The bidirectional interaction between ventral tegmental rewarding and hindbrain aversive stimulation effects in the rat

BRAIN RESEARCH ELSEVIER

Brain Research 688 (1995) 15-20

Research report

The bidirectional interaction between ventral tegmental rewarding and hindbrain aversive stimulation effects in the rat Roberta Anderson, Magali Diotte, Eleftherios Miliaressis * School of Psychology, University of Ottawa, 275 Nicholas Street, KIN 6N5 Ottawa, Canada Accepted 28 March 1995

Abstract We used the curve-shift procedure in self-stimulating rats to examine the interaction of aversive and rewarding electrical stimuli in terms of duration and direction. The subjects were implanted with two moveable electrodes, one in a region supporting self-stimulation (the ventral tegmental area, VTA) and another in a region supporting escape (the nucleus reticularis gigantocellularis, Gi). The function relating self-stimulation rate to pulse frequency (RF function) was first obtained and then replicated in a condition where each VTA pulse was followed 0.2 or 2.0 ms later by a Gi pulse. The intensity of Gi pulses was set at a value previously found to elicit escape within less than 5 sec. The following observations were made: (1) rats self-stimulated consistently, despite the presence of Gi pulses, (2) the presence of Gi pulses shifted the RF function rightward (decreased the rewarding efficacy of VTA stimulation), with little effect on the maximum rate, (3) after 2 to 5 VTA-Gi self-stimulation sessions, the Gi pulses progressively lost their ability to shift the RF function, and (4) at the end of testing, escape was no longer detectable using Gi pulses alone. It was concluded that (1) the interaction between rewarding VTA and aversive Gi stimulation effects is bidirectional, thus suggesting the presence of algebraic summation; (2) the effect of Gi on VTA reward is transient whereas that of VTA on Gi aversion cumulates and eventually results in total abolition of Gi aversion. The present study represents the first account of cumulative and long-lasting suppression of aversion following brain stimulation in the rat. Keywords: Brain stimulation reward; Self-stimulation; Aversion; Analgesia; Ventral tegmental area; Nucleus reticularis gigantocellularis

1. Introduction Several authors have reported analgesia from electrical stimulation of reward brain sites [2-4,9,12,16,17,20,26,30] although the link between reward and pain suppression has been disputed [3,4,10,17,20,27,30] (see also review by Franklin [13]). In the present study we wished to investigate the interactions of aversive and rewarding electrical stimuli in terms of duration and direction. In order to accomplish this goal, the rewarding electrode was implanted in the ventral tegmental area (VTA), a self-stimulation site. The aversive electrode was implanted in the nucleus reticularis gigantocellularis (Gi), a subset of neurons subserving responses to noxious natural somatic stimuli of, for example, the cutaneous and arterial variety [5,6,11,14]. The interaction of V T A and Gi stimulation effects was investigated using the curve-shift procedure. This procedure consists of plotting the function relating self-stimulation rate to the frequency of rewarding pulses,

* Corresponding author. Fax: (l) (613) 562-5800 extension 4592. 0006-8993/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0 0 0 6 - 8 0 9 3 ( 9 5 ) 0 0 4 6 2 - 0

before and after various treatments [19]. In the present study, the rate-frequency self-stimulation function was also obtained in a condition where each VTA pulse was followed by a Gi pulse. The duration of interaction between V T A and Gi stimuli was examined by replicating the combined rate-frequency function over several weeks.

2. Materials and methods

2.1. Subjects and surgery Seven male Sprague-Dawley rats ( ~ 300 g) were stereotaxically implanted (under anesthesia with 65 m g / k g of pentobarbital) with two monopolar moveable electrodes [18] made of a plastic guide and a 0.25 mm diameter moveal~le stainless-steel wire which was insulated except for its conically shaped tip. With the skull held horizontal, the stereotaxic coordinates for the V T A electrode were 5.3 mm posterior to bregma, 0.7 mm lateral to the midline and 8.1 mm below the cranial surface. For the Gi electrode, the

16

R. Anderson et al. / Brain Research 688 (1995) 15-20

coordinates were 11.6, 0.7 and 9.1 mm, respectively. The current returned via an Amphenol plug connected to four miniature screws fixed on the calvarium and imbedded in dental cement. Throughout the testing period, the subjects were housed individually with the lights on between 07:00 and 19:00. Food was restricted to six Purina Rat Chow pellets/day.

