Introduction

Initial first-line pharmacotherapy for major depressive disorder (MDD) is associated with a low remission rate [1]. One strategy of optimizing treatment is by the addition of dopamine (DA)/serotonin (5-HT) antagonists and/or receptor partial agonists (previously known as atypical antipsychotics [2]) to a variety of first-line medications including selective 5-HT reuptake inhibitors (SSRIs) and 5-HT-norepinephrine (NE) reuptake inhibitors (SNRIs) at regimens below their antipsychotic range. Of the adjunctive medications, addition of low-dose aripiprazole has consistently been reported to produce a robust antidepressant effect in treatment-resistant depressed (TRD) patients [3].

As aripiprazole binds to several DA and 5-HT receptors, it could interact through actions on one or several of these receptors. For instance, by blocking 5-HT2C and 5-HT2A receptors, aripiprazole addition to escitalopram normalized escitalopram-induced inhibition of firing activity of, respectively, ventral tegmental area (VTA) DA and locus coeruleus NE neurons [4]. In addition, aripiprazole combination with escitalopram rapidly normalized firing activity of dorsal raphe nucleus (DRN) 5-HT neurons due to desensitization of 5-HT1A autoreceptors and an agonist action on their excitatory D2 receptors [4,5,6]. Furthermore, the combination of aripiprazole and escitalopram enhanced tonic activation of 5-HT1A receptors in the rat hippocampus but had no effect by themselves on this parameter [7]. Despite aripiprazole possessing the highest affinity for the 5-HT2B receptor among all receptors, where it exerts an antagonist activity [8, 9], the contribution of this activity to the augmenting effect of aripiprazole when added to 5-HT reuptake inhibition has never been investigated. Interestingly, other adjunct medications (olanzapine, quetiapine, and ziprasidone) also have higher affinity for the 5-HT2B receptor relative to the D2 receptor [9], which strengthens the notion that the blockade of these receptors may contribute to their antidepressant effect.

5-HT2B receptors are expressed in brain regions that are relevant to depression and may provide an additional target to induce an antidepressant response (see [10] for a review). In the rat DRN, these receptors were found on GABA neurons [11, 12] and in the mouse VTA, about 40% of DA neurons express 5-HT2B receptor mRNA [13]. These receptors are also found in the septum, the amygdala, and on pyramidal neurons in the frontal cortex [14, 15]. 5-HT2B receptors exert a role in the antidepressant-like response in mice [16,17,18,19]. Indeed, knocking out the 5-HT2B receptor in mice produces an antidepressant-like phenotype, such that at baseline, these mice exhibit a reduced latency to feed in the novelty-suppressed feeding task, an increased sucrose consumption in the sucrose preference test, and they express increased BDNF mRNA density and protein levels in the hippocampus [18]. Furthermore, such genetic ablation of the mouse 5-HT2B receptor attenuates the effect of the SSRI fluoxetine in the forced-swim test [18].

The present study investigated whether pharmacological blockade of 5-HT2B receptors have a role in the antidepressant effect when aripiprazole is added to an SSRI. It determined whether blockade of 5-HT2B receptors by aripiprazole or the 5-HT2B antagonist LY266097, in addition to inhibition of the serotonin transporter (5-HTT) by escitalopram in vivo, induces additional change in monoaminergic activity of VTA DA neurons and DRN 5-HT neurons, as well as glutamatergic pyramidal neurons in the medial prefrontal cortex (mPFC). Furthermore, a previous study showed that activation of the 5-HT2B receptor decreases the capacity of SSRIs to bind to the 5-HTT in vitro [20]. Since combination of aripiprazole and escitalopram bind, respectively, to 5-HT2B receptors and 5-HTT, their in vivo interaction was examined in the hippocampus.

