Persistent effects of manganese on effortful responding and their relationship to manganese accumulation in the primate globus pallidus

Persistent effects of manganese on effortful responding and their relationship to manganese accumulation in the primate globus pallidus

TOXICOLOGY AND APPLIED PHARMACOLOGY 113,87-97 (1992) Persistent Effects of Manganese on Effortful Responding and Their Relationship to Manganese ...

2MB Sizes 104 Downloads 31 Views

TOXICOLOGY

AND

APPLIED

PHARMACOLOGY

113,87-97 (1992)

Persistent Effects of Manganese on Effortful Responding and Their Relationship to Manganese Accumulation in the Primate Globus Pallidus M. CHRISTOPHERNEWLAND’ AND BERNARD WEISS Environmental Health Sciences Center, University of Rochester, Rochester, New York 14642 Received May 2, 1991;accepted November26, I991

PersistentEffects of Manganeseon Effortful Respondingand Their Relationship to ManganeseAccumulation in the Primate Globus Pallidus. NEWLAND, M. C., AND WEISS,B. (1992). Toxicol. Appl. Pharmacol. 113, 87-97. Manganese produces signs and symptoms that suggestinvolvement of the basal ganglia, especially the globus pallidus and substantianigra. Overt neurologicalsignshavebeenreported in primates exposedto high levelsof manganese(over 100 mg/ kg) but little is known about the effects of lower doses.To examinetheseissues,three cebusmonkeyswere trained to operate a responsedevicewith their armsand legsby executing a rowinglike movement against a 3.9- to 4.1-kg spring through an arc length of 10 cm under a multiple fixed-ratio fixed-interval scheduleof reinforcement. Over the course of 450 days, these monkeys were administered acute dosesof 5 or 10 mg/kg iv of manganesechloride using a multiple baselineexperimental design. Dosesas low as 5 mg/kg provoked a large increasein the numberof incompleteresponses.The onsetof manganese’s effect appearedwithin days of exposureand developedover the course of several weeks. Its magnitude declined over the course of months, but after a cumulative dose of 10 to 40 mg/kg it did not return to baseline.Action tremor appeared at cumulative dosesgreater than 40 mg/kg and dystonia was never observed at the cumulative dosesexamined. Behavioral microanalysisrevealed that manganese’seffects initially appearedas increased variability of interresponsetimes and responseduration. Later, the responsepattern during the fixed ratio component shifted to one of progressivelyincreasingdurations through the course of the ratio. Magnetic resonanceimaging revealed that the behavioral effects of manganesecorrespondedto an apparent increasein the manganesecontent of the globuspallidus and substantia nigra. 0 1992 Academic Press, Inc.

INTRODUCTION

The basal ganglia are associated with the regulation of movement and the organization and expression of behavior sequences (de Long et al., 1984; Eva& and Wise, 1984). The ’ To whom correspondence shouldbe addressed at: Department of Psychology, Auburn University, Auburn, Alabama 36849.

87

various functions of the basal ganglia have been illuminated by compounds such as 1-methyld-phenyl1,2,3,6-tetrahydropyridine (MPTP) and by disease entities such as HallervordenSpatz syndrome and Parkinson’s disease that selectively lesion specific regions or chemical pathways and induce characteristic neurological signs (e.g., Langston, 1987; Marsden and Fahn, 1987). Manganese produces a constellation of signs and symptoms that strongly suggests basal ganglia involvement. Neuropathological changes associated with manganese exposure (Edsall et al., 19 19; Cook et al., 1974; Barbeau et al., 1976) and the distribution of manganese in the nervous system after high-dose exposure (Newland et al., 1989; Suzuki et al., 1975) confirm basal ganglia involvement in manganism. Indeed, manganism was once thought to elicit an entity mimicking Parkinson’s disease, a classic basal ganglia disorder (Cotzias, 1958; Mena et al., 1969). However, close examination of the neurological signs, the pathology associated with the two syndromes (Barbeau et al., 1976; Barbeau, 1984), the specificity of the MPTP model of Parkinson’s disease (e.g., Langston, 1987) and a lack of manganese accumulation in cases of Parkinson’s disease (Yamada et al., 1986; but see Bernheimer et al., 1973) all point to different neural substrates of manganism and Parkinson’s disease. In humans, manganism is often described as a three-stage process (Barbeau, 1984; Canavan et al., 1934; Cawte, 1985; Cook et al., 1974). The earliest stage lasts for weeks to months and presents as mania, hysterical laughter and crying, insomnia, and hypersexuality. Stage 2 is reflected by clumsiness, fatigue, and an expressionless face. The final stage, which may not appear for many years after the onset of exposure, presents as irreversible dystonia and hyperflexion of muscles (Cotzias et al., 1968). This description may be overly simple, as sometimes the psychological disturbances do not appear at all or appear simultaneously with the overt neurological signs. The long time course of manganese poisoning is particularly notable because of the slow rate at which manganese enters and is cleared from the central nervous system (Cotzias et al., 1968; Dastur et al., 197 1; Drown et al., 1986; Mena et al., 1967; Newland et al., 1987, 1989).

0041-008X/92

$3.00

Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.

