Optogenetic activation of septal GABAergic afferents entrains neuronal firing in the medial habenula

The medial habenula (MHb) plays an important role in nicotine-related behaviors such as nicotine aversion and withdrawal. The MHb receives GABAergic input from the medial septum/diagonal band of Broca (MS/DB), yet the synaptic mechanism that regulates MHb activity is unclear. GABA (γ -aminobutyric acid) is a major inhibitory neurotransmitter activating both GABAA receptors and GABAB receptors. Depending on intracellular chloride concentration, however, GABAA receptors also function in an excitatory manner. In the absence of various synaptic inputs, we found that MHb neurons displayed spontaneous tonic firing at a rate of about ~4.4 Hz. Optogenetic stimulation of MS/DB inputs to the MHb evoked GABAA receptor-mediated synaptic currents, which produced stimulus-locked neuronal firing. Subsequent delayed yet lasting activation of GABAB receptors attenuated the intrinsic tonic firing. Consequently, septal GABAergic input alone orchestrates both excitatory GABAA and inhibitory GABAB receptors, thereby entraining the firing of MHb neurons.

Interestingly, it has been known that lack of KCC2 in MHb neurons results in a high internal chloride concentration 20,21 . Therefore, GABAergic transmission in the MHb may function in excitatory and inhibitory manners via GABA A receptors and GABA B receptors, respectively. Given that MHb neurons generate spontaneous tonic firing 4,20 , we asked whether this rhythmic firing can be modulated by MS/DB GABAergic inputs. In the present study, exploiting the optogenetic controls of MS/DB input to the MHb, we showed that GABA released from MS/DB afferents entrains MHb neuronal firing by orchestrating both excitatory GABA A receptors and inhibitory GABA B receptors. 20,21 . Given that KCC2 is critical for extruding Cl − in adult neurons 15,22 , intracellular Cl − concentration in MHb neurons is likely to be higher. Therefore, to obtain stable intrinsic firing frequency without disturbing the intracellular ionic composition, loose-seal cell-attached recordings were made with aCSF-filled patch pipettes 23 . Recordings were performed at the ventral region of the MHb as shown in Fig. 1A and stable action currents with average firing frequency of 4.36 ± 1.06 Hz were consistently observed. When we measure any time-course change of the firing rates, the paired Student's t-test showed no significant difference between the initial value (0-5 min) and late value (15-20 min) (aCSF: 99.99 ± 6.57% compared with baseline, n = 7; Fig. 1B), indicating that stable recordings can be maintained for at least 20 min. The MHb receives GABAergic, glutamatergic and purinergic inputs [7][8][9] . In addition, the MHb has marked expression of nicotinic receptors 2,5 . Thus, we first examined whether the synaptic inputs are involved in the tonic firing of MHb neurons. Inhibition of glutamatergic (AMPA receptor antagonist, CNQX 10 μ M; NMDA receptor antagonist, D-APV 30 μ M), cholinergic (nACh receptor antagonist, mecamylamine 10 μ M), purinergic (P2X receptor antagonist, PPADS 50 μ M) or GABAergic (GABA A receptor antagonist, picrotoxin 100 μ M; GABA B receptor antagonist, CGP52432 10 μ M) inputs did not modify the spontaneous tonic firing (Fig. 1). When the firing rates recorded 10-15 minutes after the drug treatments were compared with baseline firing rates (0-5 min), the paired Student's t-test revealed no significant effect of treatments on percent changes in the firing rates (CNQX and D-APV: 98.88 ± 1.70, n = 6; mecamylamine: 107.60 ± 3.54, n = 5 ; PPADS: 107.6 ± 2.54, n = 6; picrotoxin: 100.70 ± 6.54%, n = 4; CGP52432: 107 ± 3.69, n = 5). The results demonstrate that spontaneous tonic firing is independent of these synaptic inputs. Now, we speculated that activation of GABA A receptors elicits excitatory postsynaptic currents (EPSCs) in MHb neurons due to the lack of KCC2 proteins thereby enhances MHb neuronal activity. As expected, muscimol (10 μ M), a GABA A receptor agonist, showed significant effect on firing frequency (F 1, 4 = 65.75, P = 0.0013, one-way repeated measure ANOVA). As shown in Fig. 2A, tonic firing was briefly increased 185.3 ± 19.8% in the early stage of muscimol application (P < 0.05 compared with baseline firing frequency, Bonferroni's post-hoc test) and quickly became quiescent (P < 0.001 compared with baseline firing frequency, Bonferroni's post hoc test). This might be due to the sodium channel inactivation as a result of prolonged membrane depolarization caused by lasting excitatory GABA-mediated currents.