Materials and methods

Animals

Male Sprague-Dawley rats (Charles River, St. Constant, Canada) weighing 250–350 g were housed under standard laboratory conditions (12 h light/dark cycle), with access to food and water ad libitum. Experiments were performed between 10:00 and 18:00. In vivo extracellular recordings were carried out in chloral hydrate anesthetized rats (400 mg/kg, intraperitoneally [i.p.]) preceding fixation of the rodent into the stereotaxic apparatus. A single-barrel glass micropipette (Stoelting, Wood Dale, IL, USA) preloaded with a 2 M sodium chloride solution was used to record in the VTA, DRN, and mPFC and a 5-barrel micropipette (ASI Instruments, Warren, MI, USA) preloaded with the same solution in the hippocampus. Body temperature was maintained at 37 °C throughout the experiment via a water-based heating pad. A catheter was inserted into the lateral tail vein for systemic intravenous (i.v.) injection of pharmacologic agents. All animals were handled according to the Canadian Council on Animal Care guidelines, and all protocols of this study were approved by the local Animal Care Committee (University of Ottawa Institute of Mental Health Research, Ottawa, Canada).

Minipump implantation

Escitalopram was administered subcutaneously (s.c.) using implanted Alzet minipumps. For implantation, after the rat was anesthetized with isoflurane 2–4%, the fur was shaved, and the skin was washed with alcohol and betadine. An incision and a pocket were made and a filled minipump was inserted and the wound was closed with metal clips left in place no longer than 7 days. Minipumps were implanted for 2 and 14 days, and electrophysiological experiments were carried out while minipumps were still implanted. Two different types of minipump were used: model 1003D for 2-day and model 2ML2 for 14-day administration.

Experimental groups and treatments

Acute experiments

In the VTA, the preferential 5-HT2B receptor agonist BW723c86 (pKi 5-HT2A: 6.6, pKi 5-HT2B: 7.9, pKi 5-HT2C: 6.3 [10, 21]) and the selective 5-HT2B receptor antagonist RS127445 (pKi 5-HT2A: 6.3, pKi 5-HT2B: 9.5, pKi 5-HT2C: 6.4 [10, 22] were administered i.v. and s.c., respectively. Doses used were based on previous studies: BW723c86, 1–6 mg/kg, [21] and RS127445, 2 mg/kg, [23, 24]. In the mPFC, a dose of aripiprazole 0.6 mg/kg was used based on a previous study [25].

Short-term (2-day) administration

In this group, the SSRI escitalopram was administered s.c. for 2 days with a minipump. The selective 5-HT2B receptor antagonist LY266097 (pKi 5-HT2A: 7.7, pKi 5-HT2B: 9.8, pKi 5-HT2C: 7.6 [10, 26] was administered i.p. alone or concomitantly with escitalopram for 2 days. Doses used were based on previous studies: escitalopram 2 mg/kg [27] and LY266097 0.6 mg/kg [28].

Long-term (14-day) administration experiments

In this group, the same dose of escitalopram was administered for 14 days with a minipump. LY266097 or aripiprazole were administered alone or concomitantly with escitalopram for 3 days and 14 days, respectively, preceding electrophysiological recordings. Doses of LY266097 and escitalopram were the same to those used for 2 days, and a dose of aripiprazole 2 mg/kg was used based on a previous study [4].

In vivo electrophysiological recordings

Recording of VTA DA neurons

Putative DA neurons were recorded by positioning a single-barrel glass micropipette according to the following coordinates (in millimeters [mm] from lambda): A–P, 3.2–3.7; M–L, 0.6–1.0; D–V, 7.0–9.0 [29]. DA neurons were identified according to the following electrophysiological properties: 1) a firing rate of 2–10 Hz, which may include burst firing, 2) a biphasic or triphasic action potential with a “notch” in the rising phase and a prominent negative inflection, and 3) a spike duration >1.1 ms from spike initiation to the trough of the negative inflection. DA neurons burst activity was analyzed using the following criteria: a series of 2–10 spikes of decreasing amplitude, with a maximal interspike interval (ISI) of 80 ms for initiation of the spike train, and a maximal ISI of 160 ms for the continuation of the spike train [30, 31]. Nine electrode descents were carried out in each rat and neurons were recorded for 2 min after stabilization. The number of spontaneously active neurons per track was identified by recording multiple tracks in a 400–400 µm grid [28].