88

NEWLAND

Although_ the papers noted above, combined with visible consequences of exposure to high levels of manganese in monkeys, portray the late stages of manganism, the manner in which functional effects develop and their relationship to manganese in the central nervous system remain obscure. The present experiments trace the behavioral effects of manganese at cumulative doses below those evoking clinical neurological signs. These effects are related to the apparent accumulation of manganese in the basal ganglia, especially the globus pallidus, as detected in vivo with magnetic resonance imaging (Newland et al., 1989). A behavioral preparation that enabled the simultaneous examination of effortful responding and schedule-controlled behavior was used. Cebus monkeys pulled against a spring that resisted extension with a force that approximated their body weight. This behavior was maintained by a multiple fixed-ratio fixed-interval schedule of reinforcement (Newland and Weiss, 1990). METHODS Su/+crs. Three wild-caught, male adult cebus monkeys, CM842, CM846, and CM904 weighing 3.2, 3.7, and 3.9 kg, respectively, were used. The monkeys were housed in individual cages in an AAALAC-approved facility and allowed free accessto Purina Monkey Chow. They were water deprived at about 1600 hours and tested the morning of the following day. CM842 and CM904 had been used in previous experiments (Newland and Weiss, 1990) in which they received acute doses of d-amphetamine, pentobarbital, bromocriptine, benztropine, and levodopa/carbidopa at least 6 months before the beginning of the present experiment. These drugs induced only acute behavioral effects. Apparatus. The response device is illustrated in Newland and Weiss ( 1990). Monkeys perched in a Plexiglas primate chair that was open at the front and provided only neck restraint. The chair was placed in front of a response device that the monkey could pull with its arms and push with its legs, executing a motion resembling that produced by a stationary rowing machine. The manipulandum was attached to a constant force, spring-driven reel. To be counted as a response, the device had to be moved through an angle of about 45”, corresponding to an arc length of 10 cm. This permitted nearly full extension of the arms at the beginning of the response and substantial extension of the legs at the peak of the response. The spring resisted movement in one direction with a force of 39-41 N (corresponding to 3.9 to 4.1 kg mass), and it returned the lever to resting position when the animal released the bar. Monkeys were tested in a sound-attenuated chamber, 171 by 54 by 86 cm (H X W X D, inside dimensions). Air was circulated by connecting a floor vent spanning the back of the chamber to a circulating exhaust system. This pulled fresh air through a 7-cm shaft situated at the top of the chamber. All events were controlled and monitored with 0.01~set resolution by a PDP 11 computer running the SKEDll behavioral programming language (Snapper et al., 1982). Procedure. The operant response consisted of moving the lever through an arc of 10 cm and allowing it to return to the home position. Responses that did not meet this criterion were “incomplete responses.” Limit switches detected the extremes of movement and a 0.1 -set tone burst from a Sonalert sounded when the top of the arc was reached. This discriminative stimulus was presented at the top of the arc to maximize its contiguity with the point at which the lever was fully extended. The tone sounded only if the response began from the home position. The reinforcement cycle consisted of a 1.5 set train of Sonalert bursts (0.25 set on, 0.25 set off) followed by 1 ml of

AND WEISS fruit juice delivered in 0.1 set through a spout located in front of the monkey’s mouth. The reinforcement cycle, like the O.l-set tone burst, began at the top of the arc if the response began from the home position. Final performance was maintained by a multiple fixed-ratio 20 fixedinterval 90-set (Mult FR 20 FI 90 set) schedule of reinforcement for CM842 and CM904. During the fixed ratio schedule the reinforcement cycle was triggered after 20 responses were completed. During the fixed-interval 90set schedule the reinforcement cycle was initiated by the first response to occur after 90 set had elapsed. Three white lights situated in front of the animal were lit during the Fl schedule and were dark when the FR schedule was in effect.Training is described in Newland and Weiss (1990). For CM846, final performance was maintained by a Mult FR 10 Fl90 set schedule of reinforcement. During training, this monkey ceased responding when the ratio requirement exceeded 10 responses. Because its behavior under the Mult FR 10 Fl 90 set resembled that of the other two monkeys under a Mult FR 20 Fl 90 set, and was profoundly different at higher FR values, it was maintained at the Mult FR 10 FI 90 set schedule. Dosing. All doses were administered iv, in a sterile solution, into a tail or leg vein, under the supervision of a veterinarian. Initially, concentrations of 1 mg/ml of MnClr (calculated as the salt) were used but, later, more dilute solutions were administered. The manganese solution was administered in small boli followed by saline flush over the course of 20 to 30 min. The total dosage administered during a single session was either 5 or 10 mg/kg of manganese (calculated as the base). The monkey was sedated with ketamine throughout administration. Dosings were separated by at least 1 week, and typically longer, depending upon the animals behavior. Sham exposure to an equal volume of isotonic saline administered over 20 to 30 min served as a control procedure. Because of the lengthy development of manganese neurotoxicity, the animals were dosed in a staggered manner. Thus, when CM904 received the initial series of manganese doses CM842 and CM846 served as concurrent controls. CM842 next received a series of doses while CM846 served as a concurrent control. Magnetic re.ronunce imaging. Magnetic resonance images were produced with a General Electric 2-T (proton frequency of 85.56 MHz) magnetic resonance spectrometer with a 22-cm working bore. The animal’s head was centered in a 15-cm-diameter birdcage coil manufactured by General Electric NMR. The coil was centered horizontally in the magnetic field. Images with a slice thickness of 2 mm were obtained using a 90-(TE/2)- 180-(TE/2)-echo spin-echo imaging sequence (Newland et al., 1989). Tl-weighted images using a repetition (TR) of 450 ms and an echo delay (TE) of 16 ms were used. For additional details see Newland et al. (1989).