2.2. Shaping for self-stimulation Shaping for VTA self-stimulation began one week after surgery. Each press on a lever in an operant chamber triggered a constant current generator which delivered a 0.3-s train of cathodal rectangular pulses of fixed width (0.1 ms) and variable frequency and intensity. A lock-out interval of 0.4 s was imposed between the beginning of two consecutive pulse trains. The brain site under investigation was retained only if the subject self-stimulated consistently for pulse frequency and intensity less than 20/train and 0.65 mA, respectively. If the subject failed to meet this criterion, the electrode was lowered by 0.2 mm, and testing was resumed shortly thereafter. Once consistent self-stimulation was obtained, the electrode and current intensity (from 0.4 to 0.65 mA, depending on the animal) remained fixed for all subsequent testing.

quency and a starting intensity of 0.1 mA, until it changed compartments. Immediately following escape, the stimulation was shut off for 60 s. New trials were performed with the current increased each time by 0.1 mA, until the latency to escape was lower than 5 s. We used the latency-intensity function to infer a threshold intensity for escape, defined as the pulse intensity corresponding to latency of 5 s. Latency to escape from Gi stimulation was also measured across several days in a separate group of three rats that never received any VTA stimulation.

2.5. The paired-pulse rate-frequency function In the final phase of the experiment, the rate-frequency self-stimulation function was collected again as in step one, with the exception that each VTA pulse was followed 0.2 or 2.0 ms later by a Gi pulse. The intensity for the VTA pulses was identical with that used in step one, whereas that for the Gi was set at the value defined above as the threshold intensity for escape. The paired-pulse rate-frequency function was obtained once daily over several days. Rate-frequency functions using VTA pulses only were also gathered periodically to assess the stability of this control condition. At the end of testing, escape thresholds for continual Gi-only stimulation, were measured once again.

2.3. The rate-frequency self-stimulation function 2.6. Histology In a first step, we collected the rate-frequency function, relating bar-pressing rate/min to the frequency of electrical pulses. A rate-frequency function was obtained from a series of 1-min self-stimulation trials. Two consecutive trials were separated by a 30-s period during which stimulation was not available. Between trials, the frequency was varied in ascending then descending order twice, so as to cover the entire range of self-stimulation performance. The rate-frequency function was replicated daily until it did not shift laterally for 3 consecutive days by more than 0.05 log-frequency units. These three functions were used to infer the frequency threshold, defined as the number of pulses required for half of the maximum self-stimulation rate.

After completion of testing, the animals were given a lethal dose of pentobarbital and stimulation sites, marked according to a procedure described elsewhere [25]. The brains were subsequently sliced (30 /xm thick) in a cryostat-microtome, and the brain slices containing the electrode traces were stained with thionine.

16o

~

2.4. The latency to escape-intensity function In a second step we collected the function relating latency to escape from continual Gi stimulation to pulse intensity. For a given subject, the frequency of Gi pulses was fixed at the value supporting maximum self-stimulation when VTA pulses were used. This frequency was chosen because, in a later stage, the rats were destined to receive pairs of VTA + Gi pulses and we wanted the stimulation to carry both rewarding and aversive effects. The subject was placed in one compartment of a two-compartment operant chamber. It was then administered continual Gi stimulation consisting of pulses of constant fre-

#312

140

loo

D8

o ~

80

•~

60

D45

40

~

~o 4

/2Z22 6

8

~o

~2

~

~6

~8

20

Number of pulses per train Fig. 1. Barpressing rate as a function of the frequency of VTA pulses (closed symbols) and VTA + Gi pulse-pairs (open symbols), in a representative subject. In this subject, the V T A - G i function was replicated 10 times, from day 1 (D1) to day 45 (D45) following the control (VTA) condition. For visibility, only four of these replications are shown.

