260185.qxd

Richard J. Porter · R. Hamish McAllister-Williams
Allan H. Young
Acute effects of venlafaxine and paroxetine on serotonergic transmission in human volunteers Received: 11 January 1999 / Final version: 17 May 1999 Abstract Rationale: Antidepressant drugs are thought
receptor function in man (Smith et al. 1991) and has con- to enhance serotonergic neurotransmission through post- sistently been shown to be blunted in depressive illness receptors. This effect is delayed in ani- (Power and Cowen 1992). This abnormality appears to mals and may be paralleled by a delay in the onset of a resolve following successful treatment (Upadhyaya et al.
clinical response in humans. In humans, the growth 1991) and there is evidence that specific antidepressant hormone (GH) response to intravenous L-tryptophan healthy volunteers and depressed subjects (Price et al.
and the prolactin response is blunted. Both are therefore 1990). However, the time course of this enhancement and its relationship to the onset of antidepressant action tion. Clomipramine has previously been found to en- hance the GH and prolactin responses to IV L-TRP after It has been suggested that delay in the onset of clini- only 2 h. Objective: The purpose of this study was to use cal response in antidepressant therapy may be due to a this method to investigate the effects of newer antide- delay in the increase in transmission through postsynap- receptors and that this delay is due to the in- hods: Twelve healthy male volunteers took part in a ran- fluence of negative feedback mediated by somatoden- dom order, double blind study, in which 18.75 mg venla- receptors (Blier and de Montigny 1994).
faxine, 5 mg paroxetine or placebo was administered 3 h This conflicts, however, with the results of Anderson before infusion of L-TRP. Results: Pretreatment with and Cowen (1986) which show that the tricyclic anti- venlafaxine significantly enhanced the growth hormone depressant clomipramine increases the GH response to (GH) response to the infusion compared with pretreat- IV L-TRP only 2 h after pretreatment in healthy volun- ment with placebo. There was no significant difference teers. The acute effects of other antidepressants on neu- between the GH response following paroxetine com- roendocrine responses to IV L-TRP have not, however, pared with placebo or with venlafaxine. Conclusions: The data suggest enhancement of transmission through We sought to investigate the acute effects on the GH agents with more selective profiles than clomipramine:paroxetine, a selective serotonin re-uptake inhibitor and Key words L-Tryptophan · Receptor · Serotonin ·
venlafaxine, a specific serotonin and noradrenaline re- Growth hormone · Prolactin · Human volunteer · The GH response to the intravenous infusion of L-TRP Twelve healthy male volunteers, aged 18–40 years (mean 28.1, SD appears to be a useful measure of postsynaptic 5-HT1A 5.7), gave their informed consent to the study which was approved by the local Ethics Committee. They had no history of significant R.J. Porter · R.H. McAllister-Williams · A.H. Young (✉) psychiatric or physical illness and had been on no medication for Department of Neuroscience and Psychiatry, University of Newcastle, Royal Victoria Infirmary, Subjects were tested on three occasions, at least 3 weeks apart, having taken pretreatment medication at 0700 hours on the morn- e-mail: A. H.Young@ncl.ac.uk, Fax: 44-191-227-5108 ing of testing. Pretreatment medication consisted of placebo, par- oxetine 5 mg or venlafaxine 18.75 mg administered orally in abalanced order, double-blind, cross-over design. Following anovernight fast, subjects attended the research laboratory at0900 hours, when an intravenous cannula was inserted. This waskept patent with heparinised saline. Subjects fasted throughout theexperiment, remained semi-supine and were not allowed to sleep.
After 1 h, an infusion of L-TRP (in aqueous solution 10 g/l) wasgiven, at a dose of 100 mg/kg, over 25 min. Blood samples weretaken every 15 min from 30 min before the infusion (–30 min,–15 min and time 0) and every 15 min from 5 min until 95 min af-ter completion of the infusion (+5 min, +20 min etc.).