Spontaneous tonic firing in MHb neurons. MHb neurons lack KCC2 expression
Meanwhile, GABA B receptor is G-protein coupled receptor that associates with pertussis toxin sensitive Gi/o family, that in turn regulates specific ion channels and cAMP cascades 10,11 . Consequently, activation of GABA B receptors stabilizes neuronal activity. As shown in Fig. 2B, baclofen (10 μ M), a GABA B receptor agonist, markedly blocked MHb neuronal firing (9.83 ± 4.32% compared with baseline firing rates, p < 0.0001, n = 5, paired t-test). The result is consistent with the well-known inhibitory effect of GABA B receptors on neuronal excitability. Taken together, while GABA may not be involved in basal tonic firing of MHb neurons (Fig. 1), activation of GABA transmission would actively modify neuronal firing. Therefore, we next tested the possibility that GABAergic inputs to the MHb efficiently regulate neuronal tonic firing in the MHb. The MS/DB GABAergic input to the MHb. Previous histological study using retrograde tracers show that GABAergic neurons in MS/DB project to the MHb 7 . However, functional GABAergic connections between these two regions have not been examined. Using optogenetic approaches, we tried to elucidate the functional properties of the synapses connecting between the MS/DB and the MHb. We delivered AAV expressing Chronos-GFP [Syn:: Chronos-GFP] into the MS/DB (Fig. 3A). Four to six weeks post infection, marked GFP signals were observed in the injection site (Fig. 3B). In addition, significant GFP signals were also apparent in the ventral MHb (Fig. 3C). As expected, the GFP-positive axon collaterals from the MS/DB showed the expression of vGAT1, a GABAergic presynaptic marker (Fig. 3D).
Next we measured the reversal potential of the GABA A receptor-mediated currents (E GABA ) elicited by light stimulation. To preserve intracellular Cl − concentration, gramicidin perforated patch recordings were performed. Figure 4E shows representative current traces evoked by a brief light stimulation (470 nm, 10-ms duration) at different membrane potentials: The light-evoked Cl − currents were reversed in polarity between − 40 mV and − 30 mV. We then constructed current-voltage (I-V) plots for the Cl − currents and used linear regression analysis to measure E GABA (Fig. 4F). E GABA was estimated to be − 33.27 ± 1.43 mV (n = 4), which was more depolarized than resting membrane potential of MHb neurons (− 44.50 ± 1.29 mV, n = 12). Therefore, activation of GABA A receptors is expected to produce EPSCs in MHb neurons.
Although it has been reported that GABA A receptor signaling is absent in the MHb 21 , our data clearly verify the functional GABA A receptor-mediated synapses in the MHb. As we assessed the expression of mRNA encoding GABA A receptor subtypes in the MHb, multiple subtypes appeared to be expressed in the MHb (Supplementary information Fig. S1).

Entrainment of MHb neuronal firing via both GABA A and GABA B receptors. Our immunohis-
tochemical and optogenetic approaches demonstrate that the MS/DB exerts GABAergic transmission on MHb neurons. Since agonists for GABA A receptors increased MHb neuronal firing ( Fig. 2A) and E GABA was more depolarized than resting membrane potential of MHb neurons, we tested whether the MS/DB GABAergic inputs can generate firing in MHb neurons. To this end, we delivered light stimulus (470 nm, 5 ms) to activate Chronos  expressed in MS/DB afferents upon loose-patch cell-attached recordings. Raster plot, corresponding normalized firing frequency and z-score profiles demonstrated that a single light stimulus reliably induced neuronal firing with the kinetics not different from those of the spontaneous tonic firing (Fig. 5A inset, Pearson correlation coefficient r = 0.95, P < 0.0001). A light stimulus frequently triggered a brief burst (Fig. 5A), which might be attributed to the light-evoked strong depolarization. Given the high input resistance of MHb neurons (1.29 ± 0.11 GΩ, n = 12), we expected that EPSPs generated by light-activated GABA A currents are sufficient to trigger action potentials in MHb neurons. Indeed, the light-induced firing was completely abolished by picrotoxin (Fig. 5B). Meanwhile, a light stimulus not only induced immediate firing, but also generated delayed but prolonged quiet periods when spontaneous tonic firing was suppressed (as judged by negative values of z-score, Fig. 5A). Intriguingly, the suppression of tonic firing was still maintained in the presence of pricrotoxin. This might be due to the delayed but lasting activation of GABA B receptors. Indeed, CGP52432 rendered the quiet period less prominent without affecting light-induced firing (Fig. 5C). Consequently, GABA A and GABA B receptors cooperatively participate in resetting the intrinsic tonic firing of MHb neurons.
