Psychopharmacology( 1991) 104:220-224 003331589100107C

Psychopharmacology © Springer-Verlag 1991

Effects of ritanserin on aversive classical conditioning in humans R. H e n s m a n 1, F.S. Guimar~es 2, M. W a n g 3, and J.F.W. Deakin 1

1 UniversityDepartment of Psychiatry,Withington Hospital, Manchester M20 8RL, UK 2 Department of Pharmacology,Faculty of Medicine, USP, Ribeirao Preto, SP, BR14049, Brazil 3 Department of Psychology,Universityof Hull, HU6 7RX, UK Received August 10, 1990 / Final version December 3, 1990

Abstract. According to one formulation of the behavioural functions of 5HT, aversive conditioned stimuli mediate their behavioural and emotional effects through activation of 5HT projections from dorsal raphe nucleus to receptors of the 5HT2 family in amygdala and elsewhere. To test this theory in humans, groups of ten normal volunteers received placebo, the 5HT2/lc antagonist ritanserin (10 mg PO) and no pill. Ritanserin had no effect on skin conductance level, variability (spontaneous fluctuations) or habituation to a sequence of ten neutral tones. After a conditioning trial in which tone 11 was followed by an aversive white noise, skin conductance responses to a further ten tones were enhanced. This effect was abolished by ritanserin. The results indicate a selective involvement of 5HT2/lc receptors in modulating aversively conditioned skin conductance responses. Key words: Ritanserin - Anxiety Aversive conditioning - Healthy volunteers - Skin conductance - 5HT

Evidence concerning the role of 5HT in anxiety and depression is increasingly contradictory. One paradox is that symptoms of anxiety and depression co-exist (Goldberg et al. 1987), yet anxiety has been linked with excessive 5HT function, while deficient 5HT function has been implicated in depression. The failure of 5HT theories of behaviour to explain such contradictions may be due to their failure to consider the behavioural functions of the distinct anatomical and pharmacological 5HT sub-systems (Deakin 1983). According to the "5HT receptor imbalance theory", affective disturbance results from excessive 5HT2/lc neurotransmission relative to 5HTla/ld; 5HT systems involving the 5HT2/lc receptor family mediate symptoms of anxiety, whereas deficient 5HTla flmction mediates depressive symptoms (Deakin 1988, 1989). tt was suggested that clinically effective anxiotytics Offprint requests to: J.F.W. Deakin

work by decreasing neurotransmission involving the 5HT2 receptor family. For example, tricyclic antidepressants are clinically useful anxiolytics (Johnstone et al. 1980; Khan et al. 1986) and some have significant 5HT2 antagonist properties (Fuxe et al. 1977), and all share the ability to down-regulate 5HT2 ligand binding (Peroutka and Snyder 1980). In agreement with this theory, the anxiolytic properties of ritanserin, a 5HT2/ 5HTlc receptor blocker, have been demonstrated in both animal and clinical studies (Ceulemans et al. 1985; Pangalila Ratu-Langi and Janssen 1988). According to the 5HT receptor imbalance theory, conditioned fear stimuli activate 5HT neurones in the dorsal raphe nucleus which project to 5HT2/lc receptors in amygdala and other forebrain structures (Deakin 1989). The anxiolytic effect of ritanserin would thus be mediated by blockade of conditioned fear responses. We have therefore investigated the effect of ritanserin on aversive conditioning in humans. There are no studies of the effect of 5HT-related drugs on aversive classical conditioning, although this is thought to be important in the aetology of morbid anxiety (Eysenck 1979; Gray 1987). For example, Gray proposes that anxiety is related to a "behavioural inhibition system" (BIS), which is activated by signals of novelty, innate fears, or by secondary signals of frustration and/or punishment. The latter are learned through classical conditioning (Gray 1982, 1987). The skin conductance response is a convenient, reliable and objective method of measuring reactions of individuals to significant stimuli. In addition, it has been proposed that electrodermal activity is more closely related to aversive rather than appetitive stimulation (Fowles 1988). We have recently developed a single-trial Pavlovian aversive classical conditioning paradigm using auditory stimuli and skin conductance measurements in human volunteers and anxious patients. The paradigm has been able to detect associative effects between conditioned (CS) and unconditioned (UCS) stimuli in volunteers and also to differentiate between anxious patients and volunteers (Ashcroft et al. 1991 ; Guimar~es et aI. 1991).