Blood samples were taken into EDTA tubes and centrifuged to re-move plasma. This was stored at –20°C. Plasma was also ultra-fil-tered and stored until assay. Samples were analysed for prolactin,GH and cortisol by standard radioimmunoassay. Free TRP was Fig. 1 Effect of pretreatment with paroxetine or venlafaxine or
measured using high performance liquid chromatography (Mar- placebo on GH response to IV infusion of L-TRP. GH levels shall et al. 1987). Intra- and inter-assay coefficients of variation (mIU/l) are plotted as mean±SEM against time. The time of the for prolactin were 5.7% and 6.4%, respectively, for GH 2.6% and infusion is indicated by bold arrows. ▼ Venlafaxine, ■ paroxe- 7.4% and free TRP 3.4%, 4.4% and total TRP 3.3% and 4.4%.
and data are therefore missing for this trial. Data fromthis subject are included in the post hoc comparison of SPSS for Windows Release 7 (SPSS, Chicago, Ill., USA) was paroxetine versus placebo, but not in the repeated mea- used for statistical analysis. In all cases the Kolmogorov-Smirnovtest was used to exclude any significant departure from a normal sures ANOVA, in which data from the remaining nine distribution. The biochemical and hormonal data were analysed using a two-way repeated measures analysis of variance (AN-OVA), with drug (paroxetine/venlafaxine/placebo) and time aswithin subject variables. The reported P values of all ANOVAsused the Huynh-Feldt correction factor when the sphericity as- sumption was not met. For clarity, uncorrected degrees of freedomare reported.
One set of data was excluded (venlafaxine trial) because Hormonal responses were also analysed using the trapezoid ar- of high baseline GH values (>10 mIU/l), since GH inhib- ea under the curve (AUC) method. This was measured from theaverage of the three baseline measures taken prior to infusion of its its own secretion (Checkley 1980). This subject’s TRP. AUCs and average baseline measures were analysed by one- baseline GH values after paroxetine and placebo were way repeated measures ANOVA with drug as a within-subjects within accepted limits. ANOVA is therefore reported on factor. Both AUCs and baselines were then analysed for three sep- data from eight subjects. ANOVA showed a significant arate comparisons (paroxetine versus placebo, venlafaxine versus placebo and paroxetine versus venlafaxine) using post-hoc paired t-tests (two-tailed). These data are quoted as means±SEM.
(F=4.79; df=2,14; P=0.026), a significant drug by timeinteraction (F=2.80; df=18,126; P=0.021) and a signifi-cant effect of time (see Fig. 1 and Table 1). ANOVA of AUCs showed a significant effect of drug (F=0.038;df=2,14; P=0.038). Post-hoc analysis of AUC measures Two subjects failed to complete the second trial because showed a significant difference between venlafaxine and of intolerance of side effects (nausea and vomiting) and placebo (venlafaxine 1676±338; placebo 598±146; t= because the analysis was within subjects the data from 2.71; df=1,7; P=0.030; 95% CI 138-2019) but no signifi- the first trial was not used (order of administration – par- cant difference between AUC for paroxetine versus pla- oxetine/venlafaxine and venlafaxine/paroxetine). One cebo (paroxetine 786±214; placebo 600±117; t=0.64; subject did not complete the third trial, when venlafaxine df=1,9; P=0.537; 95% CI-469-842) or venlafaxine ver- had been administered, (paroxetine/placebo/venlafaxine) sus paroxetine (venlafaxine 1676±338; paroxetine 961± Table 1 Effects of venlafaxine, paroxetine and placebo pretreatment on responses to L-TRP infusion: summary of analysis of variance
results. Significant findings are shown in bold
Table 2 Effect of pretreatment with venlafaxine and paroxetine
of missing data for two subjects from the venlafaxine arm of the on mean baseline measures and responses (calculated as trapezoid study, paired t-tests for comparisons with venlafaxine used mean area under the curve) to L-TRP infusion for all subjects. Because AUCs which were different from those given in the table Table 2 for mean values). There was no effect of drug onmean baseline values (F=0.59; df=2,16; P=0.540).
The main finding of the study is that pretreatment with18.75 mg venlafaxine 3 h before infusion, significantlyenhanced the GH response to infusion of L-TRP com-pared with placebo. There was no significant effect ofparoxetine compared with placebo and no significant ef-fect of either drug on prolactin responses.
The GH response to L-TRP has been shown to be at- 1991). Although pindolol has β-antagonistic properties Fig. 2 Effect of pretreatment with paroxetine or venlafaxine or
in dynamic tests of noradrenergic function (Aellig 1976), placebo on prolactin response to IV infusion of L-TRP. Prolactin propranolol, a β-antagonist which has a much lower af- levels (mIU/l) are plotted as mean±SEM against time. The dura- tion of the infusion is indicated by bold arrows. ▼ Venlafaxine, increases rather than decreases the GH response to L-TRP (Upadhyaya et al. 1990). This suggests that theeffect of pindolol in inhibiting the GH response to L-TRPis not mediated via β-adrenoceptor blockade but proba- 227; t=1.75; df=1,9; P=0.123; 95% CI –1681–250) (see Table 2 ). ANOVA showed no effect of drug on mean sponse was not attenuated by the non-selective 5-HT an- baseline values (F=0.29; df=2,14; P=0.653).