We next sought to establish whether various frequency activation of GABAergic input from the MS/DB can affect firing output in MHb neurons. When light with various frequencies (1-10 Hz) were applied, raster plot and corresponding z-score profile showed that MHb neurons reliably followed the stimulation frequency in a stimulus-locked manner, regardless of their original tonic firing frequency (Fig. 6A). To ensure the involvement of both GABA A and GABA B receptors on the firing entrainment, we examined the effects of picrotoxin and CGP52432. As expected, treatment of picrotoxin completely failed light-evoked firing, consequently abolished the firing entrainment (Fig. 6B). Instead, prominent light-induced suppression of intrinsic tonic firing was apparent in the presence of picrotoxin, which might be due to the activation of GABA B receptors alone. In contrast, light stimulations in the presence of CGP52432 induced limited entrainment: light elicited GABA A receptor-dependent firing, with less suppressed intrinsic tonic firings between stimulations (Fig. 6C). For this reason, CGP52432 rendered low-frequency entrainment (1 and 2 Hz) far less accurate. In conclusion, GABA A and GABA B receptors activated by GABA released from MS/DB afferents entrain MHb neuronal firing by exerting opposite effects on neuronal activity.

Discussion
In the present study, we have demonstrated functional GABAergic synaptic connectivity between the MS/DB and the MHb. More importantly, we revealed that the GABAergic transmission alone is sufficient to entrain rhythmic firing in the MHb. We hope that our findings will give insight to understand MHb activity-mediated behaviors.
It has been known that MHb neurons possess spontaneous tonic firing 4,20 . A recent study revealed that MHb neurons are equipped with hyperpolarization-activated cyclic nucleotide-gated (HCN) channels that confer them with intrinsic rhythmic firing 4 . Indeed, the tonic firing is maintained without synaptic input as shown in this study (Fig. 1) as well as the previous observations 4, 20 . This intrinsically generated tonic firing was dramatically modified following the activation of GABA A and GABA B receptors using the agonists, muscimol and baclofen, respectively: GABA A receptor activation by muscimol triggered robust firings in MHb neurons, whereas GABA B receptor activation by baclofen completely abolished the firings (Fig. 2). Previous study also reported GABA A excitation in the MHb from juvenile rats (18-to 25-day old) 20 . Since GABA A receptors are excitatory until 2

cells).
Scientific RepoRts | 6:34800 | DOI: 10.1038/srep34800 postnatal weeks in the developing brain 6 , the possibility was raised that the GABA A excitation in the MHb only reflects the immature property of the developing brain. However, we consistently observed GABA A excitation in the MHb obtained from adult mice (10-16 weeks), indicating that excitatory GABAergic activity is not attributed to immaturity of MHb neurons.
Both input and output pathways of the MHb have been well established [6][7][8]24,25 . Nevertheless, many functional studies related with the MHb have been focused on the output pathway [26][27][28] or the habenular nucleus itself 1,2,4,29,30 . Indeed, only one study, to our knowledge, has been attempted to explain functional relevance of posterial septal afferents to the MHb 31 . Here we focused on MS/DB afferents to the MHb to establish functional input pathway of the MHb. It has been reported that the MHb received GABAergic projection from MS/DB using a retrograde tracer 7 . Consistently, we observed that Chronos-GFP-expressing MS/DB afferents express vGAT1, a GABAergic presynaptic marker. Furthermore, light stimulation of Chronos-GFP-expressing MS/DB afferents evoked picrotoxin-sensitive GABA A currents, indicating functional synaptic connectivity of the GABAergic MS/DB afferents to the MHb. On the contrary, a previous study demonstrates that β 2/3 subtypes of GABA A receptors are not expressed in the MHb and GABA application evoked no measurable currents in the MHb 21 . Our study, however, identified GABA A receptor subtypes differentially expressed in the MHb using RT-PCR ( Supplementary  Fig. S1). More importantly, optogenetic stimulation evoked picrotoxin-sensitive GABA A currents in MHb neurons (Fig. 4).
