- Email: [email protected]
Carbon monoxide exposure reduces the rewarding quality of brain-stimulation reward in rats
Neurotoxicologyand Teratology,Vol. 13, pp. 175-179. ©PergamonPress plc, 1991. Printed in the U.S.A.
0892-0362/91 $3.00 + .00
Carbon Monoxide Exposure Reduces the Rewarding Quality of Brain-Stimulation Reward in Rats I J A M E S D. R O W A N A N D S T E P H E N B. F O U N T A I N 2
Department o f Psychology, Kent State University, Kent, O H 44242 R e c e i v e d 3 M a r c h 1990
ROWAN, J. D. AND S. B. FOUNTAIN. Carbon monoxide exposure reduces the rewarding quality of brain-stimulation reward in rats. NEUROTOXICOL TERATOL 13(2) 175-179, 1991.--The effect of carbon monoxide (CO) exposure on hypothalamic brainstimulation reward (BSR) was examined. Rats were trained in a procedure that daily determined their stimulus duration threshold (SDT), that is, the shortest electrical stimulus to the posterior lateral hypothalamus that would support discrete-trial leverpress responding for BSR. After a stable SDT baseline was established using a single response lever, rats were exposed to 0, 5, 10, 20, and 40 ml/kg pure CO by IP injection. The SDT was significantly elevated by the 40 ml/kg exposure (corresponding to approximately 65% carboxyhemoglobin in the blood) compared to control exposures of an equal volume, No change was observed in response rate at any dose in this 1-1ever task. No tolerance was observed when 40 ml/kg CO exposure was repeated on alternating days for 14 exposures, but a small reduction in response rate was observed in this procedure. When rats of a second group were required to alternate responses on two levers some distance apart, SDT was elevated by the highest exposure (40 ml/kg) as before. Additionally, response rate was also significantly suppressed by the highest exposure in this 2-1ever task. The results support the view that CO has a direct effect on brain reward systems assessed by the SDT task. Response rate changes due to CO exposure may be due to both direct effects on brain reward systems and other effects such as hypoxia-induced fatigue. Carbon monoxide
Brain-stimulation reward
Operant behavior
CARBON monoxide (CO) is known to reduce the rate of operant behavior. For example, inhaled CO reduces the leverpressing rate of rats and mice on a variety of schedules of reinforcement (1-3, 6, 10-12, 17-19). Significant rate-reduction effects are typically observed when CO exposure results in blood carboxyhemoglobin (HbCO) levels in excess of approximately 40% [cf. (3, 6, 10, 11, 17, 18)]. It has also been shown that intraperitoneal injections of pure CO can produce HbCO levels equivalent to those achieved with inhaled CO, and that CO exposure via this route produces behavioral rate-reducing effects comparable to those observed with inhaled CO in both rats (6) and mice (10,11). One question of interest is whether CO produces its behavioral effects by direct action on CNS systems (via tissue hypoxia or other mechanisms) or by hypoxia-induced fatigue or other peripheral effects that might modify performance levels. For example, a potentially important finding is that CO decreases food and water intake (1). This decrease in appetitive behavior, it might be argued, could indicate that CO has a direct effect on CNS motivational or " r e w a r d " systems, an effect that might also account for all or part of CO's behavioral effects on response rate. On the other hand, a decrease in appetitive behavior might be the result of some other effect, such as CO-induced nausea, dizziness, or fatigue.
Threshold procedure
The present experiments were designed to assess the impact of CO exposure on the rewarding quality of hypothalamic brainstimulation reward (BSR) in rats. Studies examining the effects of CO on BSR-supported operant performance have generally found that CO produces rate-reducing effects with the same characteristic pausing observed in operant tasks using other reinforcers (1, 2, 6). Additionally, it has been proposed that the hypothalamic BSR system can serve as a useful model system for assessing the neurotoxicity of chemicals [cf. (2,7)]. This idea would be especially true in the case where agent effects on reward quality could be dissociated from other nonspecific effects on performance. The BSR threshold procedure used in the present experiments, a variation of one described by Stein and Ray (21), allows for such a dissociation across a wide range of performance [cf. (7)]. The results, then, can potentially provide new information concerning the direct impact of CO on the brain systems subserving motivation and reward versus the effects of CO on other (as yet nonspecific) processes relevant to performance. Rats were trained in a procedure that daily determined the shortest stimulus duration of an electrical stimulus administered to the lateral hypothalamus that would support discrete-trial leverpress responding for BSR. Once an SDT baseline was established, this SDT measure was used to assess changes in the rewarding
~A portion of the results was reported at the Methods in Behavioral Toxicology/Teratology Meeting, Little Rock, AR, October, 1989, and described in the proceedings (7). 2Requests for reprints should be addressed to Stephen B. Fountain at the Department of Psychology, Kent State University, Kent, OH 44242 or [email protected]. 175
176
ROWAN AND FOUNTAIN
quality of BSR following CO exposure. CO exposures were IP injections of 0, 5, 10, 20, or 40 ml pure CO/kg body weight, corresponding to approximately 0, 25, 35, 45, or 65% HbCO in the blood (7). METHOD
Subjects The subjects were 14 male hooded rats weighing approximately 300 g at the time of implantation of electrodes. The rats were housed in individual stainless steel hanging cages and were provided free access to food and water in their home cages throughout the experiments. None of the rats failed to complete the study. Colony lights went off at 7:00 p.m. and on at 7:00 a.m. All testing was conducted during the daylight portion of the light/dark cycle. Each rat was anesthetized by intraperitoneal nembutal injection (50 mg/kg). A twisted stainless steel bipolar electrode (Plastic Products MS303/1) was then implanted unilaterally in the posterior lateral hypothalamus (coordinates with skull level: 4.5 mm posterior to the bregma, 1.5 mm lateral to the midline, and 8.5 mm below the surface of the skull). After surgery, rats received 60,000 units penicillin by IM injection. Rats were allowed at least 1 week recovery time before leverpress training.