RESULTS Figure 1 shows the number of incomplete responses over the course of 450 days. The baseline for monkeys CM842 and CM904 extends back for more than 300 days before the first day shown on the graphs and was stable during this time. The multiple baseline design is evident in this figure. When CM904 was administered manganese, CM842 and CM846 were not treated. Thus, CM904 can be compared both with its own pretreatment baseline and with the other two monkeys. A similar logic can be applied to CM842 and CM846. One sham exposure for CM842 occurred on the same day as a 5 mg/kg exposure for CM904 and one sham exposure for CM846 occurred on the same day as a 5 mg/ kg exposure for CM842. An elevated number of incomplete responses during the FR component

appeared

in all three

monkeys

after man-

MANGANESE

AND EFFORTFUL

89

RESPONDING

I

1-B

CM642

A

CM646

Day Number FIG. 1. Number of incomplete responses during the FR component (left) and FL component (right) for each monkey (rows). The scales on the ordinates were adjusted to maximize visibility. Dotted lines show the 5th and 95th percentiles taken from baseline sessions. Day 0 for each abscissa is the same day, so each abscissa is scaled equivalently in real time. A 0 immediately under the line in each graph shows sham administration of manganese; 5s and 10s show days when 5 or 10 mg/kg of manganese, respectively, were administered. An “A” at the top of a graph indicates days when magnetic resonance images shown in Fig. 2 were taken. ‘3” indicates days when images shown or described in Newland et al. (1989) were taken.

90

NEWLAND

ganese administration (Fig. 1). CM904 showed an increase in incomplete responses from about 5- 15 during control sessions to peaks of 300 to over 400 incomplete responses, a 20- to go-fold increase. This elevation in the number of incomplete responses was most notable after a second 10 mg/ kg dose, and it remained elevated over the next two doses of 5 mg/kg. The elevation in incomplete responding declined over the course of about 50 days but it never returned to pretreatment baseline. The first increase corresponded to the appearance of tremor in the lower legs. A second increase in incomplete responses during the FR component appeared about 100 days after the termination of administration, and its appearance corresponded to the appearance of action tremor in the arms. Two more doses were administered at about 260 days: the first had little effect and the second was lethal. Sham treatment had no effect on CM842 but a single dose of 5 mg/kg evoked a gradual increase in incomplete responses during the FR component. A second dose of 5 mg/kg 33 days later provoked wide fluctuations in incomplete responses over the next 70 days. The number of incomplete responses then declined but never returned to baseline levels. No tremor or other neurological signs were noted during this period. The dosing sequence beginning at Day 260 resulted in another increase in incomplete responding that declined gradually after dosing ended. Action tremor was first noticed in CM842 at about Day 300, shortly after a cumulative dose of 40 mg/kg. The pattern of CM846’s responding resembled that of the other two monkeys but differed in its overall sensitivity to manganese and in the duration that the number of incomplete responses remained elevated. Intermittent bouts of action tremor were first noticed in this monkey at about Day 310, and they appeared more persistently after Day 340 and a cumulative dose of 50 mg/kg. The FI component was less sensitive to manganese than the FR component for all three monkeys. Only the highest acute dose, 10 mg/kg, provoked an increase in incomplete responding during this component. This increase appeared once in CM904 and CM842, but not in CM846. Figure 2 shows magnetic resonance images of CM846 before manganese exposure and on Days 176, 2 11, and 350. A dose of 5 mg/kg resulted in darkening of the globus pallidus (middle row, left) and substantia nigra (middle row, right) on Day 2 11. This darkening corresponds to an increased spin-lattice relaxation time, an effect consistent with an increased concentration of a paramagnetic material, such as manganese. Later, a series of 10 mg/kg doses ultimately resulted in darkening of all of the basal ganglia on Day 350 (bottom row). Magnetic resonance images of CM904 can be seen in Newland et al. (1989). This paper also shows time- and doserelated changes in Tl in different brain regions for CM904 and CM846.

