Serotonin modulates glutamatergic transmission to neurons in the lateral habenula

The lateral habenula (LHb) is bilaterally connected with serotoninergic raphe nuclei, and expresses high density of serotonin receptors. However, actions of serotonin on the excitatory synaptic transmission to LHb neurons have not been thoroughly investigated. The LHb contains two anatomically and functionally distinct regions: lateral (LHbl) and medial (LHbm) divisions. We compared serotonin’s effects on glutamatergic transmission across the LHb in rat brains. Serotonin bi-directionally and differentially modulated glutamatergic transmission. Serotonin inhibited glutamatergic transmission in higher percentage of LHbl neurons but potentiated in higher percentage of LHbm neurons. Magnitude of potentiation was greater in LHbm than in LHbl. Type 2 and 3 serotonin receptor antagonists attenuated serotonin’s potentiation. The serotonin reuptake blocker, and the type 2 and 3 receptor agonists facilitated glutamatergic transmission in both LHbl and LHbm neurons. Thus, serotonin via activating its type 2, 3 receptors, increased glutamate release at nerve terminals in some LHb neurons. Our data demonstrated that serotonin affects both LHbm and LHbl. Serotonin might play an important role in processing information between the LHb and its downstream-targeted structures during decision-making. It may also contribute to a homeostatic balance underlying the neural circuitry between the LHb and raphe nuclei.

In this study, we examined the effects of 5-HT on the glutamate transmission in the LHb, and whether differences in the effects of 5-HT existed between the LHbm and LHbl neurons. We also examined the receptor subtypes that mediate 5-HT-induced facilitation of glutamate transmission.

Discussion
We provide here the first electrophysiological evidence that 5-HT bi-directionally and differentially regulates glutamate transmission in the LHb. Furthermore, 5-HT's facilitation of glutamate transmission is mediated by 5-HT 2 and 5-HT 3 receptors probably at the glutamatergic terminals. Finally, blockade of 5-HT reuptake facilitates glutamate transmission in most of the LHbm and LHbl neurons, suggesting that 5-HT may regulate glutamate transmission in the LHb under physiological conditions. Via altering glutamate transmission that regulates activity of LHb neurons, 5-HT may in turn alter the activity of raphe nuclei. Thus, 5-HT's effects in the LHb may provide a feedback loop to the raphe nuclei, and may help maintain homeostatic balance of serotonergic function.

5-HT bi-directionally and differentially modulates glutamate transmission in the LHbm and LHbl neurons. Our result of 5-HT inhibition of glutamate transmission in LHbl neurons generally agreed with previous reports
showing that 5-HT suppressed EPSCs in LHbl neurons in slices of rats 12 , and that 5-HT 1B agonist inhibited glutamate transmission in the LHb 28 . Our finding of 5-HT facilitation of glutamate transmission supports a recent in vivo study showing that intra-LHb injection of selective 5-HT 2C agonist Ro60-0175 increased the expression of depressive-like behaviors in rats, suggesting an excitatory effect of 5-HT in the LHb 27 . We extend this finding by demonstrating that 5-HT facilitated glutamate transmission in the majority of the LHbm neurons and a minority of the LHbl neurons.
A future study is needed to determine whether the different 5-HT effects are due to the difference in the neuronal phenotype. Using morphological analysis Weiss and Veh 31 revealed four main categories of projection neurons randomly distributed throughout the LHb. However, electrophysiological characterization of neurons within the different categories demonstrated no significant differences between groups. Based on the pattern of spontaneous activity, neurons were classified as silent, tonic or bursting. The occurrence of distinctive firing modes was not related to topographic allocation. These investigators thus concluded that the formation of functional neuronal entities within the LHb may be achieved through defined synaptic inputs to particular neurons, rather than by individual neuronal morphologies and intrinsic membrane properties. By analyzing the gene expressions of neurotransmitter markers in the habenula, Aizawa, et al. 21 showed that neurons in the LHb were almost uniformly glutamatergic. The current study revealed that EPSCs were recorded in almost all LHb neurons tested, and were sensitive to 5-HT. Thus, the different responses to 5-HT may be resulted from the heterogeneous expression of 5-HT receptors on the glutamatergic terminals in LHb neurons.
In the current study, we identified presynaptic 5-HT 2A/B/C and 5-HT 3 receptors mediate 5-HT-induced facilitation of glutamate transmission. 5-HT potentiated the frequency and amplitude of sEPSCs. Tetrodotoxin completely abolished the effect on sEPSC amplitude, indicating that 5-HT's effect on sEPSC amplitude depends on the action potentials. Action potential firing may increase the proportion of multiquantal events, thus skewing the amplitude distribution of synaptic currents towards larger size classes. 5-HT significantly increased the frequency but not the amplitude of mEPSCs, suggesting that 5-HT increases the probability of glutamate release, probably at the nerve terminals. In support, 5-HT increased the amplitude, and decreased the paired pulse ratio of the eEPSCs.
RNA labeling evidence indicates the existence of 5-HT 2 receptors in the LHb 32 . Pharmacological evidence has confirmed the existence of 5-HT 1B 28 and 5-HT 2C 27 receptors in rat LHb. In both the LHbl and LHbm, we showed that 5-HT 2A/C antagonist ritanserin and 5-HT 2B/C antagonist SB200646 substantially attenuated 5-HT-induced facilitation of sEPSCs. Conversely, the 5-HT 2C agonist mCPP increased sEPSCs, in general agreement with a recent in vivo study 27 . Our data suggest that activation of 5-HT 2 receptors increases glutamate release probability at the nerve terminals. The existence of 5-HT 3 receptors was revealed by the application of ondansetron, which significantly attenuated 5-HT-induced facilitation of sEPSCs. Accordingly, bath application of the selective 5-HT 3 agonist mCPBG increased sEPSCs. These data suggest that activation of 5-HT 3