Apparatus Four operant chambers (30 × 30 × 30 cm) were composed of clear Plexiglas walls and a floor of stainless steel rods (0.5 cm in diameter) spaced 1.25 cm apart. One wall was equipped with 2 retractable response levers located 4.0 cm from the front and rear walls and 5.0 cm above the floor. Levers required approximately 18 N of force for switch closure. Rats in the testing chamber were connected to a stimulator by way of a flexible cord (Plastic Products MS304) and a commutating device that allowed the animal free movement within the chamber. Each operant chamber was enclosed in a sound-attenuating cubicle (47 × 90 × 56 cm, inside dimensions) made of plywood. Each cubicle was equipped with a circulation fan to provide ventilation and background noise. A houselight (7.5 W) was centered on the ceiling of the cubicles. The experiment was controlled by an IBM XT-compatible computer equipped with a digital I/O interface (IBM Data Acquisition and Control Adapter).
Procedure Pretraining. Following recovery from surgery, each rat was placed in an operant chamber and connected to the stimulator by way of a flexible cord and commutating device. The rats were shaped to leverpress for BSR; a 60-Hz sinusoidal pulse train of 250-ms duration (30-100 p.A) was delivered for shaping and for each leverpress. Prior to the present experiments, all rats participated in one other experiment that involved pressing sequences of levers in an 8-1ever array for 250-ms pulses of BSR. Rats were not exposed to drugs (except those associated with surgical implantation of electrodes) or toxicants prior to the present experiments. Dose-response determination: 1-Lever task. On the two days beginning the present experiment, 6 rats were shaped to leverpress for 250-ms pulses of BSR in a discrete-trial procedure. Leverpress training sessions lasted 60 min. The left lever was inserted into the chamber at the beginning of each trial and was retracted following each response. The intertrial interval was 1 s. The houselight remained illuminated throughout these training sessions.
Following the foregoing initial training, rats were transferred to the procedure to be used for the remainder of the experiment. The discrete-trial procedure employed before was used with the exception that the stimulus duration of the BSR pulse train was decremented by 2 ms each trial until the nonresponse criterion of 30 s was met. When the rat failed to respond for 30 s, the lever was retracted from the chamber and the houselight was extinguished for 5 s. Following this 5-s "timeout," the houselight was illuminated, the lever was inserted into the chamber, the stimulus duration was reset to 250 ms, and a 250-ms priming pulse of BSR was administered. Thus rats received decreasing stimulus durations on each trial in a discrete-trial procedure until no response was produced, then the stimulus duration was reset and the procedure was started once again. The stimulus duration supporting the last response in a descending series was taken to be the SDT, and this value was recorded for each descending series of BSR values. All SDTs less than 250 ms were averaged each day to obtain the daily mean SDT. Training under these conditions was conducted for 1 h each day, 5 days each week. Once stable baseline SDTs were observed, CO exposures were implemented on Tuesdays and Thursdays of each week so that performance on days both before and after exposure could be assessed. CO exposure was by IP injection of pure carbon monoxide [cf. (6)]. On CO exposure days, rats were injected 30 min prior to behavioral testing, then replaced in their home cages to await testing. Rats were twice exposed to 0, 5, 10, 20, and 40 ml CO/kg body weight. Each rat was exposed to only one dose each day and the order of exposure to the doses across days was counterbalanced. Rats that received the control exposure of 0 ml CO/kg body weight were injected with a volume of air equal to 40 ml/kg. Effects of repeated exposure. To characterize the effects of repeated exposure to CO, the 6 rats of the foregoing study were tested again beginning the following day using the same procedure except that a) they were tested 7 days a week (1 h each day, as before) and b) rats were injected on alternating days. Two days following the end of the dose-response determination phase of the experiment, all rats received a control injection of 40 ml air/kg body weight. Two days later, and on altemating days thereafter, rats received injections of 40 ml CO/kg body weight. On noninjection days, rats were tested but received no injection. Rats received a total of 14 CO exposures, all of the 40 ml/kg dose. Dose-response determination: 2-Lever task. On the two days beginning the present experiment, 8 other rats were shaped to leverpress for 250-ms pulses of BSR in a discrete-trial procedure. Leverpress training sessions lasted 30 min. On the first day, rats were shaped to leverpress the left lever. On the second day, they were shaped to leverpress the right lever. All other shaping procedures were the same as in the preceding experiment. Following this initial training, rats were transferred to the SDT procedure for the remainder of the experiment. The SDT procedure employed during the initial dose-response determination above was used with the exception that levers were inserted in alternating left and right positions, thus requiring the rat to leverpress on alternating levers in a forced-choice procedure. All other procedures, including CO exposure protocols, were identical with the exception that rats received only one exposure to each CO dose. Thus, rats were exposed only once to 0, 5, 10, 20, and 40 ml CO/kg body weight. Each rat was exposed to only one dose each day and, as before, the order of exposure to the doses across days was counterbalanced.
Statistical Analysis Analysis of variance (ANOVA) was employed for statistical
CO EFFECTS ON BRAIN-STIMULATION REWARD
evaluation of SDT and response rate data. Daily response rates were calculated as total responses (i.e., total trials) divided by total time in the session corrected by excluding the time elapsed during trials meeting the nouresponse criterion (to account for varying numbers of SDTs and, thus, varying numbers of nonresponse criteria met in different sessions). Duncan's multiple range test was used for post hoc comparison of treatment means generated using the dose-response procedures. Only p-values less than 0.05 were considered statistically significant.
177
160 150
g
14o
E
130
Q
o~
Dose-Response Determination: 1-Lever Task
Effects of Repeated Exposure As in the dose-response determination using the 1-1ever task, exposure to 40 ml/kg CO produced a statistically significant increase in SDT. However, no effect of repeated exposure was obtained. The top panel of Fig. 1 depicts rats' daily mean SDTs following either 40 ml/kg CO injection or no injection (on alternating days) over the course of 28 days (filled symbols). The ANOVA performed on these data indicated a significant main effect of injection, F(1,63)= 55.74, p<0.001, but no other significant effects (ps>0.05). In contrast to the results of the 1-1ever task reported above, however, repeated CO exposure did significantly reduce rats' response rate, though the effect looks to be somewhat variable. The bottom panel of Fig. 1 shows rats' daily mean response rate corresponding to the data reported in the top panel of Fig. 1. Here the main effect of injection was significant, F(1,65)= 17.66, p<0.001. No other effects were found to be statistically significant (ps>0.05).