AND

WEISS

An increased number of incomplete responses during the FR component might be expected to lengthen the time required to complete the fixed ratio requirement. This appeared to be the case for CM904, but no consistent changes appeared in CM842 or CM846 (Fig. 3). Figure 3 shows run time (time to complete the FR requirement once responding began) during the same time period displayed in Fig. 1. Figure 3 also shows the mean number of responses during the FI component. FI rate decreased for CM904 but recovered shortly after exposure ended. No clear decrease in rate appeared for CM842 and CM846. Figure 4 shows computer-generated cumulative records from a representative control session and from a session 8 days after administration of 10 mg/kg of manganese. The pattern of control responding and the major effects of manganese are evident in these records. Control performance under the FR schedule is characterized by a pause followed by an unbroken train of high-rate responding (Newland and Weiss, 1990). The FI schedule maintained a much lower rate of responding: only a few complete responses (those extending the full displacement) and more incomplete responses occurred in this component during control sessions. The infrequent incomplete responses that occurred during the FR component tended to occur at the beginning of the component. The major effects of manganese are also evident in Fig. 4. Many more incomplete responses occurred during the FR component after manganese exposure. Approximately the same number of incomplete responses occurred during the FI component before and after exposure. However, response rate (which is reflected by the slope of the cumulative record) under the FR schedule was not substantially affected despite the large number of incomplete responses, except that toward the end of the session FR responding was interrupted by long pauses. FI response rate was reduced after manganese for this monkey. Often, only the single response required for reinforcement was emitted during this component. No neurological signs such as tremor or dystonia were evident at this time. The undulating pattern of incomplete responses seen in CM904’s data prompted an examination of his performance in greater detail. Inspection of daily records for this monkey revealed that the increase in incomplete responses corresponded to a shift in the distribution of both the interresponse times (IRT, the time between the end of one response and the beginning of the next) and the response durations (the time required to execute a complete response). These data are not shown. These inspections revealed that, during the FR component, long IRTs increased in frequency but no change appeared in the median IRT. The long IRTs, which can be categorized as pauses, correspond to the flat periods in the cumulative records in Fig. 1. The distribution of response durations showed a more systematic shift toward

MANGANESE

AND EFFORTFUL

RESPONDING

91

FIG. 2. Tl-weighted magnetic resonance images from Days 168, 175, and 349 for CM846. The top row shows an image taken before manganese administration, the middle row after a cumulative dose of 5 mg/kg, and the bottom row after a cumulative dose of 50 mg/kg. Images show a 2 mm-thick slice taken at the level of the globus pallidus (left) and at the level of the substantia nigra (right). Note the darkening of globus pallidus (G), pituitary gland (PI), and substantia nigra (S) after a low cumulative dose of manganese, and of the rest of the basal ganglia, including caudate (C) and putamen (P) after more extensive exposure.

longer values. These data (also not shown) prompted a closer examination of response durations and interresponse times. Figure 5 shows IRTs and response durations for CM904 from sessions on Days 1, 104, and 173: that is, a session prior to exposure, several weeks after manganese treatment began and when the first elevation in the number of incomplete responses occurred, and several months after the first dose, at a time when incomplete responses had stabilized at a low level before increasing again (see Fig. 2). These graphs show the average IRT time (top) and average duration of complete responses (bottom) plotted against ordinal position in the ratio. IRTs designate times between complete responses and may include intervening incomplete responses. Only the first 19 durations of each ratio were included in

this analysis because the monkeys sometimes perched on an extended bar to collect their juice. During control conditions, IRTs during the ratio component averaged about 0.1 set at each position and showed little variability through the session. Visual monitoring of the monkeys confirmed that after the first response the remainder of the ratio was emitted as an unbroken burst of vigorous, short duration responses. Shortly after manganese exposure IRTs lengthened and became more variable both within the ratio and through the session. This period corresponded to a period of a very high number of incomplete responses, so that the variability probably reflects the contribution of many incomplete responses. Later, the IRTs stabilized to a mean value of 0.4 set, about 4 times that seen

92

NEWLAND

AND WEISS

18

CM904 _

_

_

_

_

_

_

_

_

.

.

_

12

6

_..._..._..._..._...~.... _ _ _ _ _ _

2

25

CM642

CM646 M

BB

CM642

9

PA

A

BB

A

CM646

Day Number FIG. 3. Run time (the time to complete the fixed ratio requirement, left column) and responses/fixed interval, averaged over a session (right column) for the three monkeys. Captions and labels as in Fig. 2. The baseline for CM846 for run time did not stabilize until about Day 50.

MANGANESE

_

Last Control Session

0

15

AND EPPORTPUL CM

904

After

30

45

o

Session

15 Time

93

RESPONDING

30

Mn

45

60

75

DO

(min)

FIG. 4. Cumulative records from CM904 before and after manganese administration. The abscissa is session time and the ordinate is cumulative responses. Two lines are shown for each session; the upper line (stopping at 475 responses) comes from the fixed ratio component and the lower line from the ftxed interval component. The pen advanced for each completed response, except the response that terminated in the reinforcement cycle. A diagonal slash indicates each incomplete response. Only a few responses appear during the fixed interval component because sometimes only about one response (the reinforced one) occurred in an interval.

during control sessions but with the variability through the session as small as that seen during control sessions. A slight increase in IRT began to appear at about the 15th IRT. The first response in the ratio tended to be long and variable in duration while subsequent durations were shorter and much less variable. The mean duration for responses 3 to 16 varied only slightly around a mean of 0.3 set but this increased to 0.5 set for ordinal positions 17 to 19, an effect not detectable by visual observation. This increased duration at the end of the ratio does not represent incomplete responses; cumulative records revealed that most of these occurred in the beginning or middle of the ratio, and incomplete responses were excluded from the calculation of duration. Two weeks after the end of the first series of manganese doses (and about 10 weeks after the beginning of that series), response durations became much longer and more variable and showed no tendency to increase toward the end of the ratio. Later, durations shortened, variability decreased to about the level seen during control sessions, and the pro-

gressive increase in duration reappeared but with an earlier onset. The second peak of incomplete responses shown in Fig. 2 coincided with the appearance of action tremor in the upper limbs. Although the data are not shown, the pattern of response durations and IRTs during these sessions followed a pattern like that seen in Fig. 5. Increased IRTs and durations, coupled with greater variability, coincided with the appearance of action tremor. Later, the pattern seen in the curve identified as “late” in Fig. 5 reappeared. DISCUSSION The present report describes what is essentially a series of three experiments, each conducted on a single subject (Sidman, 1960). Two control conditions were present for each subject. First, each subject served as its own control because data were compared with a preexposure baseline. Second, unexposed or previously exposed subjects served as a con-

