The Selective Serotonin Reuptake Inhibitor Citalopram Induces the Storage of Serotonin in Catecholaminergic Terminals

Materials and Methods

Male New Zealand rabbits (2–3.5 kg) were sacrificed by decapitation. The brains were quickly removed, and olfactory tubercles were immediately dissected on ice in a cold room at 4°C, following these landmarks: diagonal band of Broca (caudal-medial border), beginning of the olfactory tract (rostral border), lateral olfactory tract (lateral border), and medial forebrain bundle (dorsal border). Next, the tissues were chopped into 0.4-mm slices by means of a McIlwain tissue chopper (Mickle Laboratory Engineering Co., Ltd., Surrey, UK). Approximately, from 20 to 30 slices were obtained from both olfactory tubercles of one rabbit.

Uptake of [³H]5-HT.

Olfactory tubercle slices were incubated for 30 min at 37°C with 0.1 μM [³H]5-HT in 4 ml of superfusion medium. The medium composition was 118 mM NaCl, 4.8 mM KCl, 1.3 mM CaCl₂, 1.2 mM MgSO₄, 25 mM NaHCO₃, 1.2 mM KH₂PO₄, 0.03 mM disodium EDTA, 0.57 mM ascorbic acid, and 11 mM glucose. The medium was continuously gassed with 95% O₂ and 5% CO₂, and the pH was set at 7.4. The slices were incubated with [³H]5-HT either in the presence or in the absence of the 5-HT-uptake pump inhibitor citalopram (1 or 10 μM) and/or the NE/DA uptake pump inhibitor nomifensine (1 or 10 μM). Subsequently, the slices were washed three times with 4 ml of cold (4°C) superfusion medium, and the radioactivity present in the slices was quantified by liquid scintillation spectrometry.

Displacement of Stored [³H]5-HT.

Immediately after incubation, two olfactory tubercle slices were transferred to a nylon mesh mounted on a polypropylene cylinder. The cylinders were then placed in 1-ml glass superfusion chambers. The slices were superfused with prewarmed oxygenated superfusion medium (37°C) at 1 ml/min. Fraction collection (5-min fractions) started after a 75-min washout period with drug-free medium. After obtaining a basal efflux of radioactivity for 20 min, slices were superfused with stepwise increasing concentrations (from 1 nM to 30 μM) of either 5-HT, DA, or NE at sequential intervals of 20 min. In a group of experiments, the effects of 5-HT, NE, or DA were studied in the presence of either 1 μM citalopram or 1 μM nomifensine. In these experiments, perfusion with citalopram or nomifensine started 30 min before collection of baseline samples (50 min before perfusion with monoamines) and continued until the end of the experiment. Radioactivity in the superfusate-fractions (efflux) and in the slices (at the end of the experiment) was quantified by liquid scintillation spectrometry. Because of differences in the extent of labeling observed between control slices and those labeled in the presence of citalopram, the efflux results were always corrected by the amount of radioactivity present in the slices. The efflux results were expressed as fractional release, i.e., as a percentage of the tissue ³H content. The radioactivity (total ³H) collected in the superfusion samples consisted of a mixture of unchanged [³H]5-HT and their respective³H-metabolites (Fozard and Berry, 1976). In this study, for convenience, the collected [³H] efflux was simply referred to as displacement of [³H]5-HT.

Statistics.

Inhibition of [³H]5-HT uptake by citalopram and/or nomifensine was expressed in absolute values (disintegrations per minute per milligram of tissue) and as a percentage of inhibition employing as control values the tissue content obtained in the absence of uptake inhibitors (controls). The efflux of [³H]5-HT was expressed in absolute values (dpm/5 min) and as a fraction of the amount of radioactivity present in the tissue (slices) at the time of sample collection. One-way analysis of variance (ANOVA) was used to establish overall effect of uptake pump inhibitors on uptake of [³H]5-HT, using different treatments (concentrations and combinations) as the levels. Repeated measurement one-way ANOVA was used to determine overall ability of increasing concentrations of nonradioactive monoamines to displace [³H]5-HT, serving monoamine concentrations as a repeated factor. Two-way ANOVA for a repeated measurement design was used to determined differences in [³H]5-HT displacement induced by nonradioactive monoamines in the presence or absence of uptake pump inhibitors, with uptake condition (with or without inhibition) as the cross-factor and monoamine concentrations as the repeated measure factor. In cases where the two-way ANOVA revealed significant interaction between the two factors, data were again analyzed with repeated measurement one-way ANOVA. Statistically significant effects detected by ANOVA were further analyzed using Duncan's multiple range test. Significance was assumed at P < 0.05.

