Several lines of evidence indicate since the early discovery of Dahlstom and Fuxe [48] that the cell bodies of serotonergic neurons are concentrated in the midbrain areas. Of the serotonin (5-HT) cell groups present in the brainstem, the dorsal and median raphe nuclei have been studied most frequently.At least 50% of all serotonergic neurons of the central nervous system are located in the dorsal raphe nucleus[175], which endows an important interface function with its extensive afferent and efferent connections [126]. Neurotransmitter systems often interact in a reciprocal fashion in the central nervous system. This may also be true for the raphe nuclei in which substantial neural elements, the medium-sized serotonergic neurons, the GABAergic interneurons and the glutamatergic axon terminals modulate each other’s function [118, 119]. The aim of this review was to overview auto- and heteroregulation of serotonergic projection neurons by various pharmacological agents, and also to discuss possible reciprocal interactions among the main neurotransmitter systems within the raphe nuclei.

Serotonergic neurons in the midbrain raphe nuclei exhibit spontaneous discharge activity of 1 to 5 spikes per second [163]. The activity of serotonergic neurons in the raphe nuclei is determined by at least three different events. These are the spontaneous, slow firing activity of 5-HT neurons, the 5-HT release-mediated autoinhibition of the neurons and the excitatory and inhibitory afferent inputs mediated by non-5-HT receptors [126]. The autoinhibitory properties of serotonergic neurons, which include regulation of transmitter release and synthesis and neural firing, are determined by various metabotropic (5-HT_1A/1B/1D, 5-HT_2A/2C, 5-HT₇) and ionotropic (5-HT₃) 5-HT receptors [11, 15, 180]. Different 5-HT₁ receptor subtypes are present on 5-HT-containing neurons in the raphe nuclei and they inhibit 5-HT release [49, 125, 149, 153] albeit their regulatory function in this process is different. In the raphe nuclei, few data are available about the types of 5-HT receptors mediating heteroreceptor functions i.e. those located on other neurons. Postsynaptic 5-HT receptors are the target of 5-HT released from dendritic varicosities into the synaptic gap or into the extracellular space as both synaptic and non-synaptic 5-HTergic neurotransmission may occur in the raphe nuclei [43, 73].

The release of 5-HT in the raphe nuclei originates from the somatodendritic area, the recurrent axon collaterals or axon terminals and the regulation of 5-HT release from these sites may be different. Dendritic 5-HT release originates from serotonergic neurons with vesicles within the dendrites [82], which then may provide neurochemical basis for synchronous firing pattern observed among dorsal raphe serotonergic cells. A further piece of evidence for the presence of 5-HT receptors on serotonergic neurons is that iontophoretically applied 5-HT decreases the firing of raphe neurons. Besides, 5-HT released from 5-HT-containing neurons, the repetitive cycles of hyper- and depolarization of 5-HTcontaining cells are altered by heteroregulation of excitatory and inhibitory projection neurons and interneurons [125].

The ascending projections of raphe serotonergic neurons influence several central nervous functions in different forebrain structures [121, 154]. In vivo electrophysiological experiments indicate that activation of ascending serotonerg pathway causes a poststimulus inhibition of 5-HT cell firing in the dorsal raphe nucleus [71]. This effect is mediated through 5-HT released from dendrites and axon varicosities within the raphe nuclei. 5-HT release can be demonstrated from both somatodendritic and axon terminal parts of these neurons, although the two areas show several differences in 5-HT content, turnover rate, and receptor density [8]. The function of 5-HT autoreceptors located on somatodendritic and axon terminal areas is also different: the former regulates dendritic 5-HT release and action potential generation, and the latter primarily governs impulse-mediated 5-HT release of vesicular origin [180].

The raphe-hippocampal pathway is one of the numerous ascending serotonergic projections in the brain [55]. The caudate nucleus also represents an important projection field for serotonergic neurons located in the raphe nuclei [162, 168]. Serotonergic neurons of the raphe nuclei form extended projections to the cerebral cortex [71, 94, 146] and the raphe-cortical 5-HT neurons mediate an inhibition in the cortical neural network [116, 154]. In many cases, forebrain structures that receive serotonergic innervations from the raphe nuclei send neuronal inputs back to the midbrain serotonergic nuclei [165]. Morphological evidence has been presented for the existence of a glutamatergic pathway arising from the median prefrontal cortex and projecting to the raphe nuclei [16]. Furthermore, it has been shown that GABA interneurons in the raphe nuclei are the primary targets for the cortico-raphe glutamatergic neurons [42] and their stimulation by glutamate release leads to inhibition of 5-HTergic cells [71, 165]. Serotonergic neurons in the raphe nuclei are regulated by a number of efferent projections. Thus, afferents to the dorsal raphe nucleus originate from the substantia nigra [91, 155] or from the hypothalamus [103]. The habenulo-raphe projection represents a major GABAergic inhibition to the raphe nuclei cells [5, 155, 174]. The habenuloraphe pathway also contains excitatory transmitters and the presence of substance P- and glutamate-containing fibers was reported. Another neural pathway that projects onto the raphe serotonergic neurons is noradrenergic and mediates its effect through postsynaptically located α-adrenoceptors [13, 23, 124]. Noradrenergic projections for the dorsal and median raphe nuclei arise from the locus coerulus and the lateral tegmental area, respectively [111, 138]. Hopwood and Stamford [84] have reported that both alpha-1 and alpha-2 adrenoceptors are involved in inhibition of stimulated 5-HT release from rat dorsal and median raphe nuclei slices.

Fig. (1) Preparation of raphe nuclei slice from rat brain. AC, cerebral aqueduct; DR, dorsal raphe; MR, median raphe; PCS, superior cerebellar peduncle. Vertical line indicates direction of brain transection.

When the release of amino acids from raphe nuclei slices is measured, the slices are incubated with either [₃H]glutamate or [³H]GABA dissolved in aerated Krebs-bicarbonate buffer at 36 °C for 30 min. For [₃H]GABA release, raphe nuclei slices are loaded in the presence of β-alanine, an inhibitor of GABA uptake in glial cells, and superfused with a buffer containing the aminotransferase inhibitor aminooxyacetic acid and the GABA uptake inhibitor nipecotic acid [76].

Fig. (2) The time course of [3H]5-HT release measured from raphe nuclei slices of the rat. Slices containing the raphe nuclei were prepared from rat brain, loaded with [3H]5-HT and superfused with Krebs-bicarbonate buffer. The slices were stimulated electrically (more ...)