94

NEWLAND F2.2

set Interresponse Times

1.2. 1.0. 0 . 8. 0 . 6. 0 . 4-

1::: -2 0

20

?+3.95

w

set Response Durations

0.24

0

2

4

6 8 10 12 14 16 18 20 Position in Ratio

FIG. 5. Interresponse times (IRT, top) and response durations (bottom) from the fixed ratio component of sessionson Days 5 (control), 104 (early), and 173 (late) for CM904. The abscissa shows the location of the response or IRT in the ratio. Each point is the average duration or IRT for that position in the ratio; postreinforcer pauses are not included. For example the response duration over point 1 is the average duration of the first response for that session. The IRT over point 1 is the average time between responses 1 and 2 for that session. Only 19 durations are shown; the reinforced duration is not shown because sometimes the subject perched on the bar to collect the reinforcer, thereby producing a very long duration. All IRTs are shown; 20 responses contain 19 IRTs. The error bars show the standard error of the mean.

current control to ensure that irrelevant seasonal or equipment-related effects were not misinterpreted as treatment effects. The principle conclusion drawn is consistent across all three subjects: cumulative doses of manganese that produced no other lasting signs of toxicity disrupted the execution of a response that required considerable exertion. In two of the three monkeys this disruption was large and irreversible, and in a third it appeared but did not persist. Although exposure was intravenous, the effect did not appear immediately but required at least several days to develop, and even continued to increase in magnitude over the course of weeks for two of the monkeys. In these two monkeys, the magnitude of disruption tended to decline over many weeks, but baseline performance was never recaptured.

AND WEISS

Two subjects were investigated with magnetic resonance imaging to localize manganese in the brain. These data are described above and in Newland et al. (1989). In CM846, the earliest behavioral effects emerged at about the time elevated manganese levels were detected in the globus pallidus and substantia nigra, but not in other regions of the basal ganglia. This effect was reversed in subsequent sessions. During the later dosing sequence a more sustained elevation in incomplete responses appeared and corresponded to persistent darkening of the globus pallidus and the rest of the basal ganglia and was accompanied by a shortened Tl in that region. In CM904, the initial, persistent rise in incomplete responses following dosing occurred when manganese was detected throughout the entire basal ganglia. The role of schedule of reinforcement. The nature of the behavioral response to manganese depended upon the schedule of reinforcement maintaining behavior. One source of this difference is the disparity in response pattern generated by the two schedules (Newland and Weiss, 1990, contains interresponse time distributions for CM842 and CM904). The fixed ratio schedule maintained a vigorous, high-rate pattern of responding with little variability, and showed pronounced effects at lower cumulative doses of manganese in all three animals. The fixed-interval schedule maintained a more relaxed, less cohesive pattern of responding: interresponse times were longer and more variable, yielding lower and more variable response rates in turn. FI performance was modified by manganese only at the highest cumulative doses, and not consistently across the three animals. The differences in sensitivity between the two reinforcement schedules probably are related to the response pattern selected by them (Newland and Weiss, 1990). They suggest that physically demanding responses are more sensitive to manganese toxicity than relatively sedentary responses. This observation may be related to complaints of fatigue and difficulty in executing demanding tasks, consistently reported as an early symptom of manganese intoxication (Schuler et al., 1957). In these miners these subtle symptoms historically have been difficult to detect and have emerged as subjective complaints. The present experiments demonstrate that such effects can be quantified and clearly observed. Similarly, reports of workers occupationally exposed to lower levels of manganese than those seen in mines have quantified subtle effects such as deficits in hand steadiness (which may reflect tremor) and a slight slowing of movement (Iregren, 1990; Roels et al., 1987), effects that are consistent with bradykinesia following pallidal lesions (Horak and Anderson, 1984). The frequency of incomplete responses displayed both short-term (days) and long-term (weeks to months) fluctuations. The source of short-term fluctuations is unknown but day-to-day fluctuations in the clinical syndrome of manganism have been reported (Edsall er al., 19 19; Schuler,