17

R. Anderson et al. // Brain Research 688 (1995) 15-20

urinated scratched at chamber walls, jumped upon initial stimulation, withdrew into the far corner of the opposite compartment and tried to escape. When the VTA and Gi were stimulated concomitantly,

3. Results Electrical stimulation of the Gi was confirmed to be aversive, based on the observations that all rats defecated, 0.30 S 01 r, S

0.25 -

"~

o.15 -'

"0 O ,C It

0.10 =

#231

O. 30

0.20 -' 0.15 I

0.10 =

0.05

0.05 "

0.00

i

0 . O0

-0.05 01 0 ~1

-0.05

01 -o.lO

-0.10 -0.15

-0.15

-0.20

w w 8 12

| ! i 11 20 24 Day

e c e P O

• 28

• 32

of

1 31

• i 40 44

m 48 52

-0.20

14

0.30

0.25

0251

0.20

0.20

0.15

0.15 I~

0.10

i

0.00

S

-

12

i

-

01 0 .I

-0.15 -0.20

-

i

-

4

i

-

8

i

•

12

i

•

i -

1(; 2 0

l

-

24

l

-

i

•

28

32

i

3(;

•

i

40

•

I

•

i

-

1

-

32

i

-

i

-

|

3(; 40

-

|

44

-

48

52

-0.15 -

0

i

-

4

i

•

8

i

•

i

-

i

- i

1 2 1(; 2 0

-

24

i -

i -

28

32

i

-

36

i

-

40

1

-

44

i

•

48

52

Day of test 0 . 3 0 ,~

i

0.15

° ' 2 5 1

#295

0.20 q 0.15 1

0.10 "

m

0.05 =

0.10 " 0.05 '

0.00 -

0.00

-0.05

-0.05

'

-0.10

'

-0.10

'

-0.15

'

-0.15

'

--0.20

I

-0.20

0

4

8

12

ll;

20

2(; 20

32

3(; 40

Day of test

0.30 )

v

28

-0.10

4(; 4 0 5 2

0.20

£

-

1

-0.20

•

#292

00 .. 32 05 't" -

it

w

2(;

#284

Day of test

m o

-

-0.05

-0.10

S 0t C e 4: u

i

1(; 20

0.05 k,,

-0.05 0 ,.I

4

0.30

0.00

01

- | - u - 1

Day of test

0.05

o

I

0

test

0.10 0

#269

0 . 2 5 -'

o,2o -

44

40

52

. i . i

4

.

0

I

12

.

1

.

1

.

1

1(; 20

.

24

i .

i .

28

32

i .

i .

3(; 40

i .

44

I

49

.

52

Day of test

#312

0.25 •

,¢ O "0 m 0 4: It

£

~

0.20 " 0.15 0.10 = 0.05 = 0.00 -0.05

'

-0.

'

l 0

-o.15-' -0.20

0

4

0

12

1(;

20

24

28

32

3(;

40

44

40

52

Day of test Fig. 2. Logarithmic change in self-stimulation threshold frequency, as a function of day following the control (VTA) condition. Positive values indicate that the Gi pulses increased threshold whereas negative values indicate that the threshold was decreased.

18

R. Anderson et al. / Brain Research 688 (1995) 15-20

escape was abolished: In fact, rats self-administered trains of Gi stimulation at the same current that originally elicited escape, as long as the V T A was stimulated concomitantly. Fig. 1 shows the rate-frequency function obtained with trains of VTA pulses (filled symbols) and pairs of V T A + Gi pulses (open symbols), according to the testing day following the control (VTA) session, in a representative subject. Note that the presence of Gi pulses shifted the self-stimulation function to the right, with relatively little effect on the maximum barpressing rate. Note also that the capacity of Gi pulses to shift the function decreased progressively with time a n d / o r repetitive testing. Note finally

70 #293

U

60

O

5O

O

I

Q. m O I

0£

20-

0

-I

| 100

m

day 1

•

day 11 day 32

.i

.

200

300

Intensity

400 (in

500

600

700

800

mlcroAmps)

Fig. 4. The location of each stimulation site is depicted by a circle on the corresponding plate of Paxinos and Watson's stereotaxic atlas [23]. The number on each individual plate indicates the millimetric distance behind bregma. The left and right plates show the anterior and posterior locations for each electrode pair. The number beside each stimulation site identifies the animal. Abbreviations according to the above-mentioned stereotaxic atlas.

70 8302 @

u~ v 0 9.

U U)

50

40 ~

e

s

~yl

•

day14

30 U C

o

e .1

~y~

2o

~o~ w

-

100

w

•

200

|

-

300

Intensity

|

-

400

(in

|

.