The release of [³H]5-HT determined from raphe nuclei slices was characterized by the resting and the electrical stimulation-induced release and they were compared to those measured from hippocampal slices, an area containing axon terminal parts of serotonergic neurons. The resting release was 2.04±0.13 and 2.00±0.05% of content released in 3 min in the raphe and hippocampal slices, the electrically evoked release was 4.32±0.13 and 5.84±0.22% of content released in 3-min fraction in these brain areas, respectively. Although the resting and the evoked [³H]5-HT release did not differ in raphe nuclei and hippocampus the net fractional release (evoked release minus resting release) was 3.67±0.42 and 7.59±0.42% of content released per stimulus in the central and peripheral parts of serotonergic neurons. These findings indicate differences between the size of storage and release pools for 5-HT in somatodendritic and axon terminal fields.

Using the de Langer-Mulder plot, the size of the bound and the efflux compartments was calculated for [³H]5-HT in the raphe nuclei. This analysis revealed that at about 76.21± 3.19% of the total amount of [³H]5-HT taken up represent the efflux compartment and the remaining 23.77±3.19% contribute to the bound fraction. It can also be determined that at about 39% of radioactivity stored in the efflux compartment is released during a 66-min superfusion period. Moreover, compartmental analysis of the radioactivity stores in raphe tissue also indicates that the efflux compartment consists of various pools for [³H]5-HT in the raphe nuclei. Desaturation curve for each compartment can be determined from the de Langer-Mulder plot and the intercept of the regression line with the abscissa estimates the size of the pool that participates in the release. The involvement of different efflux compartments of [³H]5-HT in the release process was also shown as the efflux of [³H]5-HT was plotted in function of time (Fig. 3). As shown in Fig. (3), the efflux compartment can be divided into at least three different pools from which fast and slow efflux of [³H]5-HT occur in response to electrical stimulation and a third one which represents the source of resting outflow of [³H]5-HT.

Fig. (3) Release of [H]5-HT from raphe nuclei slices, a compartmental analysis. The fractional release of [3H]5-HT was plotted in function of superfusion time. The efflux of [3H] 5-HT occurs with different fractional rate constants indicating involvement of more (more ...)

Fig. (4) Effect of selective serotonin reuptake inhibitors on resting and electrically stimulated [3H]5-HT release in raphe nuclei slices of the rat. Fluoxetine (A) and citalopram (B) were added in 10 and 30 μM concentrations from fraction 9 and maintained (more ...)

Interaction between neuronal transporters and presynaptic release-mediating autoreceptors was already reported [99, 122]. 5-HT uptake inhibitors have been shown to antagonize the inhibitory effect of LSD on the electrically evoked overflow of [³H]5-HT from rat hypothalamic slices [65]. Moreover, 5-HT_1A autoreceptors are desensitized in the dorsal raphe nucleus of knockout mice lacking the 5-HT transporter [21]. Such interaction between autoreceptor and transporter may have importance in the mechanism on how L-deprenyl desensitizes 5-HT_1B receptors [77]. The close functional connection between neurotransmitter transporters and presynaptic autoreceptors is supported by the findings that repeated administration of 5-HT reuptake blockers alters sensitivity of release-mediating 5-HT₁ receptor subtypes in the somatodendritic area as well as in the axon terminals [17, 126]. In the dorsal raphe nucleus of the rat, the 5-HT_1A receptor partial agonist pindolol attenuated the inhibitory effect of fluoxetine on 5-HT cell firing [130]. Another piece of evidence for receptor-transporter functional interaction is shown by our finding that the 5-HT_2A/2C antagonist deramciclane potentiates the 5-HT uptake inhibitory effect of fluoxetine in raphe nuclei slices (unpublished observation). This finding is in line of previous suggestions that 5-HT₂ receptor antagonists may augment 5-HT uptake inhibitory properties of certain compounds[36].

Fig. (5) Effects of 8-OH-DPAT, CGS-12066 and 5-CT on electrical stimulation-induced [3H]5-HT release from raphe nuclei slices. The calculated IC50 values to inhibit [3H]5-HT release were 0.01, 0.28, and 0.58 μM for 5-CT, 8-OH-DPAT, and CGS-12066, respectively. (more ...)

Table 1 Effects of Various 5-HT Receptor Agonists on Electrical Stimulation-Induced [3H]5-HT Release from Raphe Nuclei and Hippocampal Slices of the Rat

5-Carboxamidotryptamine (5-CT), which has an agonist effect on 5-HT₁ receptor subtypes [32, 72, 171] and 5-HT₇ receptors [33, 85], also inhibited the electrically evoked [³H]5-HT release (Fig. 5) without influencing the resting [³H]5-HT outflow from raphe nuclei slices of the rat. This effect of 5-CT can be seen in both the somatodendritic and the axon terminal fields of raphe 5-HT neurons (Table 1). The calculated IC₅₀ value for 5-CT, concentration that induced 50% decrease in [³H]5-HT overflow in raphe nuclei slices, was 3.34±0.37 nM and the rank order of [₃H]5-HT release inhibition in raphe nuclei slices was found to be 5-CT > 8-OH-DPAT > CGS-12066 [80].

Table 2 Effects of Various 5-HT Receptor Agonists on Electrical Stimulation-Induced [3H]5-HT Release from Raphe Nuclei Slices of the Rat

The inhibitory effect of 8-OH-DPAT on [³H]5-HT release from [₃H]5-HT-preloaded raphe nuclei slices was also measured in the presence of (+)WAY-100135. Fig. (6) shows that (+)WAY-100135 dose-dependently antagonized the inhibitory effect of 8-OH-DPAT on [³H]5-HT release in raphe nuclei slices. WAY-100635 only slightly influenced the inhibitory effect of 5-CT on [³H]5-HT overflow and the apparent pA₂ value for WAY-100635 against 5-CT was 4.99 (Fig. 7). SB-216641, a selective 5-HT_1B receptor antagonist with approximately 25-fold selectivity over 5-HT_1D receptors [128], also produced a concentration-dependent rightward shift of the concentration-inhibition curve of 5-CT on [³H]5-HT release with no alterations in the maximal 5-CT response (Fig. 7). The calculated apparent pA₂ value for SB-216641 on 5-CT-induced inhibition of [³H]5-HT overflow was 7.12. This finding is in accord of the general view that 5-HT_1B and 5-HT_1D receptors are inhibitory autoreceptors in the 5-HT-containing raphe nuclei neurons [150] although different regulatory role of the two receptor subtypes was suggested [153].