221 1

W e have used this p a r a d i g m to test the t h e o r y that 5 H T 2 receptors are involved in anxiety m e c h a n i s m s and to investigate the n a t u r e o f the involvement by observing the effects o f ritanserin in healthy volunteers.

(2,71

SCR

-OA (0.67) -1,8 (o.16)

Material

-3.2 (0,04)

Subjects. Thirty healthy volunteers, 15 males and 15 females, aged between 18 and 24 years and recruited from the first year of a University management sciences degree course participated in this study as paid volunteers. All subjects were drug free for at least 2 weeks (with the exception of oral contraceptives) at the time of the study. A medical examination excluded any physical and/or psychiatric disease. Written informed consent was obtained from each subject. Subjects were randomly divided into three groups and received 10 mg ritanserin, matching placebo or no pill at all. The investigator was unaware of the subject's treatment. The apparatus and procedure have been described in Guimarges et al. (1991). At the end of the experiment all subjects completed the Eysenck personality inventory (EPI) (1964).

-4.6, (0.01)

Statistical analys&. All subjects tested and all responses obtained were included in the analysis. Values for skin conductance response (SCR) amplitude and skin conductance level (SCL) were not found to be normally distributed and were converted to their natural log (logn) equivalents (with the addition of a constant value of 0.01) following the procedure suggested by Venables and Christie (1980). Separate multivariate analyses of variance (MANOVA) for SCR, SCL, spontaneous fluctuations (SF) and unconditioned response (UCR) were carried out using the Statistical Package for Social Sciences (SPSS/PC +, version 3.0). The degrees of freedom of the univariate tests were corrected by the Huynh-Feldt epsilon (*indicates adjusted degree of freedom; Jennen-Steinmetz 1989). The factors were: drug, sex, trials (ten trials) and conditioning phase (first versus second ten trials, i.e. habituation versus extinction). Confirming previous findings with the model (Guimarges et al. 1991), there was no difference on any measurement between placebo and no-pill groups. Therefore, the placebo and no-pill groups were analysed together as a control group.

Results

Skin conductance response amplitude Average S C R s in h a b i t u a t i o n and extinction are presented in Table 1, b r o k e n d o w n by drug treatment and sex. 1. Mean ( + / - S D ) skin conductance response amplitude in logn umhos before and after unconditioned stimulus. [ ] = detransformed means

Table

......."',

"~':;"-.,z.. .... 1 • ~-./ ......:'-. ~ ~

i

~

1

~

i

3

5

~

~

~

7

i 9

-'..

" ~ " - " xX// \ " i

~ 11

~' 13'

~

1 15

i

~ 17

'19'

~ 2t

STIMULI

30'

.o**,*** 25"

°.*" 20.

**°°

¢1

li

15Z < 10-

BEFORE

AFTER

Fig. 2. Mean number of spontaneous fluctuations before and after aversive conditioning in subjects treated with ritanserin (solid line) compared with healthy controls (dotted line)

The m a i n effect o f conditioning phase (/71,26=13.19, P = 0.001) indicates S C R s are greater after conditioning than before. The m a i n effect o f trials (*/'8,22o = 18.62, P < 0 . 0 0 1 ) indicates significant h a b i t u a t i o n and extinction occurred, but the conditioning x trials interaction (*Fs,zls = 1.99, P < 0 . 0 5 ) indicates h a b i t u a t i o n was significantly greater t h a n extinction. The conditioning phase effect was significantly less in ritanserin-treated volunteers (drug x conditioning phase, /'1,26 = 6.37, P = 0.0i 8; Fig. t). Ritanserin prevented conditioning to a significantly greater extent in males t h a n females (drug x s e x , F1,26 = 5.29, P = 0.03).

Control Group

"\i

Fig. 1. Mean skin conductance response amplitude in natural log gmho to successive tone presentations in ritanserin (solid line) compared to healthy controls (dotted line). Detransformed values in brackets

Drugs. Ritanserin and placebo pills were kindly supplied by Janssen Pharmaceutical, and were administered in identical gelatine capsules 16 min before the experiment.

t

"~:',,

and methods

Ritanserin Male

Female

Group

Male

Female

BEFORE UCS --2.9(1.1) --3.4(1.0) [0.061 [0.O3]

--2.4(0.9) [0.091

--3.0(1.4) [0.05t

--2.2(1.2) [0.11]