Results

Effect of Citalopram and Nomifensine on [³H]5-HT Uptake.

In slices of the olfactory tubercle, the SSRI citalopram (1–10 μM) inhibited the uptake of 0.1 μM [³H]5-HT to a maximum of 71 ± 2% (Table1). Comparable degrees of uptake inhibition were observed with 1 and 10 μM citalopram (Table 1). However, nomifensine alone (1–10 μM) weakly inhibited [³H]5-HT uptake. Complete inhibition of [³H]5-HT uptake (99.0 ± 0.01%) was achieved with a combination of 1 μM citalopram and 10 μM nomifensine (Table 1).

Release of [³H]5-HT from Superfused Slices Labeled under Control Conditions (Absence of Uptake Inhibitor).

In control slices preloaded with 0.1 μM [³H]5-HT (with no citalopram in the incubation mixture), superfusion with low concentrations of nonradioactive 5-HT (0.03–1 μM) increased the efflux of ³H (release) in a concentration-dependent manner (P = 0.0001) (Fig.1). Nearly 8% (7.8 ± 0.5%) of tissue radioactivity was released by 0.1 μM 5-HT, and radioactivity increased to 34 ± 2% with 1 μM 5-HT (Fig. 1A). Neither DA nor NE at low concentrations (0.03–1 μM) evoked the release of [³H]5-HT (Fig. 1A). The increase in³H efflux induced by 5-HT was greatly antagonized by coperfusion with 1 μM citalopram (P = 0.0001) (Fig. 2). These results suggest that when the olfactory tubercle slices are incubated with [³H]5-HT in the absence of any drugs (controls), the radioactive amine most likely labels the 5-HT stores.

Release of [³H]5-HT from Slices Labeled with [³H]5-HT in the Presence of the SSRI, Citalopram.

In this set of experiments, the slices were incubated with 0.1 μM [³H]5-HT in the presence of citalopram (1 μM). The objective was to determine whether 5-HT could be stored in other monoaminergic terminals when its neuronal uptake transport was blocked with citalopram. After the labeling period, citalopram was removed by extensive washout during the continuous superfusion of the slices. Once the basal efflux had leveled off to a monoexponential decay rate, sample collection was initiated. Compared with control slices, low concentrations of nonradioactive 5-HT (0.03–0.1 μM) in slices labeled in the presence of citalopram were no longer able to evoke the release of [³H]5-HT. Higher concentrations of 5-HT (0.3–10 μM) were now required to induce significant displacement of [³H]5-HT (Fig.3). Indeed, in control slices, 0.1 μM 5-HT released 7.8 ± 0.5% and 1.0 μM released 34 ± 2% of tissue radioactivity. In slices labeled in the presence of citalopram, 0.1 μM 5-HT failed to evoke release, and 1.0 μM released only 5.5 ± 0.6% of tissue radioactivity (P < 0.001) (compare Fig. 1, A and B).

In slices labeled in the presence of citalopram, superfusion with NE or DA at concentrations as low as 0.3 and 1 μM evoked significant increases in the release of [³H]5-HT (Fig. 1B). However, in control experiments (labeling without citalopram), 1 and 3 μM DA and NE failed to evoke the release of [³H]5-HT (Fig. 1A). In addition, in slices labeled in the presence of citalopram, when citalopram was coperfused with DA or NE, it potentiated the releasing effects of both catecholamines (Fig. 3). Both DA and NE (3–10 μM) evoked a much greater release of [³H]5-HT when perfused in the presence of citalopram (P = 0.0001) (Fig. 3). On the other hand, when citalopram was coperfused with 5-HT, it induced a very small parallel shift to the right in the concentration-release curve for 5-HT (P = 0.0593) (Fig. 3).

In slices labeled in the presence of citalopram, nomifensine reduced NE, DA, and 5-HT-evoked release of [³H]5-HT (Fig. 4). Nomifensine antagonized the releasing effect of NE to a greater extent than that of DA (Fig. 4). Although for both catecholamines, nomifensine seemed to induced a parallel shift to the right, the release evoked by 5-HT was decreased, but only at the highest concentrations, 3 to 10 μM (P= 0.0006) (Fig. 4).