MANGANESE

AND

EFFORTFUL

1957). In a particularly striking example of long-term fluctuations, Edsall et al. ( 19 19) described one man whose signs and symptoms dissipated so completely after 1 year that he returned to work until the signs eventually reappeared. The long-term cycle, most evident in CM904, may reflect a behavioral adjustment to impairment. The nature of this adjustment is revealed in the microanalysis of behavior and is discussed below. Neurotoxicity in this experiment was most directly reflected by motor impairment rather than by nonspecific effects. Motivational variables were unaffected, since each monkey continued to respond at a high rate after manganese exposure, even when the response may have been more difficult to execute. Indeed, when reinforcement rate depended upon response rate, the contingency built into the fixed-ratio schedule, response rates remained high even in the face of high numbers of incompletely executed responses. Schedulespecific response patterns also remained intact after manganese exposure. The characteristic fixed-ratio and fixed-interval patterns of responding persisted in all monkeys throughout the experiment. Microanalysis. Microanalyses of CM904’s performance suggested that the decline in incomplete responses after the initial series of doses reflected a behavioral adjustment to impairment. During control conditions, the fixed-ratio schedule maintained short-duration responses separated by short interresponse times that were almost invariant throughout the ratio. Shortly after manganese exposure, both IRTs and response durations became erratic and highly variable. This pattern was accompanied by increases in the number of incomplete responses. A new response pattern, however, eventually emerged in fixed-ratio performance alter about 2 to 3 months. IRTs shortened and attained a consistent distribution but baseline levels were not achieved. Response durations assumed a novel pattern quite different from that prevailing before exposure; within each ratio, response durations lengthened progressively from the execution of the third response to the execution of the last response required for reinforcement. These changes were accompanied by a decline in the number of incomplete responses. Although the data are not shown, this pattern repeated itself after the appearance of action tremor after Day 175. The role ofdose. Attempts to model a precise dose-effect relationship on the basis of the present data are complicated by the lengthy time-course to the emergence and decline of performance deficits and by the apparent behavioral compensation to impairment. However, some general statements seem to be valid. Acute doses of both 5 and 10 mg/kg disrupted performance under the fixed-ratio schedule but without pronounced differences in acute effects. A coarser doseeffect relationship can be inferred from the qualitative effects. The cumulative doses that elevated rates of incomplete re-

RESPONDING

95

sponses were lower than the cumulative doses at which more visible neurological signs appeared. Action tremor first appeared in all three monkeys after a cumulative dose of 40 mg/kg, a higher cumulative dose than that which elevated incomplete responses under the fixed ratio schedule. Dystonia was not observed in these monkeys, but in other, unreported, work it appeared after cumulative doses of 200 to 500 mg/kg, levels comparable with those studied by others (Eriksson et al., 1987; Suzuki et al., 1978) and reported by them to produce similar signs. The relationship between cumulative dose and the appearance of overt signs, or even death, may depend upon the rate of dosing. Suzuki et al. (1978) administered weekly doses of manganese to Macaca mullata monkeys. The doses varied from 250 to 1000 g of Mnz04, subcutaneously, each week and the monkeys weighed 3.5 to 4.5 kg.-The cumulative dose at which signs appeared was lower (and the latency to onset was longer) at the lower dosing rates. Interestingly, the only fatality in that study was an animal receiving the lowest dosing rate. Even at the lowest dosing rate, the monkeys in that study received about 2250 mg of Mn204 (or approximately 350 mg of manganese per kilogram body weight), much higher than the doses used in the present study. Although there are many differences between the present study and that reported by Suzuki et al. (1978), one may be pertinent. The dosing rate used in the present study was much lower than that reported by Suzuki et al. and, in support of one of their general conclusions, the latency to the onset of signs (including tremor and, in CM904, death) was longer and the cumulative dose at which they appeared was lower in the present study than in that reported by Suzuki et al. The appearance of action tremor late in the present experiments facilitates comparisons with other investigations using different dosing regimens, such as that reported by Suzuki et al. The appearance of tremor at 40 mg/kg supports the notion that the relationship between cumulative dose and the appearance of overt signs may depend upon the rate of dosing. This relationship may also be responsible for the fact that CM904, who received the highest initial doses, was the most severely affected. But if so, then the relationship must be a complex one since Suzuki et al. reported no clear connection between the intensity of signs and the dosing regimen. CNS distribution of manganese. The globus pallidus and substantia nigra may be especially vulnerable to manganese exposure, even though these regions do not specifically accumulate manganese in the absence of excess exposure (Bonilla et al., 1982; Barbeau et al., 1976). The changes in the globus pallidus have been described as pallidal atrophy (Jellinger, 1986; Pentschew et al., 1963; Yamada et al., 1986) but neurochemical changes have also been noted (Bird et al., 1984; Eriksson et al., 1987). Yamada et al. (1986) noted