Recording of DRN 5-HT neurons

Putative 5-HT neurons were recorded by positioning a single-barrel glass micropipette according to the following coordinates (in mm from lambda): A–P, 0.8–1.2; M–L, 0; D–V, 5.0–7.0 [29]. 5-HT neurons were identified according to the following electrophysiological properties: 1) a firing rate of 0.5–3 Hz, 2) a biphasic or triphasic action potential with steady, regular firing and, 3) a spike duration of 1.5–3.0 ms [32, 33]. Four to five electrode descents were carried out in each rat and neurons were recorded for 2 min after stabilization.

Recording of pyramidal neurons in the mPFC

Previous works have reported pyramidal neurons in the mPFC are glutamatergic in nature [34, 35]. Putative mPFC pyramidal neurons were recorded by positioning a single-barrel glass micropipette according to the following coordinates (in mm from bregma): A–P, 3.2–3.4; M–L, 0.6–0.8; D–V, 2.5–5.5 [29]. Pyramidal neurons were identified according to the following electrophysiological properties: 1) a firing rate of 0.01–3 Hz, 2) a biphasic or triphasic action potential with highly irregular firing and 3) a spike with a positive inflection duration >0.36 ms and negative inflection duration >1.08 ms to exclude any fast-spiking interneurons [36]. Burst firing of pyramidal neurons was analyzed using the following criteria: a series of 2 or more spikes, with a maximal ISI of 45 ms for the initiation and continuation of the spike train [37]. For acute experiments in the mPFC, 5–8 neurons were recorded per rat to establish baseline firing and burst activity, then aripiprazole was administered i.v. After the injection, another 5–8 neurons were recorded per rat. For long-term experiments, 4–5 electrode descents were carried out in each rat and neurons were recorded for 5 min after stabilization.

Recording and microiontophoresis in CA3 dorsal hippocampus

Glutamatergic dorsal hippocampus pyramidal neurons [38, 39] were recorded by positioning a five-barrel glass micropipette according to the following coordinates (in mm from lambda): A–P, 4.0–4.2; M–L; 4.0–4.2; D–V, 3.5–4.5 [29]. Since pyramidal neurons in the hippocampus do not discharge spontaneously under chloral hydrate anesthesia, quisqualic acid was used to activate these neurons within their physiological range (10–15 Hz). Pyramidal neurons were identified according to the following electrophysiological properties: 1) large amplitude (0.5–1.2 mV), 2) long duration (0.8–1.2 ms) simple action potentials alternating with 3) complex spike discharges [40]. The following compounds were used to fill the 5-barrel electrode: 10 mM 5-HT in 200 mM NaCl (pH 4), 10 mM BW723c86 in 200 mM NaCl (pH 1.3), 1.5 mM quisqualate in 200 mM NaCl (pH 8), and 2 M NaCl used for automatic current balancing. The central barrel filled with NaCl 2 M is used for recordings. Iontophoretic ejection of 5-HT for 50 sec (s) suppresses the firing activity of CA3 pyramidal neurons. The inhibited pyramidal neurons gradually regain their initial firing activity after the cessation of ejection due to reuptake of 5-HT. To reliably determine the activity of 5-HT transporter (5-HTT) in vivo, RT-50 index was determined. It is defined as the time elapsed from the cessation of iontophoretic application of 5-HT to 50% recovery of the initial firing rate [41, 42].

Drugs

Escitalopram was generously provided by Lundbeck A/S Ltd. (Valby, Denmark) and was dissolved in water. Aripiprazole (LKT Laboratories, St. Paul, USA) was dissolved in 2% lactic acid. BW723c86 and RS127445 (Tocris, Burlington, ON, Canada) were dissolved in 10% lactic acid. LY266097 (Tocris, Burlington, ON, Canada) was dissolved in 20% hydroxypropyl-beta-cyclodextrin. Apomorphine and haloperidol (SigmaAldrich, Oakville, ON, Canada) were dissolved in 0.5% lactic acid. In all solutions, distilled water was the main solvent, and the pH was adjusted.