--3.8(1.3) [0.021

AFTER UCS - 1.9(1.1) -2.1(1.1) [0.15] [0.121

- 1.7(1.2) [0.181

-2.8(1.4) [0.061

2.3(1.3) [0.101

-3.3(1.4) [0.041

222 Table 2. Mean ( + / - S D ) number of spontaneous fluctuations before and after unconditioned stimulus

Table 4. Mean ( + / - SD) skin conductance response in logn gmho

Control

Control Group

Male

UCR 0.8(1.2) [2.23]

0.1(1.2) 1.4(0.7) 0.5(1.3) 0.9(0.7) 0.0(1.7) [1.111 [4.05] [1.651 [2.46] [1.00]

Group

to the unconditioned stimulus. [ ] = detransformed means

Ritanserin Male

Female

Group

Male

Female

BEFORE UCS 11.6(11) 9.8(10) 13.3(12) 11.6(13) 16.8(13) 6.4(12) AFTER UCS 29.2(20) 25.7(10) 32.7(26) 16.6(15) 23.2(15) 10.0(12)

Table 3. Mean (+/--SD) skin conductance level in logn gmho before and after unconditioned stimulus. [ ]= detransformed means

Table 5, Mean scores ( + / - SD) for extraversion (EXT), neuroticism (NEU) and lie (LIE) scales of the Eysenck personality inventory

Control Group

Female

Ritanserin Group Male

Female

Ritanserin Male

Female

Group

Male

Female

BEFORE UCS 0.0(1.2) --0.3(1.6) [1.00] [0~74]

0.4(0.8) [1.49]

--0.6(2.3) [0.55]

1.0(0.5) [2.72]

--2.0(2.1) [0.14]

AFTER UCS 0.8(0.8) 0.5(0.8) [2.23] [1.65]

1.1(0.7) [3.00]

0.0(1.7) [1.00]

1.3(0.3) [3.67]

-- 1.0(1.6) [0.37]

Control Group EXT 16.3(3.1) NEU 8.3(3.5) LIE 3.1(1.6)

Spontaneous fluctuations The mean number of SF before and after the unconditioned stimulus (UCS) are shown in Fig. 2 and Table 2. SFs increased after conditioning (conditioning phase main effect, Fl,z6= 14.39, P = 0 . 0 0 1 ) but this was almost absent in ritanserin-treated subjects (drug x conditioning phase, F1,26 = 4.49, P = 0.044).

Ritanserin Male

Female

Group

Male

Female

16.5(2.5) 8.5(4.0) 2.8(I.5)

16.1(3.8) 8.2(3.2) 3.5(1.6)

16.5(4.0) 8.5(3.4) 2.0(1.2)

16.8(4.2) 7.6(2.9) 1.8(1.5)

16.2(4.3) 9.4(4.0) 2.2(1.1)

tween d r u g x s e x (F1,26=7.43, P = 0 . 0 1 1 ) due to increased SCRs in ritanserin treated males and decreased in treated females (Table 4).

Eysenck personality inventory The results of the EPI are displayed in Table 5. There was no difference between groups.

Skin conductance level The mean SCL before and after UCS occurrence m a y be seen in Table 3. There were significant effects o f sex (F1,26=7.87, P = 0 . 0 0 9 ) , conditioning (Fl,z6=16.53, P < 0 . 0 0 1 ) , trials (*F2.74= 11.39, P < 0 . 0 0 1 ) and significant interactions of drug x sex (F1,26 = 19.96, P < 0.001) and of drug x sex x trials (*F2.74=4.04, P < 0 . 0 5 ) . Although a trend was found, there was no significant drug effect (FL26=3.37, P < 0 . 1 >0.05).

Unconditioned response Although there was no drug effect in the SCR to the UCS (F1,26 = 0.47, NS), there was a significant interaction be-

Discussion

We have previously shown that increases in S C R and SF following the tone-noise pairing involve an associative mechanism, since the presentation of the aversive noise alone did not increase subsequent SCRs or SFs. In the present experiment almost no increase in SCRs and SFs occurred in subjects treated with ritanserin. This cannot be attributed to non-specific peripheral or central suppressant effects of autonomic function, since SCRs and SFs prior to the conditioning trial were indistinguishable in ritanserin and drug-free subjects. Therefore, the results suggest ritanserin interferes either with the acquisition or with the expression of aversively conditioned responses.