96

NEWLAND

that pallidal atrophy appeared in a case of manganese poisoning even in the absence of nigral damage. Nigral and pallidal accumulation of manganese is detectable at lower iv cumulative doses than other regions of the basal ganglia and the globus pallidus, particularly, retains manganese more avidly (Newland et al., 1989). Also, a larger relative reduction in the dopamine content of these regions, compared to other structures, occurs in monkeys exposed to manganese by inhalation (Bird et al., 1984) or subcutaneous administration (Eriksson et al., 1987). The increase in the number of incomplete responses occurred at a time when enough manganese had accumulated in the substantia nigra and globus pallidus to be easily detectable by magnetic resonance imaging. Such a correlation supports the view that chemical dysfunction in these regions underlies the earliest stages of manganese’s neurotoxicity. Other lines of evidence also point to these two regions. For example, excessive iron exposure as well as HallervordenSpatz disease are associated with damage to these same regions and the accompanying signs resemble those of manganism (Alberta et al., 1987; Hill and Switzer, 1984; Rutledge et al., 1987). The potential role of the globus pallidus in manganese neurotoxicity is particularly interesting because of the alleged phases seen in manganism. The early signs of manganese poisoning have repeatedly been described as emotional lability (Cotzias, 1958; Barbeau, 1984) and subtle neurological signs such as response slowing (Roels et al., 1987). Anatomical evidence shows that, in addition to linkages with other motor regions of the CNS, the globus pallidus also receives afferents from limbic structures via the striatal striosomes (Graybiel, 1990) and sends efferents to limbic structures (Nauta and Domesick, 1984; Parent, 1990) although the functionality of these afferents is still not well understood. A role for the pituitary gland in manganism is implied by the apparent affinity of this organ for manganese. In the present experiments and in Newland et al. (1989) the pituitary accumulated manganese at low cumulative doses and retained it as long as the globus pallidus. The specific contribution that the pituitary gland makes in manganism cannot be identified yet, but the intimate links between the pituitary and the hypothalamus and limbic system, its role in reproductive function, and reports that dopamine is the major neurotransmitter in the pituitary (Kannan, 1987; Mtiller and Nistico, 1989) point to many possibilities, especially in the psychological disturbances often reported in the early stages of manganism. In summary, profound disruptions of an effortful response were manifested at cumulative doses of manganese that produced no obvious, overt signs of neurological impairment. These same cumulative doses also produced detectable levels of manganese in the globus pallidus and the substantia nigra as visualized with magnetic resonance imaging. This im-

AND WEISS

pairment had a slow onset and showed both short-term and long-term fluctuations. The effects seen were restricted to the motor properties of the execution of the response; response patterns and the proclivity to respond were unalIiected. Thus, in many ways the execution of an effortful operant by nonhuman primates displays many of the characteristics of manganese exposure seen in humans. ACKNOWLEDGMENTS This research and the preparation of the manuscript was supported by National Institute of Environmental Health Sciences Grant ES10248, National Institute of Drug Abuse Grant DA06499, and a Grant-in-Aid from Auburn University.

REFERENCES Alberta, R., Rafel, E., Chinchon, I., Vadillo, J., and Navarro, A. (1987). Late onset Parkinsonian syndrome in Hallervorden-Spatz disease. J. Neural.

Neurosurg.

Psychiatry

50, 1665-1668.

Barbeau, A. (1984). Manganese and extrapyramidal disorders. Neurotoxicology 5, 13-36. Barbeau, A., moue, N.. and Cloutier. T. (1976). Role of manganese in dystonia. Adv. Neurol. 14, 339-352. Bernheimer, H., Birkmayer, W., Homykiewecz, O., Jellinger, K., and Seitelberger, F. (1973). Brain dopamine and the syndromes of Parkinson and Huntington. Clinical, morphological and neurochemical correlations. J. Neurol. Sci. 20, 415-455. Bird, E. D., Anton, A. H., and Bullock, B. (1984). The effect of manganese inhalation on basal ganglia dopamine concentrations in rhesus monkey. Neurotoxicology

5, 59-66.

Bonilla, E., Salazar, E., Joaquin, J., Villalobos, V., and Villalobos, R. ( 1982). The regional distribution of manganese in the normal human brain. Neurochem.

Res. 7,22 l-227.

Canavan, M. M., Cobb, S., and Drinker, C. K. (1934). Chronic manganese poisoning. Arch. Neural. 32, 50 l-5 13. Cawte, J. (1985). Psychiatric sequelae of manganese exposure in the adult, foetal and neonatal nervous systems.Aust. N. 2. J. Psychiatry 19, 2 ll217.

Cook, D. G., Fahn, S., and Brait, K. A. (1974). Chronic manganese intoxication. Arch. Neural. 30, 59-64. Cotzias, G. C. (1958). Manganese in health and disease. Physiol. Rev. 38, 503-532.

Cotzias, G. C., Horiuchi, K., Fuenzalida, S., and Mena, I. (1968). Chronic manganese poisoning: Clearance of tissue manganese concentrations with persistence of the neurological picture. Neurology 18, 376-382. Dastur, D. K., Manghani, D. K., and Raghavendran, K. V. (197 1). Distribution and fate of Mn54 in the monkey: Studies of different parts of the central nervous system and other organs. J. Clin. Invest. SO, 9-20. Delong, M. R., Georgopoulos, A. P., Crutcher, M. D., Mitchell, R. T., Richardson, R. T., and Alexander, G. E. (1984). Functional organization of the basal ganglia: Contributions of single-cell recording studies. In Functions of the Basal Ganglia, (D. Evered and M. O’Connor, Ed%), pp. 7477. Pitman, London. Drown, D. B., Oberg, S. G., and Sharma, R. P. (1986). Pulmonary clearance of soluble and insoluble forms of manganese. J. Toxicol. Environ. Health 17,201-212.

Edsall, D. L., Wilbur, F. P., and Drinker, C. K. (1919). The occurrence, course and prevention of chronic manganese poisoning. J. Ind. Hyg. 1, 183-193.

MANGANESE

AND EFFORTFUL

Eriksson, H., Magiste, K., Plantin, L. O., Fonnum, F., Hedstrom, K. G., Theodorsson-Norheim, E., Kristensson, K., Stalberg, E., and Heilbronn, E. (1987). Effects of manganese oxide on monkeys as revealed by a combined neurochemical, histological, and neurophysiological evaluation. Arch.