tagonist metergoline (McCance et al. 1987). The expla-nation for this may lie in a relative lack of effective an-tagonism of 5-HT been demonstrated in functional studies in animals(Koenig et al. 1987). These findings therefore support a ANOVA showed no significant effects of drug or signifi- cant drug by time interaction, but a significant effect of time (Fig. 2, Table 1). There was no effect of drug on There is evidence that the PRL response to L-TRP in- AUCs (F=2.02; df=2,16; P=0.172) and post-hoc compar- fusion may have a dopaminergic component (van Praag ison of AUCs showed no significant differences between et al. 1987). Pindolol causes markedly less attenuation of the conditions (see Table 2 for mean values). There was the PRL response than the GH response to L-TRP (Smith no effect of drug on mean baseline values (F=0.66; et al. 1991). L-TRP competes with tyramine for transport across the blood-brain barrier (Wurtman 1982) and mayreduce dopamine synthesis by reducing brain tyramine.
This is supported by evidence that an intravenous infu- sion of 5 g L-TRP causes a reduction in post probenecidcerebrospinal fluid (CSF) concentrations of the dopa- ANOVA showed a significant effect of time but no drug mine metabolite homovanillic acid (HVA) (van Praag et effect or drug by time interaction (see Table 1). There al. 1987). The PRL response to L-TRP may therefore be was no effect of drug on AUCs (F=0.36; df=2,16; mediated in part by a reduction in dopamine synthesis, P=0.625) and post-hoc comparison of AUCs showed no which releases PRL secretion from tonic inhibition by significant difference between the three conditions (see Venlafaxine is a potent serotonin reuptake inhibitor faxine is unlikely to be a major factor at the dose em- and a noradrenaline reuptake inhibitor. The active metab- ployed. It has been suggested that venlafaxine may pro- olite O-desmethyl-venlafaxine (ODV) has similar prop- duce an earlier onset of clinical response than other anti- erties and both are weak dopamine reuptake inhibitors.
depressants (Montgomery 1995) and that this may relate (Muth et al. 1991). The time to maximum concentration to early downregulation of βb adrenoceptors (Moyer et of venlafaxine is approximately 2 h and for ODV ap- al. 1992). However, more recent research suggests that proximately 4 h (Morton et al. 1995). At 3 h (the time early β-adrenoceptor downregulation may only occur in between administration and infusion in this study) the the pineal gland, which lacks serotonergic innervation, concentrations of each could therefore be expected to be and that generally venlafaxine does not produce this ef- nearly maximal. Preliminary reports in humans suggest fect (Nalepa et al. 1998). While venlafaxine is a potent that at lower doses of venlafaxine (75 mg/day), there is serotonin re-uptake inhibitor in vivo (Beique et al.
little effect on noradrenergic reuptake (Debonnel et al.
1996), it has comparatively less potency than paroxetine 1998). At the dose used in this study (18.75 mg), norad- in blocking serotonin reuptake in vitro (Bolden-Watson renergic reuptake inhibition may not therefore be a sig- and Richelson 1993). A recent study in rat brain shows a relatively low binding (2000 times less than paroxetine) This may also be the case in the study of Anderson of venlafaxine to the 5-HT transporter. It has been sug- and Cowen (1986), which demonstrated that the GH re- gested, therefore, that a different mechanism of function- sponse to L-tryptophan infusion was enhanced by pre- al 5-HT re-uptake inhibition may be involved (Beique et treatment with clomipramine. Clomipramine differs from al. 1998). This might be important in determining the venlafaxine in that it does possess anticholinergic and antihistaminergic properties (Hall and Ogren 1981).
While its metabolite, desmethylclomipramine, is a potent Both L-tryptophan and L-5-hydroxytryptophan (5- noradrenaline reuptake inhibitor, clomipramine is not HTP) have been found to depress raphé neuron firing in (Carlsson et al. 1969a, 1969b). At the time of infusion in animals (Gallager and Aghajanian 1976). Electrophysio- the Anderson and Cowen study, little desmethylclomip- logical studies also suggest that acute administration of ramine would have been present (Jones and Luscombe SSRIs inhibit firing of serotonergic neurones (Chaput et 1976), suggesting that noradrenergic reuptake inhibition al. 1986; Hajós et al. 1995) and a recent study shows that would be unlikely to contribute significantly to the re- this is also the case for venlafaxine (Gartside et al.