Fig. (6) WAY-100635 antagonized the inhibitory effect of 8-OHDPAT on [3H]5-HT release but it was ineffective on CGS-12066 induced inhibition in raphe nuclei slices of the rat. For experimental procedure see Fig. (2). 8-OH-DPAT (1 μM) and CGS-12066 (1 μM) (more ...)

Fig. (7) Effects of WAY-100635 (10 μM) and SB-216641 (1 μM) on 5-CT-induced inhibition of[3H]5-HT release in raphenuclei slices of the rat. The IC50 value of 5-CT to inhibit [3H]5-HT overflow was 3.34±0.37 nM, it was 6.65±0.56 and (more ...)

Table 3 Effects of Various K+-Channel Inhibitors on 5-HT1 Receptor Agonists-Induced [3H]5-HT Release from Raphe Nuclei Slices of the Rat

When 8-OH-DPAT or CGS-12066 was combined with different K⁺ channel blockers, the inhibitory effect of 8-OHDPAT on [₃H]5-HT release was abolished in the presence of 4-AP, whereas this effect of CGS-12066 was not influenced by the K⁺ channel blocker (Table 3). Glibenclamide, on the other hand, failed to affect the inhibitory effect of either 8-OH-DPAT or CGS-12066 on [³H]5-HT release in raphe nuclei slices. This finding suggests that stimulation by 8-OHDPAT of somatodendritic 5-HT_1A autoreceptors opens voltage-sensitive K⁺ channels in serotonergic neurons and the consequent hyperpolarization of the somatodendritic membrane may then cause a decreased [₃H]5-HT release [9]. It is possible, however, that stimulation of 5-HT_1A and 5-HT_1B receptors leads to opening of different K⁺-channels as on the contrary to 5-HT_1A receptors, the voltage-sensitive K⁺ channels did not interfere with 5-HT_1B receptor-mediated 5-HT release inhibition. In contrast to 4-AP, the ATP-sensitive K⁺ channel inhibitor glibenclamide did not modify either 8-OHDPAT- or CGS-12066A-induced inhibition of [³H]5-HT release from raphe nuclei slices.Thus, the ATP-sensitive K⁺ channels are probably not involved in the 5-HT_1A and 5-HT_1B receptor-mediated autoinhibition of somatodendritic5-HTrelease[9].

Since autoreceptor-mediated feedback inhibition of neurotransmitter release originates from the vesicular stores but not from the cytoplasm [169] we have speculated that Ca²⁺availability may be a key factor for autoregulation of somatodendritic 5-HT release. The actual state of N-type somatodendritic Ca²⁺-channels substantially regulates neurotransmitter release process: closing of voltage-sensitive Ca²⁺ channels will lead to decrease of [Ca²⁺]i required for vesicular 5-HT release, whereas opening of Ca²⁺ channels may increase [Ca²⁺]i which then makes exocytotic release of 5-HT possible. It was reported that 5-HT_1A receptor stimulation results in reduction of high threshold Ca²⁺ current in current clamp studies of acutely isolated raphe neurons [123]. Accordingly, membrane hyperpolarization that occurs after 5-HT_1A receptor stimulation and K⁺-channel opening may be accompanied by closing voltage sensitive Ca channel. Thus, both opening of K⁺-channels and closing Ca²⁺channels may be involved in development of 5-HT_1A autoreceptor-mediated feedback inhibition of [₃H]5-HT release in raphe nuclei slices [78].

Others believe that 5-HT₂ receptors are probably not located presynaptically on serotonergic neurons [36, 46]. 5-HT₂ receptors may be present postsynaptically on GABA interneurons in the dorsal raphe nucleus. This conclusion was supported by the observation that 5-HT induces an increase in the frequency of inhibitory postsynaptic current of 5-HT cells by exciting GABAergic interneurons via 5-HT_2A/2C receptors [101]. 5-HT_2A/2C receptors located on GABA interneurons within the dorsal raphe nucleus mediate a local inhibitory feedback onto adjacent 5-HT neurons [101]. 5-HT₂ receptors located on cortico-raphe glutamatergic neurons may regulate GABA cell activity by trans-synaptic mechanisms in the raphe nuclei [106]. It was, however, also suggested that 5-HT_2A/2C might be present on glutamatergic axon terminals in the raphe nuclei [12]. Moreover, it was shown that cortical 5-HT₂ receptors exert stimulation of neural circuitry of the raphe nuclei by increasing glutamate release [6].

The activation state of postsynaptic 5-HT₂ depends on biophase concentration of 5-HT. Thus, stimulation of somatodendritic 5-HT_1A receptors decreases extracellular concentrations of 5-HT, which may then lead to an attenuated 5-HT₂ receptor-mediated effect in the raphe nuclei. The role of postsynaptic 5-HT₂ receptors in the local regulation of raphe circuitry was suggested by our findings that the 5-HT₂ receptor agonists DOI and mCPP stimulated [³H]glutamate release (Table 4). This stimulation was however, not associated with influences of 5-HT release, suggesting that the 5-HT₂ receptor-mediated serotonergic-glutamatergic neural interactions are not part of the reciprocal innervations shown between these two neurotransmitter systems [79].

Table 4 Effect of 5-HT2 Receptor Ligands on Electrical Stimulation-Induced on [3H]5-HT and [3H]Glutamate Release in Raphe Nuclei Slices of the Ra

Fig. (8) The stimulatory effect of 2-methyl-5-HT (0.1 to 10 Fmol/L) on (A) resting and (B) electrically stimulated [3H]5-HT release from raphe nuclei slices. Slices containing the raphe nuclei were prepared from rat brain, loaded with [3H]5-HT and superfused with (more ...)

2-Methyl-5-HT also increased the electrically evoked[³H]5-HT release from rat raphe nuclei slices preloaded with The 5-HT₃ receptor agonist 2-methyl-5-HT increased brain areas containing serotonergic axon terminals [29, 66]. [H]5-HT release from rat raphe nuclei slices preloaded with [³H]5-HT (Fig. 8B). We found that about one magnitude lower concentration of 2-methyl-5-HT elicited an increase in depolarization-induced 5-HT release than that was required for elevation of the resting release. The calculated EC₅₀ values of 2-methyl-5-HT to induce 50% increase in [³H]5-HT overflow above the control were 0.56 and 0.25 Imol/l in raphe nuclei slices and hippocampal slices, respectively. Galzin and coworkers [66] found that 2-methyl-5-HT was capable of increasing depolarization -induced [H]5-HT release in guinea-pig cerebral cortex already in concentrations, which were ineffective in alteration of basal outflow. These data present evidence for the existence of 5-HT₃ receptors in the rat raphe nuclei: these receptors stimulate the release of 5-HT from the somatodendritic part of serotonergic projection neurons. The in vitro stimulatory effect of 2-methyl-5-HT on electrically induced 5-HT release observed in our experiments may be consonant with previous in vivo findings: Martin and coworkers [107] reported that 2-methyl-5-HT increases 5-HT release from hippocampus and the 5-HT₃ receptor antagonist MDL 72222 antagonized this effect.