Data acquisition and statistical analyses

Recordings were acquired using CED Spike2 data acquisition software (Cambridge Electronic Design, Cambridge, UK). Firing and bursts were analyzed using burstiDAtor [43], except for burst of pyramidal neurons in mPFC were analyzed using an additional package (‘bursts.s2s’) in Spike2. Data in figures were expressed as means ± S.E.M. and were analyzed using SigmaPlot 12.5. The paired t-test, the Kruskal–Wallis one-way ANOVA on Ranks, the one-way repeated-measures ANOVA, and the two-way repeated-measures ANOVA were used in this study. Pairwise comparisons of parametric and nonparametric tests were performed using, respectively, Holm–Sidak and Dunn’s method. Threshold for statistical significance was set to 0.05.

Results

Acute 5-HT2B receptor activation inhibits VTA DA firing and burst activity

RS127445 by itself did not alter firing and burst activity of DA neurons (paired samples t-test; t [7] = 1.5, p > 0.05 and t [7] = −0.4, p > 0.05; data not shown). A two-way repeated measures ANOVA conducted on the percentage of firing activity of DA neurons showed a significant effect of pre-treatment (saline or RS127445; F[1,39] = 5.1; p < 0.05) and an interaction between the dose of BW723c86 administered and pre-treatment (F[3,39] = 6.0; p < 0.01). However, Holm–Sidak pairwise comparisons showed that in rats that received RS127445 pre-treatment, acute injection of BW723c86 at 4 and 6 mg/kg did not inhibit DA neuron firing activity (p > 0.05; Fig. 1a–c). On the percentage of burst activity of DA neurons, even though there was no significant effect of pre-treatment (F[1,39] = 2; p > 0.05), an interaction was statistically significant (F[3,39] = 3.8; p < 0.05; two-way repeated measures ANOVA). Holm–Sidak pairwise comparisons indicated that at a dose of 6 mg/kg, the inhibitory effect of BW723c86 on the number of bursts per minute was blocked by pre-injection of RS127445 (p > 0.05; Fig. 1a, b, d).

Fig. 1: 5-HT2B receptors modulate firing and burst activity of DA neurons.
figure 1

a Integrated firing rate histogram illustrating the inhibitory effects of BW723c86 on a DA neuron. The upper trace in these panels corresponds to burst activity of the neuron. Apomorphine and haloperidol were injected to ascertain the DA nature of recorded neuron. b Integrated firing rate histogram illustrating RS127445 pre-administration blocking the inhibitory effects of BW723c86 on a DA neuron. The upper trace in these panels corresponds to burst activity of the neuron. Apomorphine and haloperidol were injected to ascertain the DA nature of recorded neuron. c Percentage change in the firing rate of VTA DA neurons in rats that received an acute cumulative dose of the 5-HT2B agonist BW723c86 (N = 7; dotted lines) following a pre-treatment with saline or the selective 5-HT2B antagonist RS127445 (N = 8; solid lines). d Percentage change in the burst rate of DA neurons in rats administered an acute cumulative dose of the 5-HT2B agonist BW723c86 (N = 7; dotted lines) after pre-administration of saline or the selective 5-HT2B antagonist RS127445 (N = 8; solid lines). Note that the acute inhibitory effects of BW723c86 on DA neurons firing and burst activity were prevented by RS127445. Data are presented as mean ± S.E.M. *p < 0.05, **p < 0.01,*** p < 0.001 between pretreatments for a given dose.

Full size image

Blockade of 5-HT2B receptors rescues SSRI-induced inhibition of DA but not 5-HT neurons

In the VTA, a 2-day regimen of escitalopram (10 mg/kg/day; s.c.) significantly decreased the firing activity of DA neurons (H[3] = 25.6, p < 0.001; Kruskal–Wallis One-Way ANOVA on Ranks followed by Dunn’s method; Fig. 2a). However, there was no change in the number of spontaneously active DA neurons encountered per electrode descent (population activity; H[3] = 1.8, p > 0.05; Kruskal–Wallis One-Way ANOVA on Ranks; Fig. 2b). Also, there was no alteration in the percentage of spikes occurring in burst (H[3] = 3.3, p > 0.05; Kruskal–Wallis One-Way ANOVA on Ranks; Fig. 2c) and the number of bursts per minute (H[3] = 2.8, p > 0.05; Kruskal–Wallis One-Way ANOVA on Ranks; Fig. 2d). Whereas the administration of the selective 5-HT2B receptor antagonist LY266097 (0.6 mg/kg/day for 2 days; i.p.) alone had no effect on these parameters, its co-administration counteracted the inhibitory effect of escitalopram on the firing activity of DA neurons, resulting in a recovery to control level (p > 0.05; Fig. 2a).