223

In unpublished results, we found that 4 weeks of ritanserin treatment in anxious patients reduced SCRs to tones which had been paired with the loud noise before drug treatment. This suggests that ritanserin selectively reduces the expression of aversive conditioned responses, an effect that is shared by most anxiolytics (Gray 1982). For example, in an electrodermal conditioning experiment analogous to ours, Jensen et al. (1989) found that 10 mg diazepam in normal male volunteers reduced conditioned responses to tones presented 3-9 days after 24 aversive conditioning trials. Ritanserin, therefore, has an anxiolytic profile in the aversive skin conductance conditioning paradigm characterized by the suppression of conditioned aversive responses. This is in keeping with the 5HT receptor imbalance theory of affective disturbance (Deakin 1988, 1989). Benzodiazepines decrease 5HT release (Iversen 1984). The present results suggest that decreased 5HT2 function may mediate the shared ability of benzodiazepines and ritanserin to attenuate aversively conditioned SCRs and perhaps their ability to reduce anxiety. Diazepaminduced suppression of 5HT release would also reduce 5HT1 receptor function, whereas ritanserin enhances some 5HT1 mediated behaviours (Deakin 1989). This may explain why diazepam accelerates habituation, at least in anxious patients (Lader 1975), whereas ritanserin has a more selective effect on conditioned responses. However, diazepam clearly has many actions, any of which might accelerate habituation. Experiments in animals suggest that 5HT projections mediate behavioural responses to aversive stimuli (Deakin 1983). A competing view suggests an affectively neutral role of 5HT in restraining impulsive behaviour (Soubrie 1986). However, it is difficult to see how attenuation of aversively conditioned skin conductance responses can be attributed to response disinhibition. Furthermore, the increasing evidence of anxiolytic actions of drugs with selective actions on 5HT suggests that 5HT systems have an important role in mediating affective responses. Dorsal raphe projections to the amygdata may be particularly involved in mediating conditioned fear responses (Deakin 1983; Graeff 1989). Lesions of the amygdaloid complex reduce measures of fear conditioning in several experimental paradigms (Gray 1987), and Davis (1989) proposed that "fear conditioning probably involves neural plasticity afferent to or in the amygdala rather than a change in its efferent target areas". 5HT may modulate or mediate this process, since non-selective 5HT antagonists (e.g. methysergide and cyproheptadine) and the selective 5HT2/alpha 1 blocker ketanserin have shown antipunishment effects when microinjected into the amygdala (Petersen and Scheel-Krueger 1984; Hodges et al. 1987; Kataoka et al. 1987). Neither ritanserin nor trazodone (also a 5HT2/lc antagonist) are effective in the treatment of panic (Charney et al. 1986; Der Boer and Westenberg 1990). However, panic attacks are unconditioned episodes of anxiety and may not involve the 5HT2/lc-conditioning mechanism which is probed in the present paradigm. 5HT reuptake blockers are eventually effective in the treat-

ment of panic. Thus 5HT may have distinct roles in modulating different forms of anxiety. In conclusion, our results show that ritanserin selectively impairs skin conductance responses to aversive conditioned stimuli without affecting responses to novelty. Ritanserin has an unusual non-sedative anxiolytic profile in this paradigm. The results are compatible with the 5HT imbalance theory of affective disturbance (Deakin 1989). It is argued that 5HT2/lc receptors in the amygdala may mediate or modulate conditioned fear responses and that attenuation of aversively conditioned responses might be a mechanism by which ritanserin exerts anxiolytic effect. Acknowledgement. Dr. F.S.G. is the recipient of a CAPES Fellowship.