Ondansetron was not able to antagonize the stimulatory effect of 2-methyl-5-HT on resting [₃H]5-HT outflow in raphe nuclei slices (Fig. 8A). The lack of antagonistic effect of ondansetron on 2-methyl-5-HT-induced resting [₃H]5-HT release suggests that this effect is not a receptor-mediated event. On the contrary, ondansetron reversed the stimulatory effect of 2-methyl-5-HT on electrical stimulation-induced [³H]5-HT release (Fig. 8B) and the calculated pA₂ value for ondansetron was 6.51. The fact, that the stimulatory effect of 2-methyl-5-HT on depolarization-induced but not on spontaneous [³H]5-HT release was antagonized by ondansetron, suggests the involvement of 5-HT₃ receptors in the facilitatory effect of 2-methyl-5-HT on depolarization-induced 5-HT release.

Table 5 Effect of 2-methyl-5-HT on Resting [3H]5-HT Release in Raphe Nuclei Slices of Rat in Various Experimental Conditions

Other experiments also suggest the involvement of 5-HT transporter in the effect of 2-methyl-5-HT on basal [H]5-HT outflow. Thus, the 5-HT uptake inhibitor fluoxetine completely blocked the stimulatory effect of 2-methyl-5-HT on basal [³H]5-HT efflux (Table 5). 2-Methyl-5-HT may be a substrate for the 5-HT transporter promoting exchange diffusion by reversal of the carrier that transports 5-HT from the extracellular space into the cytoplasmic storage sites. It is less possible that 2-methyl-5-HT acts as blocker of 5-HT transporter and increases outflow of radioactivity by preventing uptake of released 5-HT. 2-Methyl-5-HT-altered 5-HT outflow was also blocked by the 5-HT uptake blocker paroxetine in guinea-pig hypothalamic slices [29]. The effect of 2-methyl-5-HT on [₃H]5-HT release was further investigated in raphe nuclei slices dissected from reserpine-pretreated rats. Reserpine pretreatment, which induces redistribution of [³H]5-HT from the vesicular pool to the favor of cytoplasmic compartments, did not alter the effect of 2-methyl-5-HT on [³H]5-HT release (Table 5).

The presented data suggest that receptor-coupled ion channels as well as voltage-sensitive Ca²⁺ channels may be involved in the 5-HT₃ receptor-mediated 5-HT release. When 5-HT₃ receptor is activated, the receptor-linked ion channel permeable to Na⁺, K⁺ and Ca²⁺ opens up and Na⁺ influx will lead to local membrane depolarization [52, 178]. The consequent increase in [Na⁺]i concentrations will result in operation of 5-HT transporters in reverse mode and 5-HT outflow from the cytoplasmic pool to the synaptic cleft will occur. In addition, membrane depolarization evokes opening of voltage-sensitive Ca²⁺ channels and the elevated Ca²⁺ influx will lead to exocytotic release of 5-HT from the vesicular pool. In fact, we have found that 2-methyl-5-HT increases both resting and electrical stimulation-induced [³H]5-HT release from raphe nuclei slices, releases that originate from cytoplasmic and vesicular transmitter pools, respectively [11].

Fig. (9) Effect of SB-258719 (10 μM) on 5-CT-induced inhibition of [H]5-HT release in raphe nuclei slices of the rat. The IC50 value of 5-CT to inhibit [3H]5-HT release was 3.34±0.37 nM and it was 94.23±4.84 nM in the presence of 10 μM (more ...)

5-CT inhibited the KCl-induced [³H]5-HT release similarly to those experiments when electrical stimulation was used to evoke [³H]5-HT release (Fig. 10). Addition of TTX prevented the inhibitory effect of 5-CT on K⁺-stimulated [³H]5-HT release from raphe nuclei slices suggesting that 5-HT₇ receptors may not serve as autoreceptors but act as heteroreceptors in controlling 5-HT release. We speculated that the 5-HT₇ receptor-mediated 5-HT release inhibition may involve operation of another, possibly excitatory neurotransmitter system. These postulated excitatory neurons may be glutamatergic and therefore, the regulatory role of 5-HT₇ receptors in glutamate release was further studied in the raphe nuclei slice preparation.

Fig. (10) Effect of TTX on 5-CT-induced inhibition of [H]5-HT release in raphe nuclei slices of the rat. Slices containing the raphe nuclei were prepared from rat brain, loaded with [3H]5-HT and superfused with Krebs-bicarbonate buffer in the presence and absence (more ...)

Fig. (11) The effects of 5-CT and 5-CT and SB-258719 on [3H]glutamate release from raphe nuclei slices of the rat. The slices were stimulated electrically (20 V, 2 Hz, 2-msec for 2 min) in fraction 10. 5-CT (0.1 μM) was added 9 min and SB-258719 (10 μM) (more ...)

Fig. (12) Effects of NMDA and AMPA on (A) [3H]5-HT and (B) [H]GABA release from raphe nuclei slices of the rat. The slices were superfused with NMDA (0.3 mM) or AMPA (0.3 mM) in fractions 10 and 11. When AMPA was used, cyclothiazide was also added in a concentration (more ...)

Fig. (13) A model for interaction between serotonergic and glutamatergic neurons in the raphe nuclei of the rat. 5-HT released from dendrites or recurrent axon collaterals of serotonergic neurons inhibits its own release by release-mediating 5-HT1A and 5-HT1B/1D (more ...)

Table 6 Effects of 5-HT and GABA Receptor Ligands on Electrical Stimulation-Induced [3H]5-HT and [3H]GABA Release from Raphe Nuclei Slices of the Rat

To determine whether 5-HT_1A and 5-HT_1B receptors influencing [³H]GABA release are located on GABA neurons or they are situated on other neuronal elements, the voltage-sensitive Na⁺-channel blocker TTX was used. We found that TTX did not block [₃H]GABA overflow in response to elevated KCl, a depolarizing stimulus that appears to act directly on axon terminals [152]). Moreover, 8-OH-DPAT decreased K⁺-induced [³H]GABA release and this inhibition was completely abolished by TTX [12]. Although CGS-12066 also inhibited [³H]GABA efflux elicited by KCl depolarization, this inhibition persisted even in the presence of TTX. Experiments carried out in slices treated with TTX shows that the location of 5-HT_1A and 5-HT_1B receptors involved in GABA release inhibition may be different and this was further evidenced in experiments carried out in 5-HTdeficient raphe slices.