1997). However, both 5-HTP (Gartside et al. 1992) and Paroxetine did not significantly increase either the antidepressants (Fuller 1994) increase cortical extracel- GH or prolactin response to IV L-TRP. As discussed, at lular 5-HT following acute treatment, an increase which the doses employed, clomipramine (Anderson and is proportionally smaller in the cortex than in the raphé Cowen 1986) and venlafaxine were probably acting pri- nuclei (Artigas 1993). It could be argued that this is sec- marily via effects on serotonin reuptake. It would there- ondary to an unphysiological “spilling” of 5HT from in- fore be expected that paroxetine would have similar ef- traneuronal stores into the synapse, in the absence of cell fects. We chose dosages which were 25% of the usual firing. However, the acute increase in extracellular 5-HT daily starting dose (paroxetine 20 mg, venlafaxine seen following treatment with 5-HTP (Gartside et al.
75 mg). In both cases, the starting dose is usually an ef- 1992) and antidepressants (Rutter and Auerbach 1993) is fective antidepressant dose. A small dose was used be- cause we were concerned that the side effects of the infu- 2-(di-n-propylamino)tetralin, which shuts down seroto- sion would be amplified by the pretreatments. In fact, nin neuronal firing. The rise in extracellular 5-HT fol- even at the low doses employed, some subjects were un- lowing 5-HTP also appears to be calcium dependant, fur- able to tolerate the protocol. Clearly, a significant effect ther suggesting a dependence upon neuronal firing of paroxetine might have been seen at higher doses. In (Gartside et al. 1992). Since L-tryptophan is the precur- addition, the mean time to peak plasma concentrations of sor to 5-HTP, we would argue that the response to the in- paroxetine is 5 h (Holliday and Plosker 1993). Peak plas- fusion and its enhancement by venlafaxine are not sim- ma levels were therefore probably not reached at the ply due to leakage of 5-HT and inhibition of its reuptake.
time of infusion. It is notable that the GH response to IV Our results suggest that venlafaxine, in common with L-TRP following venlafaxine was not significantly great- clomipramine, enhances transmission through postsyn- er when directly compared with that following paroxe- receptors after only 3 h. Whether this effect tine. Using a larger number of subjects, a significant en- in humans is specific to these drugs, or occurs with other hancement by paroxetine compared with placebo or pos- classes of antidepressants and what is the exact mecha- sibly of venlafaxine compared with paroxetine, may nism by which this occurs, is unclear at present.
The finding of a significant enhancement with venla- Acknowledgements This study was supported by a grant from
faxine as opposed to paroxetine at roughly equivalent Wyeth Laboratories. We thank D. Nelson and M. Leitch for bio-chemical measurements.
doses may relate to various putative differences in theirpharmacological and clinical profile. However, we be-lieve that the noradrenaline reuptake inhibition of venla- Hall H, Ogren S (1981) Effects of antidepressant drugs on differ- ent receptors in the brain. Eur J Pharmacol 70:393–407 Holliday SM, Plosker GL (1993) Paroxetine: a review of its phar- Aellig WH (1976) β-Adrenoceptor blocking activity and duration macology, therapeutic use in depression and therapeutic poten- of action of pindolol and propranolol in healthy volunteers. Br tial in diabetic neuropathy. Drugs Aging 3:278–299 Hoyer D (1988) Functional correlates of serotonin 5-HT recogni- Anderson IM, Cowen PJ (1986) Clomipramine enhances prolactin and growth hormone responses to L-tryptophan. Psychophar- Jones RL, Luscombe DK (1976) Plasma levels of clomipramine and its N-desmethyl metabolite following oral clomipramine Artigas F (1993) 5-HT and antidepressants: new views from mi- crodialysis studies. Trends Pharmacol Sci 14:262 Koenig JI, Gudelsky GA, Meltzer HY (1987) Stimulation of corti- Beique JC, de Montigny C, Blier P, Debonnel G (1996) Blockade costerone and beta-endorphin secretion in the rat by selective of 5-HT and NE reuptake by venlafaxine: in vivo electrophysi- 5-HT receptor subtype activation. Eur J Pharmacol 137:1–8 ological studies in the rat. Soc Neurosci Abstr 22:180 Marshall EF, Kennedy