References Ashcroft K, Guimarges FS, Wang M, Deakin JFW (1991) Evaluation of a psychophysiological model of classical fear conditioning in anxious patients. Psychopharmacology 104: 215-219 Ceulemans DLS, Hoppenbrowers MLJA, Gelders YG, Reyntjens AJM (1985) The influence of ritanserin, a serotonin antagonist, in anxiety disorders: a double-blind placebo-controlled study versus lorazepam. Pharmacopsychiatry 18:303-305 Charney DS, Woods SW, Goodman WK, Rifkin B, Kinch M, Aiken B, Quadrino LM, Heninger GR (1986) Drug treatment of panic disorder: the comparative efficacy of imipramine, alprazolam, and trazodone. J Clin Psychiatry 47:580-586 Davis M (1989) The role of the amygdala and its efferent projections in fear and anxiety. In: Tyrer P (ed) Psychopharmacology of anxiety. Oxford University Press, Oxford Deakin JFW (1988) 5HT2 receptors, depression and anxiety. Pharmacol Biochem Behav 29:819-820 Deakin JFW (1989) Role of 5HT receptor subtypes in depression. In : Archer T, Bevan P, Cools A (eds) Behavioural pharmacology of 5HT. Lawrence Erlbaum, New York Den Boer JA, Westenberg HGM (1990) Serotonin function in panic disorder: a double-blind placebo controlled study with fluvoxamine and ritanserin. Psychopharmacology 102:85-94 Eysenck HJ (1979) The conditioning modet of neurosis. Behav Brain Sci 2:155-199 Eysenck HJ, Eysenck SBG (1964) Eysenck personality inventory. University of London Press, London Fowles DC (1988) Psychophysiology and psychopathology: a motivational approach. Psychophysiology 25:373-391 Fuxe K, Ogren S, Agnati L, Gustafsson JA, Jonsson G (1977) On the mechanism of action of the antidepressant drugs amitriptyline and nortriptyline. Evidence for 5-hydroxytryptamine receptor blocking activity. Neurosci Lett 6:339-343 Goldberg DP, Bridges P, Ducan-Jones P, Grayson D (1987) Dimensions of neuroses seen in primary care settings. Psychol Med 17 : 461-470 Graeff F G (t989) Serotonin and aversion. In: Bevan P, Cools AR, Archer T (eds) Behavioural pharmacology of 5-HT. Lawrence Erlbaum, New York Gray JA (I982) The neuropsychology of anxiety: an inquiry into the functions of the septo-hippocampal system. Oxford University Press, Oxford Gray JA (1987) The psychology of fear and stress, 2nd edn. Cambridge University Press, Cambridge Guimarg.es FS, Hellewell J, Hensman R, Wang M, Deakin JFW (1991) Characterization of a psychophysiological model of classical fear conditioning in healthy volunteers: influence of

224 gender, instruction, personality and placebo. Psychopharmacology 104:231-236 Hodges H, Green S, Glenn B (t987) Evidence that the amygdala is involved in benzodiazepine and serotonergic effects of punished responding but not of discrimination. Psychopharmacology 92: 491-504 Iversen SD (1984) 5-HT and anxiety. Neuropharmacology 23:1553-1560 Jennen-Steinmetz C (1989) Synopsis of repeated measurement analysis. J Psychophysiol 3:193-194 Jensen HH, Hutchings B, Poulsen JC (1989) Conditioned emotional responding under diazepam: a psychophysiological study of the state dependent learning. Psychopharmacology 98: 312-397 Johnstone EC, Cunningham Owens DG, Frith CD, McPherson K, Dowie C, Riley G, Gold A (1980) Neurotic illness and its response to anxiolytic and antidepressant treatment. Psychol Med 10: 321-328 Kataoka Y, Shibata K, Yamashita K, Ueki S (1987) Differential mechanisms involved in the anticonflic action of benzodiazepines injected into the central amygdala and mammilary body. Brain Res 416:243-247

Khan RJ, McNair DM, Lipman RS, Covi L, Rickels K, Dowing R, Fischer S, Frankenthaler LM (1986) Imipramine and chlordiazepoxide in depressive and anxiety disorders. Arch Gen Psychiatry 43 : 7%85 Lader M (1975) The psychophysiology of mental illness, Routledge & Kegan Paul, London Pangalila Ratu-Langi EA, Janssen AAI (1988) Ritanserin in the treatment of generalized anxiety disorder: a placebo-controlled trial. Hum Psychopharmacol 3 : 207-212 Peroutka SJ, Snyder SH (1980) Long-term antidepressant treatment decreases spiroperidol-labelled serotonin receptor binding. Science 210:88-90 Petersen EN, Scheel-Kruger J (1984) Anticonflict effects of 5-HT antagonists by intraamygdaloid injection. Collegium Internationale Neuro-Psychopharmacologium, 14th CINP congress book of abstracts: 654 Soubrie P (1986) Reconciling the role of central serotonin neurons in human and animal behaviour. Behav Brain Sci 9:319-364 Venables PH, Christie MJ (1980) Electrodermal activity. In: Martin I, Venables PH (ed) Techniques in psychophysiology. John Wiley, Chichester