/
\
/
•\
e-'e"e
120
1oo
40 ml/kg CO
• •/ ix
110 0
RESULTS
Increasing doses of CO produced a graded increase in SDT, that is, a graded reduction in the rewarding quality of BSR, when rats were tested in the original dose-response procedure. Rats' daily mean SDTs were 111.2, 107.8, 117.1, 122.7, and 138.1 ms following exposure to 0, 5, 10, 20, and 40 ml CO/kg body weight, respectively. The ANOVA performed on these data indicated a significant main effect of dose, F(4,20)=2.95, p = 0 . 0 4 5 . Post hoc comparisons showed that SDT was significantly elevated when rats were exposed to 40 ml CO/kg body weight compared to when they were exposed to 0, 5, 10, and 20 ml CO/kg body weight (p<0.05). Although CO reduced the rewarding quality of BSR (reflected in elevated SDT), CO did not affect rats' response rate in this procedure. Rats' daily mean response rates were 28.4, 24.6, 27.3, 27.7, and 26.5 responses per minute following exposure to 0, 5, 10, 20, and 40 ml CO/kg body weight, respectively. The main effect of dose was not significant, F(4,20) = 1.14, p =0.367. The foregoing SDT and response rate data were previously reported in . Fig. 6 of Fountain et al. (7).
/ ~
,
.O"0"O"O"O
•e.
"~
'0"
0" i
i
,o..e "O"
No Injection i
,
=
NC1
i
i
,
5
,
,
i
i
i
,
,
10
,
,
i
I
i
15
I
20
I
I
I
I
I
I
I
I
25
DAYS 35
.~
No Injection
30
c~ZXO"-O- "0" "Oz=....t;e'~O" - o - : e ~ e - - o - -o- -e&
25
v
=,=\
/
_.
=
_\ /=
• tu -~ tu tn z n° ta tr
•
=_•_=..
40 rnl/kg CO
20
15 10 5 0
NC1
5
10
15
20
25
DAYS FIG. 1. SDT (top panel) and response rate (bottom panel) produced following IP injections of 40 ml CO/kg body weight (filled triangles) versus no injection (filled circles) on alternating days in the l-lever task. Performance following control injections of 40 ml air/kg body weight (open triangles) and the day before control injections (open circles) are shown for comparison.
however, CO did reduce rats' response rate in this procedure. Rats' daily mean response rates were 26.9, 26.9, 27.5, 23.6, and 20.7 responses per minute following exposure to 0, 5, 10, 20, and 40 ml CO/kg body weight, respectively. In this instance, the main effect of dose was significant, F(4,28)= 3.10, p = 0 . 0 3 1 . Post hoc comparisons showed that response rate was significantly reduced when rats were exposed to 40 ml CO/kg body weight compared to when they were exposed to 0, 5, and 10 ml CO/kg body weight (p<0.05). The foregoing SDT and response rate data for the 2-1ever task were previously reported in Fig. 7 of Fountain et al. (7). DISCUSSION
Dose-Response Determination: 2-Lever Task As in the 1-1ever task, increasing doses of CO produced a graded increase in SDT. Rats' daily mean SDTs were 110.2, 110.5, 119.5, 118.9, and 147.7 ms following exposure to 0, 5, 10, 20, and 40 ml CO/kg body weight, respectively. The ANOVA performed on these data indicated a significant main effect of dose, F(4,28)=2.91, p = 0 . 0 3 9 . Post hoc comparisons showed that SDT was significantly elevated when rats were exposed to 40 ml CO/kg body weight compared to when they were exposed to 0, 5, 10, and 20 ml CO/kg body weight (p<0.05). In contrast to the results of the 1-1ever task reported above,
The results of both experiments support the view that CO directly affects the reward systems of the brain. In each experiment 40 ml/kg CO exposure, corresponding to approximately 65% HbCO in the blood throughout the hour of testing (6), resulted in elevation of SDT which can be interpreted as a reduction in the rewarding quality of BSR. SDT elevation following CO exposure cannot be explained by CO-induced fatigue, ataxia, or other nonspecific effects because a) the SDT is a relatively rate-independent measure [cf. (7)] and b) elevation of SDT was shown to occur when no rate reduction was observed (in the dose-response determination using a 1-1ever task). This is not to say that CO
178
ROWAN AND FOUNTAIN
does not have effects on other central or peripheral systems (in fact, one would certainly predict such effects), but only that under appropriate conditions effects on brain reward systems (viz., BSR) can be distinguished from other central and peripheral effects. It is interesting that the characteristic rate-reducing effects of CO were not obtained in the initial dose-response determination using a 1-1ever task. A possible explanation for this result is that response rate is influenced more by peripheral hypoxia than by CNS effects of CO, and that the 1-1ever task was not demanding enough to bring about sufficiently serious tissue hypoxia to cause muscle fatigue (and perhaps more serious CNS hypoxia). The results of the dose-response determination using the more demanding 2-1ever task are consistent with this idea; significant response rate suppression was observed in the latter task following 40 ml/ kg CO exposure. It should be noted that the results of the repeated exposure procedure suggest that CO did have a subde effect on response rate that could be detected in the 1-1ever task only by increasing statistical power, in this case via a repeated measures design. Two additional comparisons can be drawn between the present results and those obtained in earlier studies (3, 6, 10, 11, 17, 18). First, in the present experiments, only the highest dose of CO (viz., corresponding to approximately 65% HbCO in the blood) suppressed response rate, and that effect was observed only under conditions of a more demanding task (viz., the 2-1ever task). Several studies, including an earlier study of our own, have found that response-rate suppression and other behavioral effects typically occur when HbCO levels exceed approximately 40% (3, 6, 10, 11, 17, 18). In our earlier study (6), but not in the present one, CO exposure producing approximately 45% HbCO was sufficient to suppress response rate. As some have noted, discretetrial procedures like that employed in the present study are expected to be less sensitive to changes in response rate than free operant procedures. This idea alone, however, is not sufficient to account for our failure to detect response-rate suppression at CO doses producing 45% HbCO levels in the present experiments; both the present experiments and our earlier study (6) employed discretetrial procedures, yet the earlier study found response-rate suppression effects comparable to those reported with free operant procedures (3, 10, 11, 17, 18) whereas the present experiments did not. Currently we have no explanation for this difference be-