Discussion

Incubation of control rabbit olfactory tubercle slices with concentrations of [³H]5-HT (0.1 μM) close to the K_m for the neuronal uptake (60–180 nM) (Richelson and Pfenning, 1984; O'Reilly and Reith, 1988) , led to selective labeling of 5-HT terminals. This statement is based on the following experimental facts (present study): 1) only submicromolar concentrations of 5-HT, but not DA or NE, displaced radioactivity from slices preloaded with [³H]5-HT, 2) 5-HT-evoked release of [³H]5-HT was markedly antagonized by 1 μM citalopram, and 3) nomifensine (a preferential NE/DA uptake inhibitor) had negligible (10–15%) effects in inhibiting 5-HT uptake. Our conclusion is based on the specificity of the drugs employed and on the premise that 5-HT and catecholamine neuronal pumps do not take up 5-HT metabolites (Caccia et al., 1992; Manfridi et al., 1992; Wong et al., 1995). Thus, the tissue labeling and the displacement of³H from the slices most likely derive from unchanged [³H]5-HT. The concentration of 1 μM citalopram employed in this study is 1000-fold higher than itsK_i for 5-HT uptake, and 4 and 28 times lower than its K_i for NE and DA uptake, respectively (Richelson and Pfenning, 1984; Sitte et al., 2000), indicating that citalopram is selective for the 5-HT transporter. Consequently, 1 μM citalopram should have blocked 5-HT uptake with negligible effects on neuronal catecholamine transporters. Nomifensine, on the other hand, selectively inhibits NE and DA transporters. Its small inhibitory effect on [³H]5-HT accumulation may result from a slight effect on 5-HT transport, since nomifensineK_i for 5-HT uptake is about 1.2 μM (Richelson and Pfenning, 1984).

A striking finding of this study was the inability of high micromolar concentrations of citalopram to produce a complete block of [³H]5-HT uptake. Olfactory tubercle slices incubated in the presence of concentrations of citalopram as high as 10 μM could still take up and retain considerable amounts of [³H]5-HT (30–40% of the amount of radioactivity taken up by control slices). The concentration of citalopram used should have been sufficient to block completely the neuronal 5-HT transporter, since it was 5,000- to 10,000-fold higher than the K_i for the 5-HT uptake (Richelson and Pfenning, 1984; O'Reilly and Reith, 1988). The lack of complete inhibition of [³H]5-HT uptake by citalopram suggests that when the 5-HT uptake is blocked, [³H]5-HT may be labeling non-5-HT terminals. This is supported by the observation that addition of nomifensine (combined treatment with 1 μM citalopram and 1–10 μM nomifensine) completely blocked the labeling of the slices. Because nomifensine preferentially inhibits NE and DA transporters, our findings suggest that the labeling occurring when [³H]5-HT is incubated in the presence of high concentrations of citalopram may be due to storage of [³H]5-HT in catecholaminergic terminals. We thus propose that when the 5-HT neuronal transporter (highest affinity site for 5-HT) is blocked, 5-HT may accumulate in sites of lower affinity, such as the DA and NE terminals.

Further support for this view derives from experiments on the displacement of stored [³H]5-HT by nonradioactive monoamines. As previously discussed, when the labeling of the rabbit olfactory tubercle slices with radioactive 5-HT is conducted in the absence of drugs or treatments (controls), tissue [³H]5-HT is selectively displaced by submicromolar concentrations of nonradioactive 5-HT. However, in slices preloaded with [³H]5-HT in the presence of citalopram, the radioactivity retained by the slices (after extensive washout) is released only by higher concentrations of 5-HT. In addition, concentrations of DA and NE that failed to evoke [³H]5-HT efflux from control slices were effective in evoking [³H]5-HT efflux from slices labeled in the presence of citalopram. The marked reduction of the ability of 5-HT to displace [³H]5-HT, associated with a potentiation of the [³H]5-HT releasing action of DA and NE, further suggests that when the access of 5-HT to its nerve terminals is hindered by an SSRI, 5-HT may be stored in catecholaminergic terminals.

In slices preloaded in the presence of citalopram, the increase in efflux of [³H]5-HT induced by DA and NE was not inhibited but enhanced by coperfusion with citalopram. This finding suggests that when the slices are labeled in the presence of citalopram, DA and NE release [³H]5-HT from catecholaminergic terminals and not from 5-HT nerve terminals. If DA and NE had released [³H]5-HT from 5-HT terminals, citalopram would have inhibited the increase in³H efflux induced by the catecholamines. The fact that coperfusion with citalopram enhanced [³H]5-HT evoked by DA and NE indicates that part of the [³H]5-HT released by DA and NE is taken up by the 5-HT terminals (i.e., [³H]5-HT efflux is less than actual [³H]5-HT release). Therefore, the increase in [³H]5-HT efflux induced by DA and NE in the presence of citalopram best reflects the amount of [³H]5-HT released from catecholaminergic terminals by these monoamines. Under these conditions, NE and DA are 3 to 10 times more potent than 5-HT in evoking [³H]5-HT release, further supporting the view that DA and NE were releasing [³H]5-HT from catecholaminergic terminals.