5-HT stores in raphe nuclei slices were depleted by repeated administration of para-chlorophenylalanine (pCPA) to rats and the effects of 5-HT₁ receptor agonists on [³H]GABA release were measured [12]. We found that in 5-HT deficient slices, CGS-12066 decreased the electrically evoked [³H]GABA efflux whereas 8-OH-DPAT did not exert any inhibition on transmitter overflow (Table 7). Thus, the inhibitory effect of 5-HT_1B receptors on GABA release was independent from the tissue concentration of endogenous 5-HT whereas that of 5-HT_1A receptors was suspended after emptying 5-HT stores. This set of experiments suggests that 5-HT_1B receptors are located on GABAergic axon terminals and mediate a direct inhibition of GABA release whereas 5-HT_1A receptors are located on 5-HT cells and they exert an indirect inhibitory effect on GABA release.

Table 7 Effects of 5-HT and GABA Receptor Ligands on Electrical Stimulation-Induced [3H]5-HT and [3H]GABA Release from Para-Chlorophenylalanine (p-CPA)- or Isoniazid-Treated Rat Raphe Nuclei Slices

Our experiments indicate that both GABA_A and GABA_B receptors may be involved in the GABAergic-5-HTergic neural interaction in the raphe nuclei. The dual expression of GABA_A and GABA_B receptors on postsynaptic membranes has been demonstrated and even interactions at the level of GABA receptor subtypes were suggested [98]. The two GABA receptor subtypes may exert different types of inhibition on serotonergic neuronal activity as the GABA_B receptor antagonist phaclofen, on its own right, increased and the GABA_A receptor antagonist bicuculline did not affect somatodendritic 5-HT release. Tonic regulation of GABA_B receptors by the endogenously released GABA suggests a possible synaptic location of these receptors whereas a more phasic regulation by GABA_A receptors is in favor of non-synaptic influence on serotonergic neurons.

In a further attempt to localize GABA_A and GABA_B receptors regulating 5-HT release, raphe nuclei slices were taken from rats pretreated with the glutamic acid decarboxylase inhibitor isoniazid [12]. In GABA-deficient slices, muscimol and baclofen decreased the electrically evoked [H]5-HT efflux (Table 7). This finding indicates that 5-HT release-regulating GABA_A and GABA_B receptors may be situated postsynaptically on 5-HT neurons since depletion of releasable GABA by isoniazid failed to alter their inhibitory effects.

Fig. (14) A model for possible reciprocal interaction between serotonergic and GABAergic neurons in the raphe nuclei. 5-HT released from dendrites or recurrent axon collaterals of serotonergic neurons inhibits its own release by 5-HT1A and 5-HT1B/1D receptors. (more ...)

Several morphological and functional investigations indicate the occurrence of interaction between serotonergic and noradrenergic systems in the midbrain dorsal and median raphe nuclei. Early histological studies demonstrated noradrenergic afferent connections projecting to the raphe nuclei [138]. The dense noradrenergic innervation arises from the lower brainstem noradrenergic cell groups: retrograde tracing method labeled neurons in the locus coeruleus (A6 noradrenergic cell group), the A5 noradrenergic cell group, the caudoventrolateral medulla and the nucleus of the solitary tract (A1 and A2 noradrenegic cell groups) [124].

Neurochemical experiments showed the presence of norepinephrine (NE) and the enzymes involved in its synthesis within the raphe nuclei [137]. Autoradiographic studies have revealed the presence of alpha-1, alpha-2 and beta adrenoceptors in the dorsal raphe nucleus [148] and the presence of alpha-1 mRNA has also been demonstrated [50, 157]. The high level of mRNA encoding the alpha-1B receptors makes it likely that this adrenoceptor subtype is responsible for noradrenergic influence on serotonergic neurons. Alpha-1 heteroreceptors mediating noradrenergic control on serotonergic neurons may be located at the cell body levels.

Alpha-2 adrenoceptor binding has also been demonstrated in the raphe nuclei [50, 157]. Immuno-histochemical studies revealed the presence of alpha-2A and alpha-2C adrenoceptor immunoreactivities in the dorsal raphe nucleus [157]. Functional tests indicated that alpha-2A adrenoceptor subtype is involved in regulation of 5-HT release [133]. This was further supported by the finding that infusion of the alpha-2A adrenoceptor antagonist BRL 44408 into the dorsal raphe nucleus caused an increase of 5-HT release [129]. The majority of alpha-2 adrenoceptors may be located in noradrenergic axon terminals where they serve as autoreceptors and they regulate primarily NE release [37]. Others pointed out that alpha-2 adrenoceptors are involved in both NE and 5-HT release in rat dorsal raphe nucleus slices [84]. Guyenet and coworkers [68] concluded, however, that only a low percentage of 5-HT cells exhibit detectable alpha-2A adrenoceptor immunoreactivity in the dorsal raphe nucleus.

Fig. (15) Effects of clonidine, phenylephrine and 8-OH-DPAT on [3H]5-HT (A) and [H]NE (B) release from raphe nuclei slices of the rat. The calculated IC50 values to inhibit [3H]5-HT release were 50.26±9.27 and 98.62±3.16 nM for clonidine and 8-OH-DPAT, (more ...)

In vivo microdialysis studies also concluded that clonidine reduced and idazoxan or 2-methoxy-idazoxan enhanced 5-HT release in the rat raphe nuclei [3, 37]. These effects of alpha-2 ligands might be, however, indirect. Thus, idazoxan may enhance noradrenergic tone on 5-HT neurons through removal of the negative feedback inhibition of NE release, which then leads to activation of postsynaptic alpha-1 adrenoceptors located on 5-HT neurons. Supporting this view, it was found that the noradrenergic neurotoxin DSP-4 prevented the inhibitory effect of clonidine on 5-HT release suggesting an indirect effect on 5-HT output [37]. In raphe nuclei slice preparation, however, prazosin failed to suspend the 5-HT release-inhibitory effect of dexmedetomidine suggesting that alpha-2A heteroceptors may play a role in noradrenergic-serotonergic interaction [84].