tween the SDT procedure and our earlier discrete-trial patterntracking procedure (6) or free operant procedures (3, 10, 11, 17, 18). Second, in the present study no signs of tolerance were observed during the course of repeated exposure to the highest dose of CO (40 ml/kg CO). This result is in agreement with the results of repeated exposure of this IP dose in our earlier behavioral study (6). In conclusion, the results of these experiments demonstrate that CO can have a direct effect on brain reward systems underlying BSR. This result is consistent with the view that the CNS should be considered the primary target of CO toxicity (12) and that the hypothalamus, as well as other brain structures, may be particularly sensitive to CO toxicity [cf. (4)]. The results are also consistent with the conclusion that blood carboxyhemoglobin levels typically exceed 40% before behavioral effects are observed [cf. (3, 6, 10, 11, 17, 18)]. Our past (6) and present results, along with those of Knisely et al. (10,11) with mice, further validate IP injection as an appropriate route of administration for studying CO effects. Although the present results do support the view that CO has a direct effect on brain reward systems, the mechanism of action is as yet undetermined. Generally it has been assumed that the toxicity of CO exposure is due to the reduced oxygen carrying capacity of the blood which results in hypoxia in target tissues [e.g., (13,20)]. However, some investigators question the hypothesis that the primary action of CO is by its combination with hemoglobin (8, 9, 13), suggesting instead that CO may interfere directly with cellular physiology by combining with nonheme biomolecules such as myoglobin, cytochrome a3, or calcium [cf. (5,16)]. This question of mechanism of action remains unresolved and demands further attention. ACKNOWLEDGEMENTS This research was reported in another form by J. D. Rowan in partial fulfillment of the requirements for the Master of Arts degree, Department of Psychology, Kent State University. This work was supported in part by the U.S. National Institutes of Mental Health (MH43576). Additional support was provided by NIH Biomedical Research Support grant 2-S07RR07208 to Kent State University. We thank Pat Gallagher and Margie Sanders for assistance with surgery, Karen Esling and Tracey Housel for assistance in data collection, and Nick LeCursi for technical assistance.
REFERENCES 1. Annau, Z. The comparative effects of hypoxic and carbon monoxide hypoxia on behavior. In: Weiss, B.; Laties, V. G., eds. Behavioral toxicology. New York: Plenum Press; 1975:105-127. 2. Annau, Z. Electrical self-stimulation of the brain: A model for the behavioral evaluation of toxic agents. Environ. Health Perspect. 26: 59-67; 1978. 3. Ator, N. A.; Merigan, W. H.; Mclntire, R. W. The effects of brief exposures to carbon monoxide on temporally differentiated responding. Environ. Res. 12:81-91; 1976. 4. Bleecker, M. L. Parkinsonism: A clinical marker of exposure to neurotoxins. Neurotoxicol. Teratol. 10:475--478; 1988. 5. Coburn, R. F. Mechanisms of carbon monoxide toxicity. Prev. Med. 8:310-322; 1979. 6. Fountain, S. B.; Raffaele, K. C.; Annau, Z. Behavioral consequences of intraperitoneal carbon monoxide administration in rats. Toxicol. Appl. Pharmacol. 83:546-555; 1986. 7. Fountain, S. B.; Rowan, J. D.; Ting, Y.-L. T. Threshold procedures for assessing the impact of agents on brain reward systems. Neurotoxicol. Teratol. 12:469-475; 1990. 8. Goldbaum, L. R.; Drellano, T.; Dergel, E. Studies on the relation between carboxyhemoglobin concentration and toxicity. Aviat. Space Environ. Med. 48:969-970; 1977. 9. Goldbaum, L. R.; Ramirez, R. G.; Absalon, K. B. What is the
10. 11.
12. 13. 14. 15. 16. 17.
mechanism of carbon monoxide toxicity? Aviat. Space Environ. Med. 46:1289-1291; 1975. Knisely, J. S.; Rees, D. C.; Balster, R. L. Effects of carbon monoxide in combination with behaviorally active drugs on fixed-ratio performance in the mouse. Neurotoxicol. Teratol. 11:447--452; 1989. Knisely, J. S.; Rees, D. C.; Salay, J. M.; Balster, R. L.; Breen, T. J. Effects of intraperitoneal carbon monoxide on fixed-ratio and screen-test performance in the mouse. Neurotoxicol. Teratol. 9:221225; 1987. Laties, V. G.; Merigan, W. H. Behavioral effects of carbon monoxide on animals and men. Annu. Rev. Pharmacol. Toxicol. 19:357392; 1979. Lilienthal, J. L., Jr. Carbon monoxide. Pharmacol. Rev. 2:324-354; 1950. Lin, H.; McGrath, J. J. Carbon monoxide effects on calcium levels in vascular smooth muscle. Life Sci. 43:1813-1816; 1988. Medgan, D. E.; McIntire, R. W. Effects of carbon monoxide on responding under a progressive ratio schedule in rats. Physiol. Behav. 16:407-412; 1976. Piantadosi, C. A.; Sylvia, A. L.; Jobsis-Vandervliet, F. F. Differences in brain cytochrome responses to carbon monoxide and cyanide in vivo. J. Appl. Physiol. 62:1277-1284; 1987. Schrot, J.; Thomas, J. R. Multiple schedule performance changes
CO EFFECTS ON BRAIN-STIMULATION REWARD
during carbon monoxide exposure. Neurobehav. Toxicol. Teratol. 8: 225-230; 1986. 18. Schrot, J.; Thomas, J. R.; Robertson, R. F. TempOral changes in repeated acquisition behavior after carbon monoxide exposure. Neurobehav. Toxicol. Teratol. 6:23-28; 1984. 19. Smith, M. D.; Merigan, W. H.; Mclntire, R. W. Effects of carbon monoxide on fixed-consecutive-number performance in rats. Pharma-
179
col. Biochem. Behav. 5:257-262; 1976. 20. Smith, R. P. Toxic responses of the blood. In: Doull, J.; Klaassen, C. D.; Amdour, M. D., eds. Casarett and Doull's toxicology. New York: Macmillan; 1980:311-331. 21. Stein, L.; Ray, O. S. Brain stimulation reward 'thresholds' self-determined in rat. Psychopharmacologia 1:751-756; 1960.