500

i

-

600

i

•

700

| 300

m|croArnps)

70 #303

":" v~

6050-

T.

~

m

day 1

40-

O

9 ue,-

\\

3o -

\

.3,

2o.~

,ol O,

•

|

100

-

,

200

-

•

300

-

,

400

-

i

500

-

i

600

-

i

700

-

000

Fig. 3. Latency to escape from continual Gi stimulation in three rats, as a function of pulse intensity and replication.

that the last rate-frequency function culminated to the left of the control function. Fig. 2 plots the logarithmic change in frequency threshold for the seven animals, as a function of testing day. Positive and negative values indicate a shift of the r a t e frequency function to the right and left of the control (VTA-only) function, respectively. The following three phenomena are of interest. First, as noted above, the frequency threshold was initially increased under the effect of Gi pulses. The largest change ( + 0 . 3 log-frequency units) was noted in subject 269, indicating a twofold increase in frequency threshold. Second, the frequency threshold elevation dissipated with repeated testing: depending on the subject, the data returned progressively near the baseline or intercepted it after 2 to 5 repetitions, regardless of the day of testing. This observation indicates that the number of stimulation sessions, not the time elapsed, was the critical variable. Third, in the five subjects that were tested for longer periods, the threshold decreased below the baseline value. In other words, the combined V T A + Gi rate-frequency function was shifted

R. Anderson et al. / Brain Research 688 (1995) 15-20

to the left of the control VTA function. The largest change, noted in subject 283 ( - 0.14 log-frequency units) indicates a 28% decrease in frequency threshold. A one-way repeated measures ANOVA for the group of animals revealed that the main effect for the session was significant (F(5,27) = 6.64, P < 0.001). Post-hoc evaluation using Dunnett's test revealed a significant difference from baseline for the two first sessions only. Escape from Gi stimulation was measured a second time after completion of the paired-pulse condition. It was noted that 6 out of 7 subjects failed to escape from continual Gi stimulation, even though the VTA was no longer being stimulated. In fact, if the rats were stimulated in the operant box, they approached the lever and barpressed a few times. The contention that aversion dissipated due to repeated Gi stimulation, rather than to the contribution of VTA pulses, was investigated in a group of three subjects that had not received VTA stimulation. Fig. 3 shows the latency to escape from continual Gi stimulation as a function of pulse intensity and session. Ten to twelve latency-intensity functions were obtained per animal, in a maximum period of 76 days. In order to enhance visual inspection, only four of the functions were plotted. Note that repeated testing shifted the latency-intensity function downward rather than upward, indicating that the animals escaped more rapidly. We attribute this change to practice. Fig. 4 depicts the location of electrodes on plates reproduced from Paxinos and Watson's stereotaxic atlas [23]. Subjects 231 and 312 lost their electrode assembly. In these animals, only the VTA electrode of 312 could be correctly localized. The left and right plates show the anterior (VTA) and posterior (Gi) location for each electrode pair. The number on each individual plate refers to the millimetric distance behind bregma. Five rewarding electrode tips landed in the VTA and one in the posterior hypothalamus. The aversive electrodes landed in the vicinity of the Gi or in the intermediate reticular nucleus.

4. Discussion

The data support the view of a bidirectional interaction between VTA and Gi stimulation effects: activation of the Gi reduced the rewarding effect of VTA stimulation whereas, activation of the VTA abolished Gi aversion. We interpret this interaction as an indication of algebraic summation. By algebraic summation we mean that the combination of two opposing effects resulted in a single effect with the sign of the strongest component. One may attribute the progressive loss of Gi aversiveness to the passage of time or to prolonged Gi stimulation only, rather than to VTA stimulation. We measured escape thresholds for Gi-only pulses approximately every 2 days for more than two months in a group of three animals. No increase in escape threshold was noted in any of the