The alpha-1 adrenoceptor antagonist prazosin (0.1 to 1 μM) had neither stimulatory nor inhibitory effect on [₃H]5-HT release from raphe nuclei slices. The alpha-1 adrenoceptor agonist phenylephrine inhibited electrical stimulation induced [³H]5-HT release (Fig. 15A) and this inhibition was associated by an enhanced basal 5-HT outflow. This effect of phenylephrine is surprising as this drug is widely used to increase single neuron firing rate [89]. The opposing effect of phenylephrine on 5-HT release may be a consequence of its dual electrophysiological effect: phenylephrine evokes depolarization and increases duration of after-hyperpolarization following action potential [120]. In another series of experiments, bath addition of phenylephrine also reduced 5-HT release assessed by voltammetry in raphe slices [84].

Although in vitro experiments provided some evidence for the inhibitory properties of alpha-1 adrenoceptors on somatodendritic 5-HT release, in vivo investigations led to opposite conclusion. Using microdialysis technique in freely moving rats, phenylephrine did not modify 5-HT output in the median raphe nucleus, whereas prazosin reduced the release of 5-HT [3]. This finding suggests that endogenous NE exerts a direct tonic stimulatory control on 5-HT release through alpha-1 adrenoceptors and these receptors may be maximally stimulated in resting conditions [129].

Phenylephrine, except at high concentration, did not influence electrical stimulation-induced [₃H]NE release from slices containing raphe nuclei indicating probably lack of alpha-1 adrenoceptors in the regulation of NE release (Fig. 15B). Moreover, prazosin (0.1 to 1 μM) was also without effect on [³H]NE release. Prazosin was also ineffective on NE release in a microdialysis experiments carried out with probes implanted into the dorsal raphe nuclei of conscious rats [37].

Regulation of 5-HT release by 5-HT_1A autoreceptors has been discussed in Section 3.1.1.

It has been shown that both A and B forms of monoamine oxidase (MAO) are present in serotonergic neurons, MAO-B being present in cell bodies of the raphe nuclei and MAO-A migrate to the nerve terminals [18]. Using antiserum directed against MAO-B enzyme led to the conclusion that this enzyme is specifically located in 5-HT-containing neurons and astrocytes in the rat brain [100]. MAO-B enzyme was preferentially detected not only in the rat but also in human raphe nuclei [140].

MAO inhibitors increase extracellular 5-HT concentrations in the raphe nuclei after an acute treatment [19, 40]. The large increase in 5-HT output in the raphe nuclei after systemic administration of MAO inhibitors triggers an activation of 5-HT_1A receptors by accumulation of an excess 5-HT in the extracellular space. Acute administration of nonselective MAO inhibitors (phenelzine, tranylcypromine) produces a decrease in 5-HT neuronal firing [25] and reduction of neural firing in the raphe nuclei leads to decreased 5-HT release from the nerve terminal fields [40]. Long-term administration of MAO inhibitors also leads to an elevation of extracellular 5-HT concentrations [41, 59] and repeated injections of MAO inhibitors induces desensitization of 5-somatodendritic HT_1A autoreceptors [25, 125]. This desensitization of 5-HT_1A autoreceptors is accompanied by recovery of firing of 5-HT neurons.

In axon terminal regions, the release of 5-HT is increased by chronic administration of MAO inhibitors and this effect seems to be independent from changes in somatodendritic 5-HT_1A receptor function [26]. This finding indicates that after long-term administration of MAO inhibitors, terminal autoreceptors are probably not desensitized. The MAO-A reversible inhibitor befloxatone enhanced the electrically induced release of [³H]5-HT after a 21-day treatment, whereas the effectiveness of the terminal 5-HT autoreceptor agonist 5-methoxytryptamine was not altered in the hypothalamus, hippocampus and frontal cortex [24].

The selective MAO-A inhibitor clorgyline induces similar changes as nonselective MAO inhibitors. Thus, clorgyline increases extracellular 5-HT concentration in the raphe area. Sustained increase of endogenous 5-HT in clorgyline-treated rats induces desensitization of somatodendritic 5-HT_1A autoreceptors [25-27]. Injection of clorgyline for three weeks failed to modify the function of the terminal counterpart as no down-regulation of 5-HT_1B receptors occurred. This finding shows that long-term clorgyline treatment also does not alter the function of axon terminal 5-HT autoreceptors.

Less data are available for the selective MAO-B inhibitor L-deprenyl, probably because 5-HT is preferentially deaminated by A form of MAO enzyme [179]. It has been reported that L-deprenyl when administered repeatedly, did not modify the firing activity of 5-HT neurons in the raphe nuclei [25, 26]. In addition, extracellular 5-HT concentrations in the raphe nuclei or in the frontal cortex were also not altered by doses of L-deprenyl selective for B form MAO enzyme inhibition [40]. However, after complete MAO-A inhibition by clorgyline, administration of L-deprenyl induced a large increase in dialysate 5-HT from the raphe nuclei pointing to the importance of MAO-B in the control of 5-HT release even when MAO-A is inhibited [40]. In our experiments, we have investigated the influence of MAO inhibition on the function of different 5-HT₁ receptor subtypes involved in regulation of 5-HT neurons in the raphe nuclei and found that L-deprenyl desensitizes 5-HT release-mediating 5-HT_1B autoreceptors in the rat raphe nuclei leaving 5-HT_1A receptor function unaltered.

Table 8 Effects of In Vitro Administration of L-Deprenyl and Clorgyline on 5-HT and 5-Hydroxyindoleacetic Acid (5-HIAA) in Raphe Nuclei Slices of the Rat

Whether selective MAO inhibitors influence 5-HT_1A and 5-HT_1B receptor sensitivity in 5-HT release inhibition, rat raphe nuclei slices were loaded with [³H]5-HT and super-fused with either clorgyline or L-deprenyl. The effects of the 5-HT_1A receptor agonist 8-OH-DPAT or the 5-HT_1B receptor agonist CGS-12066 were determined on electrically stimulated [₃H]5-HT release from clorgyline or L-deprenyl pretreated raphe nuclei slices.

Superperfusion of raphe nuclei slices with L-deprenyl reduced the inhibitory effect of CGS-12066 on electrically stimulated [₃H]5-HT release whereas the inhibitory effect of 8-OH-DPAT was still observed (Table 9). Clorgyline, on the other hand, did not alter the inhibitory effects of either 8-OH-DPAT or CGS-12066 on [³H]5-HT release. These data indicate that in vitro administration of L-deprenyl, in conditions where MAO-A enzyme is not inhibited, reduces sensitivity of 5-HT_1B autoreceptors to inhibit 5-HT release whereas the activity of 5-HT_1A remains unaltered. Neither Ldeprenyl nor clorgyline when added to superfused raphe nuclei slices influenced [³H]5-HT release.