0892-0362/91 $3.00 + .00
Carbon Monoxide Exposure Reduces the Rewarding Quality of Brain-Stimulation Reward in Rats I J A M E S D. R O W A N A N D S T E P H E N B. F O U N T A I N 2
Department o f Psychology, Kent State University, Kent, O H 44242 R e c e i v e d 3 M a r c h 1990
ROWAN, J. D. AND S. B. FOUNTAIN. Carbon monoxide exposure reduces the rewarding quality of brain-stimulation reward in rats. NEUROTOXICOL TERATOL 13(2) 175-179, 1991.--The effect of carbon monoxide (CO) exposure on hypothalamic brainstimulation reward (BSR) was examined. Rats were trained in a procedure that daily determined their stimulus duration threshold (SDT), that is, the shortest electrical stimulus to the posterior lateral hypothalamus that would support discrete-trial leverpress responding for BSR. After a stable SDT baseline was established using a single response lever, rats were exposed to 0, 5, 10, 20, and 40 ml/kg pure CO by IP injection. The SDT was significantly elevated by the 40 ml/kg exposure (corresponding to approximately 65% carboxyhemoglobin in the blood) compared to control exposures of an equal volume, No change was observed in response rate at any dose in this 1-1ever task. No tolerance was observed when 40 ml/kg CO exposure was repeated on alternating days for 14 exposures, but a small reduction in response rate was observed in this procedure. When rats of a second group were required to alternate responses on two levers some distance apart, SDT was elevated by the highest exposure (40 ml/kg) as before. Additionally, response rate was also significantly suppressed by the highest exposure in this 2-1ever task. The results support the view that CO has a direct effect on brain reward systems assessed by the SDT task. Response rate changes due to CO exposure may be due to both direct effects on brain reward systems and other effects such as hypoxia-induced fatigue. Carbon monoxide
Brain-stimulation reward
Operant behavior
CARBON monoxide (CO) is known to reduce the rate of operant behavior. For example, inhaled CO reduces the leverpressing rate of rats and mice on a variety of schedules of reinforcement (1-3, 6, 10-12, 17-19). Significant rate-reduction effects are typically observed when CO exposure results in blood carboxyhemoglobin (HbCO) levels in excess of approximately 40% [cf. (3, 6, 10, 11, 17, 18)]. It has also been shown that intraperitoneal injections of pure CO can produce HbCO levels equivalent to those achieved with inhaled CO, and that CO exposure via this route produces behavioral rate-reducing effects comparable to those observed with inhaled CO in both rats (6) and mice (10,11). One question of interest is whether CO produces its behavioral effects by direct action on CNS systems (via tissue hypoxia or other mechanisms) or by hypoxia-induced fatigue or other peripheral effects that might modify performance levels. For example, a potentially important finding is that CO decreases food and water intake (1). This decrease in appetitive behavior, it might be argued, could indicate that CO has a direct effect on CNS motivational or " r e w a r d " systems, an effect that might also account for all or part of CO's behavioral effects on response rate. On the other hand, a decrease in appetitive behavior might be the result of some other effect, such as CO-induced nausea, dizziness, or fatigue.
Threshold procedure
The present experiments were designed to assess the impact of CO exposure on the rewarding quality of hypothalamic brainstimulation reward (BSR) in rats. Studies examining the effects of CO on BSR-supported operant performance have generally found that CO produces rate-reducing effects with the same characteristic pausing observed in operant tasks using other reinforcers (1, 2, 6). Additionally, it has been proposed that the hypothalamic BSR system can serve as a useful model system for assessing the neurotoxicity of chemicals [cf. (2,7)]. This idea would be especially true in the case where agent effects on reward quality could be dissociated from other nonspecific effects on performance. The BSR threshold procedure used in the present experiments, a variation of one described by Stein and Ray (21), allows for such a dissociation across a wide range of performance [cf. (7)]. The results, then, can potentially provide new information concerning the direct impact of CO on the brain systems subserving motivation and reward versus the effects of CO on other (as yet nonspecific) processes relevant to performance. Rats were trained in a procedure that daily determined the shortest stimulus duration of an electrical stimulus administered to the lateral hypothalamus that would support discrete-trial leverpress responding for BSR. Once an SDT baseline was established, this SDT measure was used to assess changes in the rewarding
~A portion of the results was reported at the Methods in Behavioral Toxicology/Teratology Meeting, Little Rock, AR, October, 1989, and described in the proceedings (7). 2Requests for reprints should be addressed to Stephen B. Fountain at the Department of Psychology, Kent State University, Kent, OH 44242 or [email protected]. 175
176
ROWAN AND FOUNTAIN
quality of BSR following CO exposure. CO exposures were IP injections of 0, 5, 10, 20, or 40 ml pure CO/kg body weight, corresponding to approximately 0, 25, 35, 45, or 65% HbCO in the blood (7). METHOD
Subjects The subjects were 14 male hooded rats weighing approximately 300 g at the time of implantation of electrodes. The rats were housed in individual stainless steel hanging cages and were provided free access to food and water in their home cages throughout the experiments. None of the rats failed to complete the study. Colony lights went off at 7:00 p.m. and on at 7:00 a.m. All testing was conducted during the daylight portion of the light/dark cycle. Each rat was anesthetized by intraperitoneal nembutal injection (50 mg/kg). A twisted stainless steel bipolar electrode (Plastic Products MS303/1) was then implanted unilaterally in the posterior lateral hypothalamus (coordinates with skull level: 4.5 mm posterior to the bregma, 1.5 mm lateral to the midline, and 8.5 mm below the surface of the skull). After surgery, rats received 60,000 units penicillin by IM injection. Rats were allowed at least 1 week recovery time before leverpress training.