Toxicol.

61,46-52.

Evarts, E. V., and Wise, S. P. (1984). Basal ganglia outputs and motor control. In Functions ojthe Basal Ganglia (D. Evered and M. O’Connor, Eds.), pp. 83-95. Pitman, London. Graybiel, A. M. (1990). Neurotransmitters and neuromodulators in the basal ganglia. Trends NeuroSci. 13, 244-253. Hill, J. M., and Switzer, R. C. (1984). The regional distribution and cellular localization of iron in the rat brain. Neuroscience 11, 595-603. Horak, F., and Anderson, M. E. (1984). Influence of globus pallidus on arm movements in monkeys. I. Effects of kainic acid-induced lesions. J. Neurophysiol.

52, 290-304.

Iregren, A. (1990). Psychological test performance in foundry workers exposed to low levels of manganese. Neurotoxicol. Teratol. 12, 673-675. Jellinger, K. (1986). Exogenous lesions of the pallidurn. Handb. ojClin. Neural.

5, 465-49

1.

Kannan, C. R. (1987). The Pituitary Gland. Plenum, New York. Langston. J. W. (1987). MPTP: The promise of a new neurotoxin. In Movement Disorders 2 (C. D. Marsden and S. Fahn, Ed%). Butterworths, London. Marsden, C. D., and Fahn, S. (1987). Problems in Parkinson’s disease and other akinetic-rigid syndromes. In Movement Disorders 2 (C. D. Marsden and S. Fahn, Eds.), pp. 65-72. Butterworths, London. Mena, I., Fuenzalida, S., and Cotzias, G. C. (1967). The metabolism of manganese in chronic manganese poisoning. J. Nucl. Med. 8, 300-301. Mena, I., Horiuchi, K., Burke, K., and Cotzias, G. C. (1969). Chronic manganese poisoning. Neurology 19, 1OOO-1006. Muller, E. E., and Nistico, G. (1989). Brain Messengers and the Pituitary. Academic Press, New York. Nauta, W. J. H., and Domesick, V. B. (1984). Afferent and efferent relationships of the basal ganglia. In Functions ojthe Basal Ganglia (D. Evered and M. O’Connor, Eds.), pp. 3-22. Pitman, London. Newland, M. C.. Ceckler, T. L., Kordower, J. H., and Weiss, B. (1989).

97

RESPONDING

Visualizing manganese in the primate basal ganglia with magnetic resonance imaging. Exp. Neural. 106,25 l-258. Newland, M. C., Cox, C., Hamada, R., Oberdorster, G., and Weiss, B. (1987). The clearance of manganese chloride in the primate. Fundam. Appl. Toxicol. 9, 3 14-328. Newland, M. C., and Weiss, B. (1990). Drug effectson an effortful operant: Pentobarbital and amphetamine. Pharmacol. Biochem. Behav. 36,38 l387. Parent, A. (1990). Extrinsic connections of the basal ganglia. Trends Neurosci. 13,254-258.

Pentschew, A. F., Ebner, F., and Kovatch, R. M. (1963). Experimental manganese encephalopathy in monkeys: A preliminary report. J. Neuropathol. Exp. Neurol. 22,488-499. Roels, H., Lauwerys, R., Buchet, J. P., Genet, P., Jawad Sarhan, M. J., Hanotiau, I., de Fays, M., and Bernard, A. (1987). Epidemiological survey among workers exposed to manganese: Effects on lung, central nervous system, and some biological indices. Am. J. Znd. Med. 11, 307-327. Roels, H., Sarhan, M. J., Hanotiau, I., de Fays, M., Genet, P., Bernard, A., Buchet, J. P., and Lauwerys, R. (1985). Preclinical toxic effects of manganese in workers from a Mn salts and oxides producing plant. Sci. Total Environ. 42, 201-206. Rutledge, J. N., Hill, S. K., Silver, J., Defendini, R., and Fahn, S. (1987). Study of movement disorders and brain iron by MR. Am. J. Neuroradiol. 8,397-411. Schuler, P., Oyanguren, H., Maturana, V., Valenzuela, A., Cruz, E., Plaza, V., Schmidt, E., and Haddad, R. (1957). Manganese poisoning: Environmental and medical study at a Chilean mine. Ind. Med. Surg. 26, 167173. Sidman, M. (1960). Tactics ofScientific Research. Basic Books, New York. Snapper, A. G., Kadden, R. M., and Inglis, G. B. (1982). State notation of behavioral procedures. Behav. Res. Methods Instrum. 14,329-342. Suzuki, Y., Mouri, T., Suzuki, Y.. Nishiyama, K., Fujii, N., and Yano, H. (1975). Study of subacute toxicity of manganese dioxide in monkeys. Tokushima

J. Exp. Med.

22,5-IO.

Yamada, M., Ohno, S., Okayasu, R., Okeda, S., Hatakeyama, H., Watanabe, H., Ushio, K., and Tsukagoshi, H. (1986). Chronic manganese poisoning: A neuropathological study with determination of manganese distribution in the brain. Acta Neuropathol. 70, 273-278.