Table 9 Effects of 5-HT1A and 5-HT1B Receptor Agonists on Electrical Stimulation-Induced [3H]5-HT Release in Raphe Nuclei Slices of the Rat After L-Deprenyl or Clorgyline Pretreatment In Vitro

The 5-HT_1A receptor agonist 8-OH-DPAT decreased the electrically induced [³H]5-HT release from superfused raphe nuclei slices of rats repeatedly injected with L-deprenyl (Table 10). The 5-HT_1B receptor agonist CGS-12066 did not influence, however, [₃H]5-HT release in raphe nuclei slices obtained from L-deprenyl-pretreated rats. In vivo treatment with L-deprenyl also reduced the inhibitory effect of CGS-12066 on [³H]5-HT release in hippocampal slices (Table 10). These data demonstrate that not only acute in vitro administration of L-deprenyl but also its repeated in vivo injections led to selective desensitization of somatodendritic and axon terminal 5-HT_1B autoreceptors. Moreover, loss of 5-HT_1B receptor sensitivity in 5-HT release inhibition was not accompanied by inhibition of MAO-A enzyme activity.

Table 10 Effects of 5-HT1A and 5-HT1B Receptor Agonists on Electrical Stimulation-Induced [3H]5-HT Release from Raphe Nuclei and Hippocampal Slices Dissected from Rats Repeatedly Treated with L-Deprenyl

There might be several possible explanations for the mechanism by which L-deprenyl selectively modifies 5-HT_1B autoreceptor functions. An interaction between inhibition of neuronal uptake and inhibition of neurotransmitter release through the effect of an agonist acting on autoreceptors was already reported [65, 99, 122]. It is possible to speculate therefore that the 5-HT uptake inhibitory effect of L-deprenyl [35, 74] may have a role in reducing the inhibitory effect of CGS-12066 on [³H]5-HT release. The effect of L-deprenyl-induced 5-HT uptake blockade on 5-HT_1B receptor-mediated [³H]5-HT release inhibition may indicate a close location of the 5-HT transporter to the autoreceptor organized in a serotonergic synapse. 5-HT transporter and 5-HT_1A autoreceptors, on the other hand, are probably not in a close link as somatodendritic 5-HT_1A receptors may have a preferential role in regulation of non-synaptic 5-HT release and the transporter may not be in a close vicinity of the receptor.

It has been shown that [H]imipramine binding site and terminal 5-HT autoreceptors are interactive and activation of this site leads to decrease autoreceptor function in regulation of 5-HT release [99]. Since L-deprenyl has been reported to increase the number of [³H]imipramine recognition site [181] the possibility arises that interaction exists between Ldeprenyl bound to 5-HT uptake site and activity of 5-HT_1B autoreceptors. The decreased function of 5-HT_1B autoreceptors in response to L-deprenyl-treated rats appears to result from a functional interaction between this type of release mediating autoreceptor and an effect on [H]imipramine binding sites.

Besides MAO inhibitors, 5-HT reuptake blockers when administered in a long-term, also exert profound effects on the sensitivity of the release-mediating 5-HT₁ autoreceptors in the somatodendritic area as well as in the axon terminals. It has been shown that sustained administration of paroxetine induces desensitization 5-HT_1A autoreceptors and also decreases effectiveness of 5-HT_1D receptor activation in reducing 5-HT release from rat raphe nuclei slices [125]. Desensitization of 5-HT_1D receptors in mesencephalic raphe and in hippocampus of the guinea pig may also occur after long-term paroxetine administration [55]. Repeated administration of 5-HT reuptake blockers leads to desensitization of terminal 5-HT_B/1D autoreceptors in different 5-HT projection areas and this was indicated by a reduced efficacy of 5-HT_1B/1D agonists to inhibit 5-HT release [126]. In another series of experiments, the 5-HT uptake inhibitor fluvoxamine, after sustained administration, was found to induce desensitization of somatodendritic 5-HT_1A receptors but not of 5-HT_1B/1D terminal autoreceptors [17]. Our experiments provide evidence that, in contrast to 5-HT reuptake inhibitors, L-deprenyl evokes desensitization of central and peripheral 5-HT_1B/1D receptors leaving the sensitivity of 5-HT_1A receptors unaltered.

In the raphe nuclei, 5-HT is released from serotonergic projection neurons as action potential propagation invades dendrites, recurrent axon collaterals or axon terminals. This release may occur from vesicles by exocytosis or from cytoplasmic pool with the mechanism of reverse-mode operation of neurotransmitter transporters. 5-HT release in the raphe nuclei may take place in synaptic organizations or directly into the biophase and thus, both synaptic and non-synaptic neurotransmission are present.

The release of 5-HT in the raphe nuclei is regulated by various 5-HT receptors. These receptors may have various locations in central part of serotonergic neurons: the releaseand impulse-mediating 5-HT_1A autoreceptors may be located on the somatodendritic part; the release-inhibitory 5-HT_1B/1D receptor may be on recurrent axon collaterals and axon terminals. The subtypes of 5-HT₁ receptors mediate feedback inhibition of 5-HT release. 5-HT₁ receptor subtypes can exert both synaptic and non-synaptic influences on 5-HT release. 5-HT_1B/1D receptors may be under tonic influence of 5-HT released whereas 5-HT_1A receptors may influence raphe 5-HT release with a more phasic mode.

The selectivity of 5-HT₁ receptor subtypes was differently altered by L-deprenyl, a selective inhibitor of B type MAO. Thus, acute in vitro and repeated in vivo administration of L-deprenyl reduced sensitivity of 5-HT_1B receptors in the raphe nuclei and the hippocampus without inducing changes in the sensitivity of somatodendritic 5-HT_1A receptors. This effect is probably not attributed to inhibition of MAO enzyme activity but rather to 5-HT uptake inhibitory effect of L-deprenyl. The inhibitory effects L-deprenyl on 5-HT carrier molecules and 5-HT_1B/1D receptors may indicate a close location of 5-HT transporter and autoreceptor organized in serotonergic synapses.