Apparatus Four operant chambers (30 × 30 × 30 cm) were composed of clear Plexiglas walls and a floor of stainless steel rods (0.5 cm in diameter) spaced 1.25 cm apart. One wall was equipped with 2 retractable response levers located 4.0 cm from the front and rear walls and 5.0 cm above the floor. Levers required approximately 18 N of force for switch closure. Rats in the testing chamber were connected to a stimulator by way of a flexible cord (Plastic Products MS304) and a commutating device that allowed the animal free movement within the chamber. Each operant chamber was enclosed in a sound-attenuating cubicle (47 × 90 × 56 cm, inside dimensions) made of plywood. Each cubicle was equipped with a circulation fan to provide ventilation and background noise. A houselight (7.5 W) was centered on the ceiling of the cubicles. The experiment was controlled by an IBM XT-compatible computer equipped with a digital I/O interface (IBM Data Acquisition and Control Adapter).
Procedure Pretraining. Following recovery from surgery, each rat was placed in an operant chamber and connected to the stimulator by way of a flexible cord and commutating device. The rats were shaped to leverpress for BSR; a 60-Hz sinusoidal pulse train of 250-ms duration (30-100 p.A) was delivered for shaping and for each leverpress. Prior to the present experiments, all rats participated in one other experiment that involved pressing sequences of levers in an 8-1ever array for 250-ms pulses of BSR. Rats were not exposed to drugs (except those associated with surgical implantation of electrodes) or toxicants prior to the present experiments. Dose-response determination: 1-Lever task. On the two days beginning the present experiment, 6 rats were shaped to leverpress for 250-ms pulses of BSR in a discrete-trial procedure. Leverpress training sessions lasted 60 min. The left lever was inserted into the chamber at the beginning of each trial and was retracted following each response. The intertrial interval was 1 s. The houselight remained illuminated throughout these training sessions.
Following the foregoing initial training, rats were transferred to the procedure to be used for the remainder of the experiment. The discrete-trial procedure employed before was used with the exception that the stimulus duration of the BSR pulse train was decremented by 2 ms each trial until the nonresponse criterion of 30 s was met. When the rat failed to respond for 30 s, the lever was retracted from the chamber and the houselight was extinguished for 5 s. Following this 5-s "timeout," the houselight was illuminated, the lever was inserted into the chamber, the stimulus duration was reset to 250 ms, and a 250-ms priming pulse of BSR was administered. Thus rats received decreasing stimulus durations on each trial in a discrete-trial procedure until no response was produced, then the stimulus duration was reset and the procedure was started once again. The stimulus duration supporting the last response in a descending series was taken to be the SDT, and this value was recorded for each descending series of BSR values. All SDTs less than 250 ms were averaged each day to obtain the daily mean SDT. Training under these conditions was conducted for 1 h each day, 5 days each week. Once stable baseline SDTs were observed, CO exposures were implemented on Tuesdays and Thursdays of each week so that performance on days both before and after exposure could be assessed. CO exposure was by IP injection of pure carbon monoxide [cf. (6)]. On CO exposure days, rats were injected 30 min prior to behavioral testing, then replaced in their home cages to await testing. Rats were twice exposed to 0, 5, 10, 20, and 40 ml CO/kg body weight. Each rat was exposed to only one dose each day and the order of exposure to the doses across days was counterbalanced. Rats that received the control exposure of 0 ml CO/kg body weight were injected with a volume of air equal to 40 ml/kg. Effects of repeated exposure. To characterize the effects of repeated exposure to CO, the 6 rats of the foregoing study were tested again beginning the following day using the same procedure except that a) they were tested 7 days a week (1 h each day, as before) and b) rats were injected on alternating days. Two days following the end of the dose-response determination phase of the experiment, all rats received a control injection of 40 ml air/kg body weight. Two days later, and on altemating days thereafter, rats received injections of 40 ml CO/kg body weight. On noninjection days, rats were tested but received no injection. Rats received a total of 14 CO exposures, all of the 40 ml/kg dose. Dose-response determination: 2-Lever task. On the two days beginning the present experiment, 8 other rats were shaped to leverpress for 250-ms pulses of BSR in a discrete-trial procedure. Leverpress training sessions lasted 30 min. On the first day, rats were shaped to leverpress the left lever. On the second day, they were shaped to leverpress the right lever. All other shaping procedures were the same as in the preceding experiment. Following this initial training, rats were transferred to the SDT procedure for the remainder of the experiment. The SDT procedure employed during the initial dose-response determination above was used with the exception that levers were inserted in alternating left and right positions, thus requiring the rat to leverpress on alternating levers in a forced-choice procedure. All other procedures, including CO exposure protocols, were identical with the exception that rats received only one exposure to each CO dose. Thus, rats were exposed only once to 0, 5, 10, 20, and 40 ml CO/kg body weight. Each rat was exposed to only one dose each day and, as before, the order of exposure to the doses across days was counterbalanced.
Statistical Analysis Analysis of variance (ANOVA) was employed for statistical
CO EFFECTS ON BRAIN-STIMULATION REWARD
evaluation of SDT and response rate data. Daily response rates were calculated as total responses (i.e., total trials) divided by total time in the session corrected by excluding the time elapsed during trials meeting the nouresponse criterion (to account for varying numbers of SDTs and, thus, varying numbers of nonresponse criteria met in different sessions). Duncan's multiple range test was used for post hoc comparison of treatment means generated using the dose-response procedures. Only p-values less than 0.05 were considered statistically significant.
177
160 150
g
14o
E
130
Q
o~
Dose-Response Determination: 1-Lever Task
Effects of Repeated Exposure As in the dose-response determination using the 1-1ever task, exposure to 40 ml/kg CO produced a statistically significant increase in SDT. However, no effect of repeated exposure was obtained. The top panel of Fig. 1 depicts rats' daily mean SDTs following either 40 ml/kg CO injection or no injection (on alternating days) over the course of 28 days (filled symbols). The ANOVA performed on these data indicated a significant main effect of injection, F(1,63)= 55.74, p<0.001, but no other significant effects (ps>0.05). In contrast to the results of the 1-1ever task reported above, however, repeated CO exposure did significantly reduce rats' response rate, though the effect looks to be somewhat variable. The bottom panel of Fig. 1 shows rats' daily mean response rate corresponding to the data reported in the top panel of Fig. 1. Here the main effect of injection was significant, F(1,65)= 17.66, p<0.001. No other effects were found to be statistically significant (ps>0.05).