19

subjects with repeated testing. Rather, a slight decrease was noted, an effect which we attribute to practice. In other words, stimulation of the Gi alone and the passage of time were not sufficient to induce loss of aversion. To date, we obtained escape thresholds for Gi stimulation before and after several VTA self-stimulation sessions in one rat, and found no change in escape threshold. At least in this subject, development of the aversion-suppressive effect seems to require simultaneous activation of the two brain structures. Subjects 269, 295 and 312 (Fig. 3) who were tested with a V T A - G i pulse interval of 2.0 ms showed no striking difference in magnitude of the aversion-suppressive effect, compared to the rest of subjects that were tested with a shorter (0.2 ms) interval. A systematic investigation using both pulse presentation orders and a large range of intervals should provide information on how close the two stimuli must be for the anti-aversion effect to occur. Such data may, in turn, be useful in understanding the V T A - G i interaction at a synaptic level. Several studies using stimulation, lesion or pharmacological treatments reported that the VTA modulates peripheral pain [1,20,21,28]. On the other hand, the Gi is known to contain neurons that respond to noxious stimuli [5,6,11,14]. This set of observations support the contention that our VTA electrodes produced analgesia. What is remarkable is that this effect considerably outlasted the length of stimulation. When we tested the animals weeks after the last VTA + Gi stimulation session, escape was no longer detectable. This finding is to be contrasted with previous findings that analgesia from stimulation of the central and periventricular grey matter and dorsal raphe generally lasted for seconds or minutes [17,22]. After several sessions of VTA + Gi stimulation, the rate-frequency function culminated to the left of the control VTA function, a finding suggesting a reversal of Gi effect, from strong aversion to mild reward. Consistent with this interpretation is the finding that when the rats were stimulated in the operant chamber with Gi pulses only, they approached the lever and did a few barpresses. One explanation may be that Gi stimulation acquired rewarding properties through repetitive association with VTA reward. An alternative, more parsimonious explanation, is that the Gi electrode also activated a small population of reward neurons, whose contribution could not be observed before the aversive effect has dissipated. In agreement with this interpretation, mice were observed to select the arm of a Y-maze apparatus where they could receive stimulation at various aversive brain regions, although they would press very little, if not at all, in the traditional self-stimulation box [7,8]. A reversal effect was also noted by Bishop et al. [2] who found that their aversive electrodes in the amygdala became rewarding upon repeated stimulation, as reported by patients. It is worth noting that, at least in rats, most of the amygdaloid complex contains also brain stimulation reward elements [151.

20

R. Anderson et al. / Brain Research 688 (1995) 15-20

In summary, activation of Gi transiently reduced the rewarding efficacy of VTA stimulation whereas, the activation of VTA permanently reduced Gi aversion. Previous studies seem to support the contention that our VTA electrodes reduced pain. If this is true then, to our knowledge, the present study represents the first account of cumulative and long-lasting analgesia following brain stimulation. Our stimulation parameters are known to excite a variety of myelinated neurons [24,29]. Future studies should demonstrate if the rewarding or some other component of VTA stimulation is responsible for the aversionsuppressive effect.

Acknowledgements This work was supported by a grant from the National Science and Engineering Research Council of Canada.

References [1] Altier, N. and Stewart, J., Intra-VTA infusions of the substance P analogue DiMe:C7 and intra-accumbens infusions of amphetamine induce analgesia in the formalin test for tonic pain, Brain Res., 628 (1993) 279-285. [2] Bishop, M.P., Elder, S.T. and Heath, R.G., Attempted control of operant behavior in man with intracranial self-stimulation, In R.G. Heath (Ed.), Role of Pleasure in Behavior, Harper and Row, New York, 1964, pp. 55-81. [3] Carl K.D. and Bak, T.H., Lateral hypothalamic stimulation-produced analgesia: inferred refractory period of directly stimulated neurons and resistance to pimozide antagonism, Physiol. Behav., 41 (1987) 371-378. [4] Carr, K.D. and Coons, E.E., Rats self-administer nonrewarding brain stimulation to ameliorate aversion, Science, 215 (1982) 1516-1517. [5] Casey, K.L., Somatic stimuli, spinal pathways, and size of cutaneous fibres influencing unit activity in the medial medullary reticular formation, Exp. Neurol., 25 (1969) 35-56. [6] Casey, K.L., Somatosensory responses of bulboretieular units in awake cat: relation to escape-producing stimuli, Science, 173 (1971) 77-80. [7] Cazala, P., Bendani, T. and Zielinsli, A., Self-stimulation of an "aversive" brain structure, Brain Res., 327 (1985) 53-60. [8] Cazala, P., Self-stimulation behavior can be elicited from various "aversive" brain structures, Behav. Brain Res., 22 (1986) 163-171. [9] Cox, V. and Valenstein, E.S., Attenuation of aversive properties of peripheral shock by hypothalamic stimulation, Science, 149 (1965) 323-325. [10] Cox, V. and Valenstein, E.S., Effects of stimulation intensity on behavior elicited by hypothalamic stimulation, J. Comp. Physiol. Psychol., 69 (1969) 730-733.