Fig. 2: Blockade of 5-HT2B receptors reverses escitalopram-induced inhibition of DA neurons.
figure 2

Firing rate (a) and population activity (b) of VTA DA neurons following a two-day administration of vehicle, escitalopram (10 mg/kg/day; s.c.), the 5-HT2B antagonist LY266097 (0.6 mg/kg/day; i.p.) and their combination. DA neurons burst rate expressed as % spikes occurring in burst (c) or in number of bursts per min (d) after a two-day administration of vehicle, escitalopram (10 mg/kg/day; s.c.), the 5-HT2B antagonist LY266097 (0.6 mg/kg/day; i.p.) and their combination. Data are presented as mean ± S.E.M. *p < 0.05 relative to control group. Numerators in the histogram represent the total number of rats used, and denominators represent the total number of neurons recorded. Open circles represent the mean of firing, population, and burst activity in each rat.

Full size image

In the DRN, a 2-day regimen of escitalopram significantly decreased the firing activity of 5-HT neurons (H[3] = 102, p < 0.001; Kruskal–Wallis One-Way ANOVA on Ranks followed by Dunn’s method, p < 0.05; Fig. 3). Despite an increase of 40%, the administration of LY266097 alone for two days had no significant effect on the firing activity of 5-HT neurons. The co-administration of LY266097 did not rescue escitalopram-induced inhibition of 5-HT neuron firing activity (p < 0.05; Fig. 3).

Fig. 3: Blockade of 5-HT2B receptors does not rescue escitalopram-induced inhibition of 5-HT neurons.
figure 3

Firing rate of DRN 5-HT neurons in rats following 2-day administration of vehicle, escitalopram (10 mg/kg/day; s.c.), the 5-HT2B antagonist LY266097 (0.6 mg/kg/day; i.p.) and their combination. Data are presented as mean ± S.E.M. *p < 0.05 relative to control group. Numerators within the histogram represent the total number of rats used, and denominators represent the total number of neurons recorded. Open circles represent the mean of firing activity in each rat.

Full size image

The increase of mPFC pyramidal neuron activity by aripiprazole alone and in combination with escitalopram may be mediated by 5-HT2B receptors

Acute aripiprazole administration (0.6 mg/kg; i.v.) increased the firing activity of mPFC pyramidal neurons by 30%, but it was not statistically significant (t [7] = −1.8, p > 0.05; paired samples t-test; data not shown). However, it significantly increased their burst activity by 50% (t [7] = −3.0, p < 0.05; paired samples t-test; data not shown).

Long-term administration of escitalopram had no effect on the firing and burst activity of mPFC pyramidal neurons. After a 14-day regimen, aripiprazole alone and in combination with escitalopram significantly increased the firing and burst activity of pyramidal neurons by 105 and 150%, respectively (For firing: H[3] = 15.8, p < 0.01 and for bursts: H[3] = 12.1, p < 0.01; Kruskal–Wallis One-Way ANOVA on Ranks followed by Dunn’s method; Fig. 4a, b).

Fig. 4: Blockade of 5-HT2B receptors increases mPFC pyramidal neurons firing and burst activity.
figure 4

Firing (a) and burst rate (b) of mPFC pyramidal neurons following 14-day administration of vehicle, escitalopram (10 mg/kg/day; s.c.), aripiprazole (2 mg/kg/day; s.c.) and their combination. c Firing and burst rate (d) of pyramidal neurons in rats administered vehicle, escitalopram for 14 days (10 mg/kg/day; s.c.). LY266097 was added for the last 3 days of saline or escitalopram treatment, which lasted for 14 days. Data are presented as mean ± S.E.M. *p < 0.05, relative to control group. Numerators within the histogram represent the total number of rats used, and denominators within the same histogram represent the total number of neurons recorded. Open circles represent the mean firing and burst activity in each rat.

