. B. Dews, eds. New York: Academic Press. Dietz, D. D., D. E. McMillan, and P. Mushak. 1979. Effects of chronic lead administration on acquisition and performance of serial position sequences by pigeons. Toxicol. Appl. Pharmacol. 47:377-384. Driscoll, J. W. and S. E. Stegner. 1976. Behavioral effects of chronic lead ingestion on laboratory rats. Pharmacol. Biochem. Behav. 4:411-417. Fox, D. A., and L. Chu. 1988. Rods are selectively altered by lead: II. Ultrastructure and quantitative histology. Exp. Eye Res. 46:613~25. Fox, D. A., and D. B. Farber. 1988. Rods are selectively altered by lead: I. Electrophysiology and biochemistry. Exp. Eye Res. 46:597-611. Fox, D. A., A. A. Wright, and L. G. Costa. 1982. Visual acuity deficits following neonatal lead exposure: Cholinergic interactions. Neurobehav. Toxicol. Teratol. 4:689-693. Fulton, M., G. Thompson, R. Hunter, G. Raab, D. Laxen, and W. Hepburn. 1987. Influ- ence of blood lead on the ability and attainment of children in Edinburgh. Lancet 1:1221-1226. Geist, C. R., and B. R. Mattes. 1979. Behavioral effects of postnatal lead acetate exposure in developing laboratory rats. Physiol. Psychol. 7:399-402. Hastings, L., G. Cooper, R. L. Bornschein, and I. A. Michaelson. 1977. Behavioral effects of low level neonatal lead exposure. Pharmacol. Biochem. Behav. 7:37~2. Hastings, L., G. Cooper, R. L. Bornschein, and I. A. Michaelson. 1979. Behavioral deficits in adult rats following neonatal lead exposure. Neurobehav. Toxicol. 1:227- 231. Holland, J. G. 1958. Human vigilance. Science 128:61-67. Kelleher, R. T., and W. H. Morse. 1969. Determinants of the behavioral effects of drugs. Pp. 383005 in Importance of Fundamental Principles in Drug Evaluation, D. J. Tedeschi and R. E. Tedeschi, eds. New York: Raven Press. Klein, A. W., R. T. Louis-Ferdinand, and D. R. Brown. 1977. Measurement of behav- ioral, chemical and anatomical changes in the developing rat brain from exposures to low levels of lead. Anat. Rec. 187:626. Lanthorn, T. and R. L. Isaacson. 1978. Effects of chronic lead ingestion in adult rats. Physiol. Psych. 6:93. Laties, V. G. 1975. The role of discriminative stimuli in modulating drug action. Fed. Proc. 34:1880-1888.
BEHAVIORAL TOXICOLOGY STUDIES 157 Laties, V. G., and B. Weiss. 1963. Effects of a concurrent task on fixed-interval re- sponding in humans. J. Exp. Anal. Behav. 6:431-436. Laughlin, N., R. E. Bowman, E. D. Levin, and P. J. Bushnell. 1983. Neurobehavioral consequences of early exposure to lead in rhesus monkeys: Effects on cognitive behaviors. Pp. 497-516 in Reproductive and Developmental Toxicity of Metals, T. W. Clarkson, G. F. Nordberg and P. R. Sager, eds. New York: Plenum. Levin, D. E., and R. E. Bowman. 1983. The effect of pre- or post-natal lead exposure on Hamilton search task in monkeys. Neurobehav. Toxicol. Teratol. 5:391-394. Moerschbacher, J. M., and D. M. Thompson. 1980. Effects of d-amphetamine, cocaine, and phencyclidine on the acquisition of response sequences with and without stimulus fading. J. Exp. Anal. Behav. 33:369-381. Nation, J. R., A. E. Bourgeois, and D. E. Clark. 1983. Behavioral effects of chronic lead exposure in the adult rat. Pharmacol. Biochem. Behav. 18:833-840. Needleman, H. D., C. E. Gunnoe, A. Leviton, R. Reed, H. Peresie, C. Maher, and P. Barrett. 1979. Deficits in psychologic and classroom performance of children with elevated lead levels. N. Engl. J. Med. 300:689-695. Ogilvie, D. M. 1977. Sublethal effects of lead acetate on the Y maze performance of albino mice (Mus musculus L.). Can. J. Zool. 55:771-775. Ogilvie, D. M. 1978. Effect of lead acetate on memory in mice. Bull. Environ. Contam. Toxicol. 19:143-146. Overmann, S. R. 1977. Behavioral effects of asymptomatic lead exposure during neonatal development in rats. Toxicol. Appl. Pharmacol. 41:459~71. Padich, R., and H. Zencik. 1977. The effects of developmental and/or direct lead exposure on FR behavior in the rat. Pharmacol. Biochem. Behav. 6:371-375. Paule, M. G., and D. E. McMillan. 1986. Effects of trimethyltin on incremental re- peated acquisition (learning) in the rat. Neurobehav. Toxicol. Teratol. 8:245-254. Penzien, D. A., D. R. C. Scott, and J. P. Motiff. 1982. The effects of lead toxication on 1eaMing in rats. Arch. Environ. Health 37:83-87. Pollard, G. T., S. T. McBennet, K. V. Rohrbach, and J. L. Howard. 