/
\
/
•\
e-'e"e
120
1oo
40 ml/kg CO
• •/ ix
110 0
RESULTS
Increasing doses of CO produced a graded increase in SDT, that is, a graded reduction in the rewarding quality of BSR, when rats were tested in the original dose-response procedure. Rats' daily mean SDTs were 111.2, 107.8, 117.1, 122.7, and 138.1 ms following exposure to 0, 5, 10, 20, and 40 ml CO/kg body weight, respectively. The ANOVA performed on these data indicated a significant main effect of dose, F(4,20)=2.95, p = 0 . 0 4 5 . Post hoc comparisons showed that SDT was significantly elevated when rats were exposed to 40 ml CO/kg body weight compared to when they were exposed to 0, 5, 10, and 20 ml CO/kg body weight (p<0.05). Although CO reduced the rewarding quality of BSR (reflected in elevated SDT), CO did not affect rats' response rate in this procedure. Rats' daily mean response rates were 28.4, 24.6, 27.3, 27.7, and 26.5 responses per minute following exposure to 0, 5, 10, 20, and 40 ml CO/kg body weight, respectively. The main effect of dose was not significant, F(4,20) = 1.14, p =0.367. The foregoing SDT and response rate data were previously reported in . Fig. 6 of Fountain et al. (7).
/ ~
,
.O"0"O"O"O
•e.
"~
'0"
0" i
i
,o..e "O"
No Injection i
,
=
NC1
i
i
,
5
,
,
i
i
i
,
,
10
,
,
i
I
i
15
I
20
I
I
I
I
I
I
I
I
25
DAYS 35
.~
No Injection
30
c~ZXO"-O- "0" "Oz=....t;e'~O" - o - : e ~ e - - o - -o- -e&
25
v
=,=\
/
_.
=
_\ /=
• tu -~ tu tn z n° ta tr
•
=_•_=..
40 rnl/kg CO
20
15 10 5 0
NC1
5
10
15
20
25
DAYS FIG. 1. SDT (top panel) and response rate (bottom panel) produced following IP injections of 40 ml CO/kg body weight (filled triangles) versus no injection (filled circles) on alternating days in the l-lever task. Performance following control injections of 40 ml air/kg body weight (open triangles) and the day before control injections (open circles) are shown for comparison.
however, CO did reduce rats' response rate in this procedure. Rats' daily mean response rates were 26.9, 26.9, 27.5, 23.6, and 20.7 responses per minute following exposure to 0, 5, 10, 20, and 40 ml CO/kg body weight, respectively. In this instance, the main effect of dose was significant, F(4,28)= 3.10, p = 0 . 0 3 1 . Post hoc comparisons showed that response rate was significantly reduced when rats were exposed to 40 ml CO/kg body weight compared to when they were exposed to 0, 5, and 10 ml CO/kg body weight (p<0.05). The foregoing SDT and response rate data for the 2-1ever task were previously reported in Fig. 7 of Fountain et al. (7). DISCUSSION
Dose-Response Determination: 2-Lever Task As in the 1-1ever task, increasing doses of CO produced a graded increase in SDT. Rats' daily mean SDTs were 110.2, 110.5, 119.5, 118.9, and 147.7 ms following exposure to 0, 5, 10, 20, and 40 ml CO/kg body weight, respectively. The ANOVA performed on these data indicated a significant main effect of dose, F(4,28)=2.91, p = 0 . 0 3 9 . Post hoc comparisons showed that SDT was significantly elevated when rats were exposed to 40 ml CO/kg body weight compared to when they were exposed to 0, 5, 10, and 20 ml CO/kg body weight (p<0.05). In contrast to the results of the 1-1ever task reported above,
The results of both experiments support the view that CO directly affects the reward systems of the brain. In each experiment 40 ml/kg CO exposure, corresponding to approximately 65% HbCO in the blood throughout the hour of testing (6), resulted in elevation of SDT which can be interpreted as a reduction in the rewarding quality of BSR. SDT elevation following CO exposure cannot be explained by CO-induced fatigue, ataxia, or other nonspecific effects because a) the SDT is a relatively rate-independent measure [cf. (7)] and b) elevation of SDT was shown to occur when no rate reduction was observed (in the dose-response determination using a 1-1ever task). This is not to say that CO
178
ROWAN AND FOUNTAIN
does not have effects on other central or peripheral systems (in fact, one would certainly predict such effects), but only that under appropriate conditions effects on brain reward systems (viz., BSR) can be distinguished from other central and peripheral effects. It is interesting that the characteristic rate-reducing effects of CO were not obtained in the initial dose-response determination using a 1-1ever task. A possible explanation for this result is that response rate is influenced more by peripheral hypoxia than by CNS effects of CO, and that the 1-1ever task was not demanding enough to bring about sufficiently serious tissue hypoxia to cause muscle fatigue (and perhaps more serious CNS hypoxia). The results of the dose-response determination using the more demanding 2-1ever task are consistent with this idea; significant response rate suppression was observed in the latter task following 40 ml/ kg CO exposure. It should be noted that the results of the repeated exposure procedure suggest that CO did have a subde effect on response rate that could be detected in the 1-1ever task only by increasing statistical power, in this case via a repeated measures design. Two additional comparisons can be drawn between the present results and those obtained in earlier studies (3, 6, 10, 11, 17, 18). First, in the present experiments, only the highest dose of CO (viz., corresponding to approximately 65% HbCO in the blood) suppressed response rate, and that effect was observed only under conditions of a more demanding task (viz., the 2-1ever task). Several studies, including an earlier study of our own, have found that response-rate suppression and other behavioral effects typically occur when HbCO levels exceed approximately 40% (3, 6, 10, 11, 17, 18). In our earlier study (6), but not in the present one, CO exposure producing approximately 45% HbCO was sufficient to suppress response rate. As some have noted, discretetrial procedures like that employed in the present study are expected to be less sensitive to changes in response rate than free operant procedures. This idea alone, however, is not sufficient to account for our failure to detect response-rate suppression at CO doses producing 45% HbCO levels in the present experiments; both the present experiments and our earlier study (6) employed discretetrial procedures, yet the earlier study found response-rate suppression effects comparable to those reported with free operant procedures (3, 10, 11, 17, 18) whereas the present experiments did not. Currently we have no explanation for this difference be-