[11] Cross, S.A., Pathology of pain, Mayo Clin. Proc., 69 (1994) 375383. [12] Denis, S.G., Choiniere, M. and Melzack, R., Stimulation-produced analgesia in rats: assessments by two pain tests and correlation with self-stimulation, Exp. Neurol., 68 (1980) 295-309. [13] Franklin, K.B.J., Analgesia and the neural substrate of reward, Neurosci. Biobehav. Rev., 13 (1989) 149-154. [14] Guilbaud, G., Besson, J.M., Oliveras, J.L. and Wyon-Maillard, M.C., Modifications of the firing rate of bulbar reticular units (nucleus gigantocellularis) after intra-arterial injection of bradykinin into the limbs, Brain Res., 63 (1973) 131-140. [15] Kane, F., Coulombe, D. and Miliaressis, E., Amygdaloid self-stimulation: A moveable electrode mapping study, Behav. Neurosci., 105 (1991) 926-932. [16] Lopez, R. and Cox, V.C., Analgesia for tonic pain by self-administered lateral hypothalamic stimulation, NeuroReport, 3 (1992) 311314. [17] Mayer, D.J. and Liebeskind, J.C., Pain reduction by focal electrical stimulation of the brain: an anatomical and behavioral analysis, Brain Res., 68 (1974) 73-93. [18] Miliaressis, E., A miniature moveable electrode for brain stimulation in small animals, Brain Res. Bull. 7, (1981) 715-718. [19] Miliaressis, E., Rompr6, P.-P., Laviolette, P., Philippe, L. and Coulombe, D., The curve-shift paradigm in self-stimulation, Physiol. Behav., 37 (1986) 85-91. [20] Moreau, J.-L., Cohen, E. and Lieblich, I., Ventral tegmental analgesia in two strain of rats: Effects of amphetamine, naloxone and parachlorophenylalanine, Brain Res., 300 (1983) 1-8. [21] Morgan, M.J. and Franklin, K.B.J., 6-hydroxydopamine lesions of the ventral tegmentum abolish d-amphetamine and morphine analgesia in the formalin test but not in the tail flick test, Brain Res., 519 (1990) 144-149. [22] Oliveras, J.L., Besson, J.M., Guilbaud, G. and Liebeskind, J.C., Behavioral and electrophysiologieal evidence of pain inhibition from rnidbrain stimulation in the cat, Exp. Brain Res., 20 (1974) 32-34. [23] Paxinos, G. and Watson, C., The Rat Brain in Stereotaxic Coordinates, 2rid edn., Academic Press, Toronto, 1986. [24] Ranck, J.B. Which elements are excited in electrical stimulation of the mammalian central nervous system: a review, Brain Res., 98 (1875) 417-440. [25] Rompr6, P.-P. and Miliaressis, E., Pontine and mesencephalic substrates of self-stimulation, Brain Res., 359 (1985) 246-259. [26] Rose, M.D., Pain reducing properties of rewarding electrical stimulation in the rat, J. Comp. PhysioL PsychoL, 87 (1974) 107-117. [27] Sanberg, D.F. and Segal, M., Pharmacological analysis of analgesia and self-stimulation elicited by electrical stimulation of catecholamine nuclei in the rat brain, Brain Res., 152 (1978) 529-542. [28] Szreniawski, Z. Meszaros, J. Gajewska, S. and Tarchalska, B., The effects of lesion of mesolimbic dopamine neurons on pain threshold and morphine analgesia in rats, Polish J. Pharmacol. Pharm., 29 (1977) 521-525. [29] Yeomans, J.S. Principles of Brain Stimulation, Oxford University Press, New York, 1990, 182 pp. [30] Yunger, L.M., Harvey, J.A. and Lorens, S.A., A dissociation of the analgesic and rewarding effects of brain stimulation in the rat, Physiol. Behac., 10 (1973) 909-913.