Meanwhile, it has been reported that GABA B receptors are expressed substantially in MHb 20 (Allen institute, experiment number 68862120, 71247614). Consistent with these observations, we found that baclofen completely abolished spontaneous tonic firing (Fig. 2B). Considering that GABA A and GABA B receptors exert the opposite effects on the activity of MHb neurons, we supposed that GABAergic MS/DB input alone entrains thereby synchronize MHb neuronal firings. We found that GABA optogenetically released from MS/DB afferents immediately elicited firing via fast activation of GABA A receptors and subsequently suppressed intrinsic tonic firings via delayed but lasting activation of GABA B receptors (Fig. 5). As a result, GABAergic MS/DB input entrained firing of MHb neurons (Fig. 6).
The MS/DB plays a key role in generating theta oscillations in the hippocampus 32 . And most MS/DB GABAergic neuronal firing is phase locked to hippocampal theta in vivo 33 . Intriguingly, hippocampal input to the MS/DB preferentially generates rhythmic firing of GABAergic neurons in the MS/DB 34 . In brain slices ex vivo, spontaneous tonic firings of MHb neurons are not synchronized with other adjacent neurons 4 . Taking into account the facts that MS/DB GABAergic neuronal firing is tuned to theta frequency in vivo 33 and that GABAergic MS/DB input entrains neuronal firing in the MHb (Fig. 6), MS/DB input may synchronize MHb neuronal firing locked to theta rhythm in vivo. MHb neurons mainly project to the interpeduncular nucleus (IPN) 25,35,36 . Supposedly, unsynchronized intrinsic tonic firing in MHb neurons per se may produce subthreshold postsynaptic activity in IPN neurons. Now the synchronized MHb neuronal firing by MS/DB input may cause postsynaptic spatial summation, allowing IPN neurons to generate faithful suprathreshold activity.
Several studies have revealed that the MHb-IPN pathway play roles in nicotine-related behaviors. Activation of IPN GABAergic neurons that receive direct projection from the MHb triggers physical nicotine withdrawal symptoms 26 . Mice lacking nAChR α 5 subunit exhibit decreased MHb input to IPN, which results in attenuated nicotine aversion 1 . Conversely, elevated expression of the nAChR β 4 subunit increases nicotine aversion in mice by enhancing activity of the MHb to the IPN 2 . Experiments in animal models have demonstrated directly that the MHb-IPN pathway participates in nicotine withdrawal 37 . Therefore, it is plausible that MS/DB-MHb pathway also plays a role in nicotine-related behaviors because synchronized MHb neuronal firing would reliably activate IPN neurons. Although behavioral relevance still remains to be investigated, to our best knowledge, we have first demonstrated the MS/DB GABAergic entrainment of MHb neuronal firing.