receptors increases glutamate
Scientific RepoRts | 6:23798 | DOI: 10.1038/srep23798 release probability and partly mediates 5-HT's facilitation of glutamate transmission in the LHb. Notably, the cocktail containing 5-HT 2 and 5-HT 3 receptor antagonists completely abolished 5-HT-induced potentiation of glutamate transmission. Our results thus revealed the presence of functional 5-HT 2A/B/C and 5-HT 3 receptors that mediate 5-HT-induced facilitation of glutamate transmission in the LHb. 5-HT may modulate glutamate transmission in the LHb under physiological conditions. There is recent evidence 33 that the 5-HT reuptake blocker citalopram affected the activity of the synapse connecting the basal ganglia with the LHb neuron. In general agreement with their finding, we found that citalopram potentiated glutamate transmission in LHb neurons, and the antagonists of 5-HT 2A/C, 5-HT 2B/C , and 5-HT 3 receptors significantly attenuated this potentiation, and the cocktail containing 5-HT 2 and 5-HT 3 receptor antagonists almost completely abolished this potentiation. These results suggest that citalopram, via the accumulated extracellular 5-HT, activates presynaptic 5-HT 2 and 5-HT 3 receptors, and increases probability of glutamate release.
Medial and lateral subdivisions of the LHb were recognized in an early rat study 34 . Ultrastructural 35 and immunohistological 36 studies in rats have defined as many as four medial and five lateral LHb regions. A corresponding subnuclear structure has been described in the mouse 37 . As mentioned, the neurons in the LHbl and LHbm are heterogeneous with different connectivity 31 . The subregions of the medial and lateral nuclei give rise to distinct projections to midbrain areas 15 , both in rat 14,30 and mouse 38 . Since the LHbm sends glutamatergic projections mainly to interneurons in the raphe nucli 15,16 , LHbm activation may suppress raphe nuclei and 5-HT release. Conversely, since the LHbl mainly projects to the RMTg, which in turn sends GABAergic projections to the midbrain dopaminergic neurons and raphe serotoninergic neurons 13,14 , LHbl activation may reduce the activity of dopaminergic and serotoninergic neurons.
Notably, although 5-HT inhibited glutamate transmission in a subset of LHbl neurons, we have recently shown that 5-HT increases firing of the majority of the LHbl neurons by activating the postsynaptic 5-HT receptors 29 . We therefore speculate that the net effect of 5-HT may increase the activity of LHb neurons, which may in turn inhibit dopaminergic and serotoninergic neurons. This may provide an additional explanation at the cellular level for the previous functional discoveries; changes of the LHb activity lead to an opposite reaction of raphe cell activity, i.e., lesion of the LHb is followed by an increase in 5-HT in the dorsal raphe nucleus 20,[39][40][41] ; where electrical 42,43 as well as chemical 16,25 stimulation of the habenula markedly suppressed serotonergic neurons in the raphe nuclei.
The LHb itself is a hub in the forebrain, which plays critical roles in a variety of brain functions, such as depression, addiction, and sleep cycle disorders, as well as decision making 11,[44][45][46][47] . At the neuronal level, integration of synaptic inputs and intrinsic properties sets the frequency and pattern of neuronal firing activity. Through the influence over glutamate transmission in the LHb, 5-HT may change the LHb output to the downstream regions. In view of the extensive innervation on midbrain monoaminergic nuclei by LHb neurons, these dual actions of 5-HT may have a profound effect on the operation of the entire midbrain network. Taken together, our data suggest a feedback loop between the LHb and the raphe nuclei, which may have a role in the balance of the reciprocal neural activity. Disruption of this fine natural balance may be involved in many neuropathology like drug abuse and mood disorders: depression and anxiety.