Full size image

To assess whether 5-HT2B receptors were involved in the increase of mPFC pyramidal neurons firing and burst activity induced by aripiprazole, the 5-HT2B receptor antagonist LY266097 was administered alone or concomitantly with escitalopram. The long-term administration of LY266097 both alone and when combined with escitalopram, significantly enhanced the firing and burst activity of pyramidal neurons by 165 and 125%, respectively (For firing: H[3] = 27.3, p < 0.001 and for burst: H[3] = 15.2, p < 0.01; Kruskal–Wallis One-Way ANOVA on Ranks followed by Dunn’s method; Fig. 4c, d), similar to the combination of aripiprazole and escitalopram.

5-HT2B receptor agonism impairs the effect of escitalopram on the 5-HTT in vivo

CA3 pyramidal neuronal firing activity was suppressed by microiontophoretic application of 5-HT and displayed a recovery time to 50% of baseline firing (RT-50) with an average of 70 ± 12 s. Because 5-HTTs were inhibited following i.v. injection of escitalopram (0.2 mg/kg; i.v.), the suppressant effect of 5-HT was significantly prolonged to a RT-50 of 140 ± 14 s (F[3,8] = 7.3, p < 0.05; One-Way Repeated Measures ANOVA followed by Holm–Sidak method; Fig. 5a, b). This escitalopram-induced prolongation was blocked by the subsequent microiontophoretic application of the 5-HT2B receptor agonist BW723c86 (RT-50 = 69 ± 12 s; Holm–Sidak method; p > 0.05, Fig. 5a, b). Following blockade of 5-HT2B receptors by RS127445 (2 mg/kg; s.c.), the RT-50 recovered back to 104 ± 10 s, a duration that was not significantly different from that induced by escitalopram (Holm–Sidak method; p > 0.05, Fig. 5a, b).

Fig. 5: 5-HT2B receptor agonism but not antagonism impairs 5-HTT function.
figure 5

a Integrated firing rate histogram of a representative pyramidal neuron in the CA3 region of the hippocampus, illustrating the effect of i.v. injection of escitalopram (0.2 mg/kg) on the inhibitory effect of microiontophoretic application of 5-HT on firing activity. Note that the duration of 5-HT-induced inhibition was prolonged in the presence of escitalopram. While 5-HTTs were still blocked, the 5-HT2B receptor agonist BW723c86 was co-ejected with 5-HT. Note that BW723c86 blunted the prolongation induced by escitalopram. The 5-HT2B receptor antagonist RS127445 was then administered s.c. and reversed the effect of BW723c86 on 5-HT uptake. b RT-50 values in seconds as a measure of 5-HTT activity following the microiontophoretic application of 5-HT in the presence of the aforementioned treatments. Data are presented as mean ± S.E.M. *p < 0.05, relative to baseline. Numbers within the histograms represent the total number of neurons recorded. Only one neuron was recorded per rat. Open circles represent the RT-50 values in each rat.

Full size image

Discussion

In the present experiments, the 5-HT2B receptor agonist BW723c86 significantly decreased firing and burst activity of DA neurons. Previous studies, however, did not report a significant effect of BW723c86 on DA neuron firing or DA dialysate levels [44, 45], probably because these investigators used a lower dose than herein. Furthermore, several of the acute anxiolytic effects of BW723c86 were only observed after the administration of a relatively higher dose [21, 46]. Di Matteo et al. [44] showed that another agonist with equivalent affinity for 5-HT2B and 5-HT2C receptors, Ro 60-0175, decreased the firing activity of DA neurons. Considering that BW723c86 has similar affinity for 5-HT2B and 5-HT2C receptors [10], it is possible that the dose used in the present experiments (6 mg/kg, i.v.), acted on both receptor subtypes. While RS127445 administration on its own has been previously reported to decrease the firing of DA neurons [23], it had no effect in the present study. Although the basis for this discrepancy is unclear, it is possible that the electrophysiological effects of RS127445 on DA neurons stemmed from the fact that heterogeneous populations of DA neurons were recorded, as previously reported [47]. Nonetheless, RS127445 prevented the inhibition of DA neuronal firing and burst activity induced by BW723c86. This indicated that the 5-HT2B receptor mediated the inhibition of DA neurons firing and burst activity.