1981. Repeated acqusition of three-response chains for food reinforcement in the rat. Drug Develop. Res. 1:67-75. Rice, D. C. 1984. Behavioral deficits (delayed matching to sample) in monkeys ex- posed from birth to low levels of lead. Toxicol. Appl. Pharmacol. 75:337-345. Rice, D. C. 1985. Chronic low-level lead exposure from birth produces deficits in discrimination reversal in monkeys. Toxicol. Appl. Pharmacol. 77:201-210. Rice, D. C. 1988. Schedule-controlled behavior in infant and juvenile monkeys exposed to lead from birth. Neurotoxicol. 9:75-88. Rice, D. C, and S. G. Gilbert. 1985. Low lead exposure from birth produces behavioral toxicity (DRL) in monkeys. Toxicol. Appl. Pharmacol. 80:421-426. Rice, D. C., and R. F. Willes. 1979. Neonatal low-level lead exposure in monkeys (Macacafascicularis): Effect on two choice non-spatial form discrimination. J. Environ. Pathol. Toxicol. 2:1195-1203. Rice, D. C., S. G. Gilbert, and R. F. Willes. 1979. Neonatal low-level lead exposure in monkeys: Locomotor activity, schedule controlled behavior, and the effects of am- phetamine. Toxicol. Appl. Pharmacol. 51:503-513. Richelle, M. 1969. Combined action of diazepam and d-amphetamine on mixed interval performance in cats. Exp. Anal. Behav. 12:989-998. Rosen, J. B., R. F. Berman, F. C. Beuthin, and R. T. Louis-Ferdinand. 1985. Age of testing as a factor in the behavioral effects of early lead exposure in rats. Pharmacol. Biochem. Behav. 23:49-54. Schroeder, S. R., and B. Hawk. 1986. Child-caregiver environmental factors related to lead exposure and IQ. In Toxic Substances and Mental Retardation, S. R. Schroeder, ed. AAMD Monograph, Washington, D.C.
158 DEBORAH A. CORY-SLECHTA Schroeder, S. R., B. Hawk, D. A. Otto, P. Mushak, and R. E. Hicks. 1985. Separating the effects of lead and social factors on IQ. Environ. Res. 38:144-154. Schrott, J., J. R. Thomas, and R. A. Banaurd. 1980. Repeated acquisition of four- member response sequences in rats. Psych. Reports 47:503-509. Snowdon, C. J. 1973. Learning deficits in lead-injected rates. Pharmacol. Biochem. Behav. 1:599~03. Suessbrick, A. 1983. A procedure for studying repeated learning by mentally retarded individuals. Dissertation Abstracts International 43(8-B):2733. Tews, P. A., and M. W. Fischman. 1982. Effects of d-amphetamine and diazepam on fixed-interval, fixed-ratio responding in humans. J. Pharmacol. Exp. Therap. 221:373- 383. Thompson, D. M. 1980. Selective antagonism of the rate-decreasing effect of d-am- phetamine by chlorpromazine in a repeated acquisition task. J. Exp. Anal. Behav. 34:87-92. Thompson, D. M., and J. M. Moerschbacher. 1979. Drug effects on repeated acquisi- tion. Pp. 229-259 in Advances in Behavioral Pharmacology, Vol. 2, T. Thompson and P. B. Dews, eds. New York: Academic Press. Thompson, D. M., J. Mastropaolo, P. J. Winsauer, and J. M. Moerschbacher. 1986. Repeated acquisition and delayed performance as a baseline to assess drug effects on retention in monkeys. Pharmacol. Biochem. Behav. 25:201-207. Van Gelder, G. A., R. Carson, R. M. Smith, and W. B. Buck. 1973. Behavioral toxico- logic assessment of the neurologic effect of lead in sheep. Clin. Toxicol. 6:405-418. Wigg, N. R., G. V. Vimpani, A. J. McMichael, P. A. Baghurst, E. F. Robertson, and R. J. Roberts. 1988. Port Pirie Cohort Study: Childhood blood lead and neuropsychological development at age two years. J. Epid. Comm. Health 42:213-219. Winneke, G., A. Brockhaus, and R. Baltissen. 1977. Neurobehavioral and systemic effects of long-term blood lead elevation in rats. I. Discrimination learning and open field behavior. Arch. Toxicol. 37:247-263. Winneke, G., H. Lilienthal, and W. Werner. 1982. Task dependent neurobehavioral effects of lead in rats. Arch. Toxicol. Suppl. 5:84-93. Yule, W., R. Landsdown, I. B. Millar, and M. A. Urbanowicz. 1981. The relationship between blood lead concentrations, intelligence and attainment in a school population: A pilot study. Dev. Med. Child. Neurol. 23:567-576. Yule, W., M. A. Urbanowicz, R. Lansdown, and I. B. Millar. 1984. Teachers' ratings of children's behavior in relation to blood lead levels. Br. J. Dev. Psychol. 2:295-305. Zenick, H., R. Padich, T. Takarek, and P. Aragon. 1978. Influence of prenatal and postnatal lead exposure on discrimination learning in rats. Pharmacol. Biochem. Behav. 8:347-350. Zenick, H., W. Rodriquez, J. Ward, and B. Elkington. 1979. Deficits in fixed-interval performance following prenatal and postnatal lead exposure. Dev. Psychobiol. 12:509-514.