Evidence was presented for the involvement of the ionotropic 5-HT₃ receptors in mediation of raphe 5-HT release. 5-HT₃ receptors stimulate exocytotic vesicular release of 5-HT and that mediated by reverse mode operation of 5-HT transporter. The release-facilitatory 5-HT₃ receptors may be sited on the proximal dendrites, dendritic shaft or on cell body with an extrasynaptic organization. 5-HT₃ receptors mediate feedback stimulation of raphe 5-HT release.

There may be an integrated operation of 5-HT receptors in the regulation of 5-HT release in the raphe nuclei (Fig. 16). As 5-HT release occurs, the subtypes of 5-HT₁ autoreceptors in the dendrites, axon collaterals and nerve endings are activated leading to inhibition the further release of 5-HT. If the autoreceptor-mediated inhibition is attenuated (autoreceptor is preoccupied by 5-HT or 5-HT transporter is down-regulated) or initiated with a delay, the concentration of 5-HT in the gap of dendritic synapse may reach high enough concentrations to diffuse out into the extrasynaptic space. 5-HT traveling a considerable distance in the biophase may activate 5-HT₃ receptors located extrasynaptically and as a consequence, changes in electrical properties of dendritic membranes will occur. 5-HT₃ receptors may be located more proximally on dendritic tree and therefore, depolarization of the membranes will not allow propagation of inhibitory inputs from distal dendrites to the dendritic shaft and cell body. GABA may be a possible candidate for this kind of inhibition as GABAergic neurons in the raphe nuclei are known to synapse with serotonergic neurons [62]).

Fig. (16) Possible neuronal interaction in the raphe nuclei. 5-HT may be released into synapses formed between vesicle-containing dendrites of serotonergic neurons and impinging neurons. 5-HT release in dendritic synapses is inhibited by presynaptic 5-HT1B/1D autoreceptors. (more ...)

Another possible consequence of 5-HT₃ receptor stimulation in the raphe nuclei is an increase of 5-HT release originated either from the vesicular or the cytoplasmic pools or both of serotonergic dendrites. The facilitation of 5-HT release by non-synaptic 5-HT₃ receptors will result in high 5-HT concentrations in the extrasynaptic space leading to transmitter diffusion toward 5-HT-containing dendrites and recurrent axon collaterals equipped with inhibitory 5-HT_1A and 5-HT_B/D autoreceptors. These dendrites or nerve endings will then be hyperpolarized and not be invaded by action potential and 5-HT release will be inhibited. Therefore, facilitation of 5-HT release by extrasynaptic 5-HT₃ receptors may induce release-inhibition from a large number of dendrites or recurrent axon collaterals (Fig. 16). The facilitatory 5-HT₃ and inhibitory 5-HT₁ receptor subtypes exert opposing role in the control of somatodendritic 5-HT release. Thus, there may be an interaction between inhibitory 5-HT₁ and facilitatory 5-HT₃ receptors in the regulation of somatodendritic 5-HT release in the raphe nuclei.

The local neuronal circuit in the raphe nuclei consists of serotonergic projection neurons, GABAergic interneurons and glutamatergic axon terminals. 5-HT inhibits GABA release in synaptic or non-synaptic communications and both 5-HT_1A and 5-HT_1B/1D receptors are involved in this inhibition. The location of the two 5-HT₁ receptor subtypes may be, however, different: 5-HT_1B/1D heteroreceptors are probably located on GABAergic axon terminals and they exert a direct inhibition on GABA release. 5-HT_1A receptors that inhibit GABA release are probably those located on 5-HT cells and mediate autoinhibition of 5-HT release. Thus, inhibition of raphe GABA release by 5-HT_1A receptors is due to an indirect effect.

GABA, if once released from axon terminals of interneurons or projection neurons, inhibits 5-HT release by stimulating GABA_A and GABA_B receptors located on serotonergic projection neurons. GABA_B receptors may mediate a tonic inhibition whereas GABA_A receptors exert a more phasic inhibition on 5-HT release. These data indicate a reciprocal interaction between serotonergic and GABAergic neurotransmission in the raphe nuclei. Moreover, reciprocal innervation may exist between serotonergic and glutamatergic nerve endings; 5-HT released into the biophase may inhibit glutamate release either in a synaptic or non-synaptic communication by 5-HT₇ heteroreceptors located on glutamatergic axon terminals.

Activation of 5-HT₇ heteroreceptors on glutamatergic axon terminals by 5-HT inhibits glutamate release in the raphe nuclei. This inhibition reduces excitatory influence to serotonergic neurons, which then leads to inhibition of 5-HT release. 5-HT released from projection neurons exerts dual inhibitory and excitatory influences on glutamate release; these effects may be mediated by 5-HT₇ and 5-HT₂ receptors expressed on glutamatergic nerve endings. The postulated glutamatergic-serotonergic interaction in the raphe nuclei was further evidenced by the finding that NMDA and AMPA, agonists for ionotropic glutamatergic receptors, enhance 5-HT release in raphe nuclei slices.

Within the raphe nuclei, serotonergic projection neurons and axon terminals of noradrenergic neurons from the locus coeruleus are in reciprocal interactions at the somatodendritic and axon terminal levels. The noradrenergic influence of serotonin cells is mediated by pre- and postsynaptic alpha-2 and alpha-1 heteroceptors, which regulate cell firing as well as 5-HT and NE release. Serotonergic-noradrenergic interplay in the raphe nuclei is an important target for a number of psychoactive drugs, including selective and nonselective serotonin reuptake inhibitors, MAO inhibitors and atypical antipsychotic drugs [31]. Several clinical observations conclude that neuronal network of the raphe nuclei mediates a number of psychiatric disorders. Ascending serotonergic pathways originating from the raphe nuclei and projecting to the amygdala and frontal cortex or to the dorsal periaqueductal gray matter have been implicated in the pathology of generalized anxiety disorder and panic disorder, respectively [127]. Of the various 5-HT receptors, 5-HT_1A and 5-HT_2A receptors were suggested to be involved in development of various anxietic disorders. Evidence for reduced serotonergic input from the dorsal raphe nuclei to the prefrontal cortex has been reported in patients suffering from major depression or attempted suicide [34, 96]. We believe that detailed investigation of neural network and neurotransmitter interactions within the raphe nuclei may lead not only to better understanding drugs’ mode of action but also to discover clinically potent new compounds for treatment of psychiatric disorders.

This research was supported in part by the Research Council for Health Sciences, Hungarian Ministry of Health and Welfare (ETT-482/2003) and the Hungarian Science Research Fund (OTKA, grant No T-43511). The author acknowledges the excellent technical assistance of Mrs. Erika Hajdune-Gosi and Zsuzsa Major and the editorial work of Ms. Judit Puskas.