WN, Eccleston D (1987) Whole blood se- Beique JC, Lavoie N, de Montigny C, Debonnel G (1998) Affini- rotonin and plasma tryptophan using high-pressure liquid ties of venlafaxine and various reuptake inhibitors for the sero- chromatography with electrochemical detection. Biochem tonin and norepinephrine transporters. Eur J Pharmacol 349: McCance SL, Cowen PJ, Waller H, Grahame-Smith DG (1987) Blier P, de Montigny C (1994) Current advances and trends in the The effects of metergoline on endocrine responses to L-trypto- treatment of depression. Trends Pharmacol Sci 15:220– 226 Bolden-Watson C, Richelson E (1993) Blockade by newly-devel- Montgomery SA (1995) Rapid onset of action of venlafaxine. Int oped antidepressants of biogenic amine uptake into rat brain Morton W, Sonne S, Verga M (1995) Venlafaxine: a structurally Carlsson A, Corrodi H, Fuxe K, Hökfelt T (1969a) Effects of unique and novel antidepressant. Ann Pharmacother 29:387– some antidepressant drugs on depletion of intraneuronal brain catecholamine stores caused by 4-alpha-dimethyl-metatyra- Moyer JA, Andree TH, Haskins JT, Husbands G, Muth EA (1992) The preclinical pharmacological profile of venlafaxine: a nov- Carlsson A, Corrodi H, Fuxe K, Hökfelt T (1969b) Effect of anti- el antidepressant agent. Clin Neuropharmacol 15:435B depressant drugs on the depletion of intraneuronal brain 5-hy- Muth E, Moyer J, Haskins J, Andree T, Husbands G (1991) Bio- droxytryptamine stores caused by 4-alpha-dimethyl-metatyra- chemical, neurophysiological and behavioural effects of Wy- 45233 and other identified metabolites of the antidepressant Chaput Y, Blier P, de Montigny C (1986) In vivo electrophysiolog- ical evidence for the regulatory role of autoreceptors on sero- Nalepa I, Manier DH, Gillespie DD, Rossby SP, Schmidt DE, Sul- tonergic terminals. J Neurosci 6:2796–2801 ser F (1998) Lack of beta adrenoceptor desensitisation in brain Checkley SA (1980) Neuroendocrine tests of monoamine function following the dual noradrenalin and serotonin reuptake inhibi- in man: a review of basic theory and its application to the tor venlafaxine. Eur Neuropsychopharmacol 8:227–232 study of depressive illness. Psychol Med 10:35–53 Power AC, Cowen PJ (1992) Neuroendocrine challenge tests: as- Debonnel G, Blier P, Saint-Andre E, Hebert C, de Montigny C sessment of 5-HT function in anxiety and depression. Mol As- (1998) Comparison of the effects of low and high doses of venlafaxine on serotonin and norepinephrine reuptake process- Price LH, Charney DS, Delgado PL, Goodman WK, Krystal JH, es in patients with major depression and healthy volunteers.
Woods SW, Heninger GR (1990) Clinical studies of 5-HT function using IV L-tryptophan. Prog Neuro-Psychopharmacol Fuller RW (1994) Uptake inhibitors increase extracellular seroto- nin concentration measured by brain microdialysis. Life Sci Smith CE, Ware CJ, Cowen PJ (1991) Pindolol decreases prolactin and growth hormone responses to intravenous L-tryptophan.
Gallager DW, Aghajanian GK (1976) Inhibition of firing of raphe neurones by tryptophan and 5-hydroxytryptophan: blockade Upadhyaya AK, Deakin JF, Pennell I (1990) Hormonal response by inhibiting serotonin synthesis with Ro-4-4602. Neurophar- to L-tryptophan infusion: effect of propranolol. Psychoneuro- Gartside SE, Cowen PJ, Sharp T (1992) Effect of 5-hydroxy-L- Upadhyaya AK, Pennell I, Cowen PJ, Deakin JF (1991) Blunted tryptophan on the release of 5-HT in rat hypothalamus in vivo growth hormone and prolactin responses to L-tryptophan in as measured by microdialysis. Neuropharmacology 31:9–14 depression; a state-dependent abnormality. J Affect Disord Gartside SE, Umbers V, Sharp T (1997) Inhibition of 5-HT cell firing in the DRN by non-selective 5-HT reuptake inhibitors: van Praag HM, Lemus C, Kahn R (1987) Hormonal probes of cen- tral serotonergic activity: do they really exist? Biol Psychiatry mechanisms. Psychopharmacology 130:261–268 Hajós M, Gartside SE, Sharp T (1995) Inhibition of median and Wurtman RJ (1982) Nutrients that modify brain function. Sci Am dorsal raphé neurones following administration of the selective serotonin reuptake inhibitor paroxetine. Naunyn-Schmiede-berg’s Arch Pharmacol 351:624–629

Source: http://blenny.ncl.ac.uk/r.h.mcallister-williams/reprints/ven_par.pdf