tween the SDT procedure and our earlier discrete-trial patterntracking procedure (6) or free operant procedures (3, 10, 11, 17, 18). Second, in the present study no signs of tolerance were observed during the course of repeated exposure to the highest dose of CO (40 ml/kg CO). This result is in agreement with the results of repeated exposure of this IP dose in our earlier behavioral study (6). In conclusion, the results of these experiments demonstrate that CO can have a direct effect on brain reward systems underlying BSR. This result is consistent with the view that the CNS should be considered the primary target of CO toxicity (12) and that the hypothalamus, as well as other brain structures, may be particularly sensitive to CO toxicity [cf. (4)]. The results are also consistent with the conclusion that blood carboxyhemoglobin levels typically exceed 40% before behavioral effects are observed [cf. (3, 6, 10, 11, 17, 18)]. Our past (6) and present results, along with those of Knisely et al. (10,11) with mice, further validate IP injection as an appropriate route of administration for studying CO effects. Although the present results do support the view that CO has a direct effect on brain reward systems, the mechanism of action is as yet undetermined. Generally it has been assumed that the toxicity of CO exposure is due to the reduced oxygen carrying capacity of the blood which results in hypoxia in target tissues [e.g., (13,20)]. However, some investigators question the hypothesis that the primary action of CO is by its combination with hemoglobin (8, 9, 13), suggesting instead that CO may interfere directly with cellular physiology by combining with nonheme biomolecules such as myoglobin, cytochrome a3, or calcium [cf. (5,16)]. This question of mechanism of action remains unresolved and demands further attention. ACKNOWLEDGEMENTS This research was reported in another form by J. D. Rowan in partial fulfillment of the requirements for the Master of Arts degree, Department of Psychology, Kent State University. This work was supported in part by the U.S. National Institutes of Mental Health (MH43576). Additional support was provided by NIH Biomedical Research Support grant 2-S07RR07208 to Kent State University. We thank Pat Gallagher and Margie Sanders for assistance with surgery, Karen Esling and Tracey Housel for assistance in data collection, and Nick LeCursi for technical assistance.
REFERENCES 1. Annau, Z. The comparative effects of hypoxic and carbon monoxide hypoxia on behavior. In: Weiss, B.; Laties, V. G., eds. Behavioral toxicology. New York: Plenum Press; 1975:105-127. 2. Annau, Z. Electrical self-stimulation of the brain: A model for the behavioral evaluation of toxic agents. Environ. Health Perspect. 26: 59-67; 1978. 3. Ator, N. A.; Merigan, W. H.; Mclntire, R. W. The effects of brief exposures to carbon monoxide on temporally differentiated responding. Environ. Res. 12:81-91; 1976. 4. Bleecker, M. L. Parkinsonism: A clinical marker of exposure to neurotoxins. Neurotoxicol. Teratol. 10:475--478; 1988. 5. Coburn, R. F. Mechanisms of carbon monoxide toxicity. Prev. Med. 8:310-322; 1979. 6. Fountain, S. B.; Raffaele, K. C.; Annau, Z. Behavioral consequences of intraperitoneal carbon monoxide administration in rats. Toxicol. Appl. Pharmacol. 83:546-555; 1986. 7. Fountain, S. B.; Rowan, J. D.; Ting, Y.-L. T. Threshold procedures for assessing the impact of agents on brain reward systems. Neurotoxicol. Teratol. 12:469-475; 1990. 8. Goldbaum, L. R.; Drellano, T.; Dergel, E. Studies on the relation between carboxyhemoglobin concentration and toxicity. Aviat. Space Environ. Med. 48:969-970; 1977. 9. Goldbaum, L. R.; Ramirez, R. G.; Absalon, K. B. What is the
10. 11.
12. 13. 14. 15. 16. 17.
mechanism of carbon monoxide toxicity? Aviat. Space Environ. Med. 46:1289-1291; 1975. Knisely, J. S.; Rees, D. C.; Balster, R. L. Effects of carbon monoxide in combination with behaviorally active drugs on fixed-ratio performance in the mouse. Neurotoxicol. Teratol. 11:447--452; 1989. Knisely, J. S.; Rees, D. C.; Salay, J. M.; Balster, R. L.; Breen, T. J. Effects of intraperitoneal carbon monoxide on fixed-ratio and screen-test performance in the mouse. Neurotoxicol. Teratol. 9:221225; 1987. Laties, V. G.; Merigan, W. H. Behavioral effects of carbon monoxide on animals and men. Annu. Rev. Pharmacol. Toxicol. 19:357392; 1979. Lilienthal, J. L., Jr. Carbon monoxide. Pharmacol. Rev. 2:324-354; 1950. Lin, H.; McGrath, J. J. Carbon monoxide effects on calcium levels in vascular smooth muscle. Life Sci. 43:1813-1816; 1988. Medgan, D. E.; McIntire, R. W. Effects of carbon monoxide on responding under a progressive ratio schedule in rats. Physiol. Behav. 16:407-412; 1976. Piantadosi, C. A.; Sylvia, A. L.; Jobsis-Vandervliet, F. F. Differences in brain cytochrome responses to carbon monoxide and cyanide in vivo. J. Appl. Physiol. 62:1277-1284; 1987. Schrot, J.; Thomas, J. R. Multiple schedule performance changes
CO EFFECTS ON BRAIN-STIMULATION REWARD
during carbon monoxide exposure. Neurobehav. Toxicol. Teratol. 8: 225-230; 1986. 18. Schrot, J.; Thomas, J. R.; Robertson, R. F. TempOral changes in repeated acquisition behavior after carbon monoxide exposure. Neurobehav. Toxicol. Teratol. 6:23-28; 1984. 19. Smith, M. D.; Merigan, W. H.; Mclntire, R. W. Effects of carbon monoxide on fixed-consecutive-number performance in rats. Pharma-
179
col. Biochem. Behav. 5:257-262; 1976. 20. Smith, R. P. Toxic responses of the blood. In: Doull, J.; Klaassen, C. D.; Amdour, M. D., eds. Casarett and Doull's toxicology. New York: Macmillan; 1980:311-331. 21. Stein, L.; Ray, O. S. Brain stimulation reward 'thresholds' self-determined in rat. Psychopharmacologia 1:751-756; 1960.