Methods
Animals. The Animal Care and Utilization Committee of Rutgers, the State University of New Jersey, in accordance with National Institutes of Health guidelines, approved all procedures, minimizing the number of animals used and their suffering. Sprague-Dawley (SD) rats (n = 200) at postnatal days 25-35 of both sexes were housed under standard conditions at 22-24 °C, 50-60% humidity, and a 12 h light/dark cycle. Food and water are available to all rats ad libidum unless otherwise indicated. Since the data from juvenile male rats did not differ significantly from those from female rats, the data were pooled.
Brain slice preparation and electrophysiology. Coronal epithalamic slices (250 μm) were cut in ice-cold glycerol-based artificial cerebrospinal fluid (GaCSF) containing (in mM): 252 glycerol, 2.5 KCl, 1.25 NaH 2 PO 4 , 1 MgCl 2 , 2 CaCl 2 , 25 NaHCO 3 , 0.3 L-ascorbate, and 11 glucose, and saturated with 95%O 2 /5%CO 2 (carbogen). Slices were incubated for > 1-hr at 24-25 °C in carbogenated aCSF of similar composition as GaCSF, but with 126 mM NaCl replacing glycerol. Electrophysiological recordings (from ~700 LHb neurons) were performed at ~33 °C aCSF perfused at 1.5-2 ml/min, as described 48 . Patch pipettes (6)(7)(8) were filled with internal solutions containing (in mM) 140 cesium methanesulfonate, 5 KCl, 2 MgCl 2 , 10 HEPES, 2 MgATP, 0.2 GTP for recordings under voltage-clamp. Both evoked and spontaneous events were recorded at a holding potential (V H ) of − 70 mV in the presence of gabazine (10 μM) and strychnine (1 μM), which block GABAA and glycine receptors, respectively. These events were blocked by 6,7-dinitroquinoxaline-2,3-dione (DNQX), an antagonist of α -amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA) receptors; indicating that they were excitatory postsynaptic currents (EPSCs), mediated by AMPA receptors. Electrical stimuli (100-200 μs in duration, 0.05 Hz) elicited EPSCs via a nichrome wire bipolar electrode positioned within 200 μm of the soma. Near the start of the recording an input/output curve was obtained and the stimulation was then set to 20-30% of the maximum, an intensity that resulted in stable responses with no failures. Paired eEPSCs were elicited with a pair of identical stimuli separated by an interval of 50 milliseconds. Drugs. We purchased common salts and 1-(3-Chlorophenyl) biguanide hydrochloride (mCPBG); gabazine; Scientific RepoRts | 6:23798 | DOI: 10.1038/srep23798 Data analysis and statistics. All data are expressed as means ± SEM. Baseline electrophysiological data were recorded for 10 min, before drug superfusion, and during the washout. To calculate the percent change in EPSC frequency/amplitude for a given cell, recordings during the initial control period (baseline) were averaged and normalized to 100%. Comparisons between the LHbm and LHbl were made using two-tailed unpaired Student's t-tests. Possible significant differences in the percent distribution of EPSCs were compared by Chi-square test. The different concentrations of 5-HT on sEPSCs were analyzed using two-way ANOVA with "subregions" (LHbm vs LHbl) as between-group factors and "dose" (from 0.1 to 30 μM) as within subject factor. Tukey's post hoc test was used for multiple dose comparisons. The effects of 5-HT antagonists on changes in sEPSCs induced by 5-HT/citalopram were assessed by paired t-test or one-way ANOVA. Dose-response data were fitted to the logistic equation: y = 100x α /(x α + x o α ), where y is the percentage change, x is the concentration of 5-HT, α the slope parameter, and x o the 5-HT concentration which induces a half-maximal change. Values of p < 0.05 were considered significant.