Electrophysiology. Electrophysiological recordings were made using an EPC10 amplifier (HEKA elektronik). Patch-clamp pipettes were pulled (PP-83; Narishige) from borosilicate glass (Warner Instruments) and had a tip resistance of 3-6 MΩ when filled with internal solution. After recovery periods, acute slices were then transferred to the recording chamber, were fully submerged at a flow rate of 1.4-1.6 mL/min and maintained at 30 ± 1 °C in aCSF. Cells were visualized using epifluorescence and infrared differential interference contrast (IR-DIC) video microscopy with a 40X magnification water-immersion objective (BX51WI, Olympus). Tonic firings were measured in a loose cell-attached mode (8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25) to prevent internal dialysis and aCSF was used for pipet solution. Chronos was stimulated by brief 470 nm light (5-ms duration) through the optic fiber (NA = 0.35) using light-emitting diode (LED; Doric lens, LEDC2) powered by an LED driver (Thorlabs, LEDD1B) under control of pulse generator (AMPI, Master-8). After the recordings, some slices were fixed in 4% paraformaldehyde in PBS to perform immunohistochemistry. Synaptic currents evoked by light were recorded at − 70 mV in a whole-cell mode using pipet solution containing (in mM): 100 K-Gluconate, 20 KCl, 10 HEPES, 0.2 EGTA, 10 Na 2 -phosphocreatine, 4 MgATP, 0.3 Na 4 GTP; pH was adjusted to 7.2-7.3 with KOH. Pipet solution for gramicidin perforated patch recording contained (in mM) 140 KCl, 10 NaCl, 10 HEPES; pH was adjusted to 7.2-7.3 with KOH. Gramicidin (Sigma-Aldrich) was first dissolved in DMSO (10 mg/ml) to prepare a stock solution and then diluted to a final concentration of 10 μ g/ml in the pipet solution. The gramicidin-containing solution was prepared and sonicated immediately before use. After E GABA measurements, the integrity of the perforated patch was confirmed by rupturing the underlying sealed membrane and observing an abrupt change in access resistance and a shift in E GABA . Data were sampled at 10 kHz and filtered at 2.9 kHz with Bessel filter of the amplifier. Data were analyzed using Patchmaster (HEKA), Igor 6.0 (Wavemetrincs) or Minianalysis (synaptosoft).
Immunohistochemistry. 30 μ m cryosected brain slices were permeabilized in 0.6% Triton X-100 and blocked in 3% normal donkey serum in PBS for 30 minutes in free floating condition. Rabbit anti-vGat1 antibody (1:500, Synaptic systems) was incubated for overnight in 1% normal donkey serum and 0.1% triton X-100 in PBS at 4 °C. For visualization, slices were incubated with Cy3-conjugated anti-rabbit secondary antibody (Jackson ImmunoResearch Laboratories) for 2 hours. Immunostained slices were scanned using a confocal laser microscope (LSM510, Carl Zeiss).

RT-PCR.
Total RNA was extracted using RNeasy Mini Kit (QIAGEN #74104) according to the manufacturer's instruction. To synthesize first strand cDNA, 1 μ g of total RNA was incubated at 70 °C for 5 min with 0.5 μ g of oligodT and deionized water (up to 15 μ l). The reverse transcription reaction was performed using 200 units of M-MLV reverse transcriptase (Promega, Madison, WI, USA) in 5X reaction buffer (250 mmol/l Tris-HCl; pH 8.3, 375 mM KCl, 15 mM MgCl 2 , 50 mM DTT), 28 units of RNasin inhibitor, and 2.5 mM dNTP mixtures at 42 °C for 90 min. The expression of GABA A receptor subunits was examined by PCR as previously described 38,39 . Two microliters of the cDNA product were amplified in a mixture containing 5 pmole of GABAA receptor subtype-specific primers, 0.2 mM dNTPs and 1 unit Taq DNA polymerase (Promega, Madison, WI, USA) with reaction buffer in a final volume of 25 μ l. The PCR amplification was carried out for 35 cycles of 94 °C for 30 sec, 52 °C for 30 sec and 72 °C for 1 min. The primers used were: beta-actin sense primer, 5′ -TCACCCACACTGTGCCCATCTACGAG-3′; beta-actin anti-sense primer, 5′ -GTGGTGAAGCTGTAGCCACGCTC-3′ ; GABA A receptor subtype α 1, α 2, α3, α 4, α 5, α 6, β 1, β 2, β 3, γ 1, γ 2, γ 3, δ primers were manufactured as previously described 38 . GABA A receptor subtype ε , θ , ρ 1, ρ 2, ρ 3 primer sequences were referred to in the previous paper 39 . Statistics. All data are presented as mean ± SEM. Data were analyzed by one-way analysis of variance (ANOVA) or paired Student's t-test. Mean differences between groups were considered significant when P < 0.05. To quantify firing entrainment, loose seal cell-attached recording was normalized using firing probability and standard z-score transformation (bin size, 10 ms). Firing probability was calculated by following equations; sum of firing in each bins/total number of sweeps. Neuronal firing was normalized to the firing rates during 500 ms (Fig. 5) or 6 s (Fig. 6) just prior to light stimulus train.