Escitalopram induced a decrease in the firing activity of DA neurons after a 2-day regimen as shown here and in line with previous results [4, 48]. However, 5-HT2B receptor blockade did not change firing and burst activity of DA neurons after 2-day administration of LY266097. Since 5-HT2B receptor mRNA is present in the VTA [13], it is possible that blockade of these receptors underlies the rescue of inhibition exerted by escitalopram on DA neurons firing activity herein. Indeed, the addition of LY266097 to escitalopram for 2 days resulted in recovery of firing of DA neurons, indicating that 5-HT2B receptors are implicated, at least in part, in this normalization of firing activity. Interestingly, a similar rescue of firing activity of DA neurons was also demonstrated when escitalopram was co-administered with aripiprazole, which exerts an antagonist activity at 5-HT2B receptors [4, 8]. A potential role for 5-HT2C receptors in this rescue had been previously established since escitalopram-induced inhibition of DA neurons was restored to the control level by the administration of the 5-HT2C receptor antagonist SB242084 or aripiprazole, which also has a high affinity for 5-HT2C receptors [4, 8, 48]. Altogether, these results indicate that 5-HT2B receptor blockade is probably involved in the rescue of escitalopram-induced inhibition of DA neurons, hence possibly involving this receptor as a target in the antidepressant effect, particularly in the combination of 5-HT reuptake inhibition and 5-HT/DA partial agonists possessing high affinity for 5-HT2B receptors such as aripiprazole.

A previous study has shown that blockade of 5-HT2B receptors located on rat DRN GABA neurons by RS127445 [12], results in a significant increase of 5-HT neuronal firing activity [49]. Although an increase of similar magnitude (40%) was observed with sub-acute administration of LY266097, herein, this did not reach significance. However, the addition of LY266097 did not rescue escitalopram-induced inhibition of 5-HT neuronal firing activity. Similar combination composed of RS127445 and citalopram significantly enhanced levels of 5-HT around DRN 5-HT neurons [12], hence triggering the negative feedback exerted by the 5-HT1A autoreceptor. Therefore, the lack of effect of LY266097 addition to escitalopram may be due to the predominance of the activity of the autoreceptor over the 5-HT2B receptor in modulating 5-HT neuronal firing. It is thus possible that the desensitization of the autoreceptor would unmask the reversing effect mediated by 5-HT2B receptor blockade. Nevertheless, the overall serotonergic effects of combination of RS127445 and citalopram on 5-HT concentrations in projection areas such as the mPFC were found to be augmented [12]. It is noteworthy that, in mice however, similar regimen combining RS127445 and paroxetine did not increase 5-HT levels in the hippocampus [17]. This apparent discrepancy may stem from the fact that 5-HT2B receptors are located on DRN 5-HT neurons in mice [19], but on GABA neurons that inhibit the firing of 5-HT neurons in the rat DRN [12]. Thus, in rats, RS127445 would block 5-HT2B receptors located on GABA neurons, hence resulting in a disinhibition of 5-HT neurons and an increase in 5-HT levels in DRN and mPFC [12]. In mice, RS127445 would antagonize excitatory 5-HT2B receptors on 5-HT neurons, which results in dampening 5-HT concentrations in the hippocampus [17]. While several findings show that 5-HT neurons are under complex regulation by 5-HT1A, 5-HT1B, 5-HT1D, 5-HT2A, 5-HT6, 5-HT7 receptors [50,51,52,53], these results, altogether, point to a role for 5-HT2B receptors in this modulation and consequently the antidepressant response. This also shows that blocking 5-HT2B receptors does not blunt reuptake inhibition by an SSRI in DRN.