Experiments employing cold 5-HT as a releasing agent indicate that the efflux of [³H]5-HT evoked by micromolar concentrations of cold 5-HT results from two effects: 1) displacement of stored [³H]5-HT and 2) inhibition of [³H]5-HT reuptake into 5-HT nerve terminals by competition with cold 5-HT. Therefore, even if cold 5-HT has a lower affinity than DA and NE to release [³H]5-HT from catecholaminergic terminals, any [³H]5-HT released by cold 5-HT would contribute to the efflux of radioactivity because it would not be taken up by the 5-HT terminals. This dual action of cold 5-HT accounts for the negligible effects of citalopram on [³H]5-HT efflux evoked by cold 5-HT, since both citalopram and cold 5-HT would competitively inhibit the reuptake of [³H]5-HT.

The observation that nomifensine, an inhibitor of the catecholamine neuronal transporter, antagonized the increase in efflux of [³H]5-HT evoked by DA, NE, and 5-HT, suggests that the three monoamines were releasing [³H]5-HT from catecholaminergic terminals. All the above-described findings support the view that incubation of olfactory tubercle slices with [³H]5-HT in the presence of citalopram leads to accumulation of [³H]5-HT in catecholaminergic neurons.

Information, albeit indirect, about a possible preferential accumulation of [³H]5-HT in NE neurons derives from experiments conducted in the presence of nomifensine. In slices preloaded with [³H]5-HT in the presence of citalopram, nomifensine produced a much greater antagonism of the releasing effect of NE than of DA. These results may in part be explained by the greater affinity of nomifensine for NE than for the DA transporter. In fact, the K_i of nomifensine to inhibit NE uptake is 10 times lower than itsK_i for inhibition of DA uptake (Richelson and Pfenning, 1984). However, it is also possible that differences in density and spatial distribution between dopaminergic and noradrenergic neurons in relationship to the serotonergic neurons may also account for the findings. A closer anatomical association between noradrenergic and serotonergic terminals, than of dopaminergic with serotonergic terminals, may also account for the greater effects of nomifensine on NE than on DA.

In summary, we propose that when the 5-HT uptake pump is blocked with citalopram, extracellular [³H]5-HT is shunted into NE and possibly DA terminals, where it could exert multiple actions. 5-HT may by competition inhibit DA and NE uptake and metabolism, and be stored in synaptic vesicles from which it could be coreleased with the catecholamines, acting as a false neurotransmitter. In agreement with our findings, the in vivo microdialysis application of clomipramine, a 5-HT uptake inhibitor, increases DA efflux from the striatum (Santiago et al., 1998). This releasing effect was inhibited by nomifensine, suggesting that when 5-HT uptake is blocked, 5-HT may be shunted to DA neurons inducing DA release (Santiago et al., 1998). 5-HT release from DA neurons has also been demonstrated in the intermediate lobe of the rat pituitary by a classical calcium-dependent exocytotic mechanism (Vanhatalo and Soinila, 1995). Notably, the release of 5-HT as a false neurotransmitter can have functional postsynaptic consequences. For example, canine coronary adrenergic nerves can take up 5-HT secreted from platelets and subsequently release it to cause contraction of the coronary arteries (Palkovits et al., 1986). In the central nervous system, chronic selective blockade of 5-HT uptake increases functional and biochemical DA parameters in rats, such as DA agonist-induced psychomotor response (Maj, 1990) and aggressiveness (Maj et al., 1979), DA receptor number, and DA agonist efficacy in the DA mesolimbic system (Serra et al., 1992). Although these findings may be explained by other mechanisms, the possibility that some of these effects may be related to accumulation of 5-HT in DA neurons and false neurotransmission cannot be ruled out. For example, compensatory changes in catecholaminergic neurons, especially up-regulation of postsynaptic DA and/or NE receptors, may occur if 5-HT instead of DA or NE is released from DA and/or NE neurons. If so, such an effect may account for some of the mood-elevating effects of antidepressants. The mild amphetamine-like effects of some SSRIs may be due to sensitization of the DA system induced by prolonged false neurotransmission and/or by increased extracellular levels of DA and NE due to release and uptake inhibition by shunted 5-HT. In fact, acute inhibition of DA uptake in superfused rat striatal synaptosomes by amineptine increases the overflow of DA, 5-HT, and NE (Garattini and Mennini, 1989), suggesting that selective blockade of catecholamine uptake may cause a similar interaction with other monoamine terminals.