It was shown that activation of the 5-HT2B receptor by the 5-HT2B receptor agonist BW723c86 induces a hyperphosphorylation of 5-HTTs [20] that was abolished in the presence of LY266097, thus confirming the role of the 5-HT2B receptor in post-translational modification of the 5-HTT. This hyperphosphorylated state was postulated to decrease the capacity of SSRIs to bind to the 5-HTT [20, 54]. In the present in vivo study, the iontophoretic application of BW723c86 reversed the escitalopram-induced blockade of 5-HTT in the hippocampus, which was blocked by RS127445. This provided further evidence that 5-HT2B receptor activation may disrupt SSRI binding to the 5-HTT. On the contrary, escitalopram-induced inhibition of 5-HT reuptake was unaltered in the presence of 5-HT2B receptor blockade when aripiprazole was added [7]. This confirms that in the presence of 5-HT2B receptor antagonism, reuptake inhibition by an SSRI was not impaired in the hippocampus, as was the case in the DRN.

In the mPFC, an acute systemic low dose of aripiprazole, previously shown to increase DA levels [25, 55], elicited an increase in the number of bursts per minute but not the mean firing activity of pyramidal neurons. Long-term drug administration regimens, however, are more relevant to the therapeutic effects of drugs since their beneficial actions take place after repeated treatment in the clinic. Long-term administration of escitalopram did not induce any change in the firing and burst activity of pyramidal neurons, which is in line with previous results [36]. However, sustained administration of aripiprazole alone or in combination with escitalopram increased firing and burst activity of pyramidal neurons herein. Since 5-HT2B receptors are expressed on pyramidal neurons [15] and the mixed 5-HT2B/2C receptor agonist meta-chlorophenylpiperazine (mCPP) [56] inhibited their firing activity [57], it is, thus, possible that this enhancement in neuronal activity involved blockade of these receptors. This is strengthened by the present results showing that blockade of 5-HT2B receptors by the selective 5-HT2B receptor antagonist, LY266097, alone or when added to escitalopram resulted in an increase of firing and burst activity with comparable magnitude to that induced by aripiprazole administration alone and in combination with escitalopram. Similarly, a 3-week administration of olanzapine, which has strong affinity for 5-HT2B receptors [9, 58], increased the basal firing activity of pyramidal neurons and reversed fluoxetine-induced inhibition of these neurons [59]. These data show that 5-HT2B receptor blockade plays a role in the augmenting effects of aripiprazole alone or when added to 5-HTT inhibition. As is the case for 5-HT neurons, mPFC pyramidal neurons are modulated by several 5-HT receptor subtypes with the best characterized being the inhibitory 5-HT1A and the excitatory 5-HT2A receptors [60]. While several medications used in the treatment of MDD engage these receptors, 5-HT2B receptor blockade may thus also be a relevant target for antidepressant drugs.

Although a role of 5-HT2B receptors in the antidepressant-like response was already reported, its involvement in the augmenting effect stemming from both blockade of 5-HT2B receptors and 5-HTTs was never investigated. Indeed, the present results add the following: 1) 5-HT2B receptor antagonism reversed an SSRI-induced decrease in DA but not 5-HT neuron activity, and 2) antagonism of 5-HT2B receptors alone or in combination with an SSRI increased mPFC pyramidal neuron activity. 3) Lastly, it was shown herein that 5-HT2B receptor activation appears to interfere with the 5-HT reuptake process, while their blockade does not [7]. The present experiments have thus unveiled the potential role of 5-HT2B receptor antagonism in commonly used augmentation treatment strategies in patients with MDD having an inadequate response to first-line medications. It is noteworthy to mention that besides aripiprazole, three other medications used as adjuncts in TRD have a higher affinity for the 5-HT2B receptor than the D2 receptor (namely olanzapine, quetiapine, and ziprasidone [9]). Whether antagonism at 5-HT2B receptors by itself is sufficient to induce a beneficial therapeutic effect remains to be elucidated.

Funding and disclosure

This research was conducted as part of the Canadian Biomarker Integration Network for Depression (CAN-BIND) program through the Ontario Brain Institute, an independent non-profit corporation, funded partially by the Ontario Government. The opinions, results, and conclusions are those of the authors and no endorsement by the Ontario Brain Institute is intended or should be inferred. RH and ElM declare no conflict of interest. PB received grant funding and/or honoraria for lectures, expert testimony, and/or participation in advisory boards from Allergan, Bristol Myers Squibb, Eli Lilly, Janssen, Lundbeck, Otsuka, Pfizer, Pierre Fabre Médicaments, Takeda, and Valeant.