T-Type Calcium Channels Contribute to Burst Firing in a Subpopulation of Medial Habenula Neurons

Abstract Action potential (AP) burst firing caused by the activation of low-voltage-activated T-type Ca2+ channels is a unique mode of neuronal firing. T-type channels have been implicated in diverse physiological and pathophysiological processes, including epilepsy, autism, and mood regulation, but the brain structures involved remain incompletely understood. The medial habenula (MHb) is an epithalamic structure implicated in anxiety-like and withdrawal behavior. Previous studies have shown that MHb neurons fire tonic APs at a frequency of ∼2–10 Hz or display depolarized low-amplitude membrane oscillations. Here, we report in C57BL/6J mice that a subpopulation of MHb neurons are capable of firing transient, high-frequency AP bursts mediated by T-type channels. Burst firing was observed following rebounding from hyperpolarizing current injections or during depolarization from hyperpolarized membrane potentials in ∼20% of MHb neurons. It was rarely observed at baseline but could be evoked in MHb neurons displaying different initial activity states. Further, we show that T-type channel mRNA, in particular Cav3.1, is expressed in the MHb in both cholinergic and substance P-ergic neurons. Pharmacological Cav3 antagonism blocked both burst firing and evoked Ca2+ currents in MHb neurons. Additionally, we observed high-frequency AP doublet firing at sustained depolarized membrane potentials that was independent of T-type channels. Thus, there is a greater diversity of AP firing patterns in MHb neurons than previously identified, including T-type channel-mediated burst firing, which may uniquely contribute to behaviors with relevance to neuropsychiatric disease.


Introduction
Burst action potential (AP) firing mediated by the activation of low-voltage-activated T-type Ca 21 channels is a unique mode of neuronal activity that occurs in various neuronal populations, particularly in thalamic, septal, and sensory neurons (Perez-Reyes, 2003;Cheong and Shin, 2013;Lambert et al., 2014). This burst firing has been shown to mediate important physiological and pathophysiological processes, including the synchronization of the thalamocortical circuit and the generation of rhythmic neuronal activity during normal sleep (Huguenard and McCormick, 2007), spike-wave discharges in absence epilepsy (Huguenard and McCormick, 2007), and neuropathic pain (Bourinet et al., 2016). Additionally, mutations in T-type Ca 21 channel genes have been implicated in epilepsy, autism spectrum disorder, and schizophrenia (Weiss and Zamponi, 2020). There are three distinct Ttype Ca 21 channel genes: CACNA1G, CACNA1H, and CACNA1I, which encode the pore-forming a1 subunits known as Ca v 3.1, Ca v 3.2, and Ca v 3.3, respectively (Perez-Reyes, 2003). In contrast to the other voltagegated Ca 21 channels, T-type Ca 21 channels have unique features, including low-voltage activation, rapid voltagedependent inactivation, and the generation of transient currents that can trigger a brief burst of high-frequency APs (Llinás and Jahnsen, 1982;Perez-Reyes, 2003;Cheong and Shin, 2013). As a large proportion of T-type Ca 21 channels are inactivated at typical resting membrane potentials (V m ), neuron hyperpolarization can de-inactivate T-type Ca 21 channels and trigger a burst of APs, a phenomenon called "rebound bursting" (Llinás and Jahnsen, 1982;Perez-Reyes, 2003;Cheong and Shin, 2013). Burst firing can have unique functional consequences, including information encoding by enhancing signalto-noise ratio, facilitation of neuropeptide release, and increasing the reliability of synaptic transmission (Lisman, 1997;Schultz et al., 1997;van den Pol, 2012).
In this study, we identified a subpopulation of MHb neurons that are capable of firing high-frequency AP bursts. Although only ;7% of MHb neurons displayed burst firing at baseline, ;20% displayed burst firing following membrane hyperpolarization or with depolarization from a hyperpolarized V m . Burst firing from a hyperpolarized V m was blocked by the selective T-type Ca 21 channel antagonist Z944 (Tringham et al., 2012). We identified robust expression of Ca v 3.1 mRNA predominantly in the lateral MHb using RNAscope in situ hybridization, which co-localized with tachykinin-1 (Tac1) and choline acetyltransferase (ChAT), markers which define neurons of the dorsal and ventral MHb, respectively. These neurons have distinct anatomic projections to the interpeduncular nucleus and have unique roles in regulating anxiety-like and depressive-like behavior and nicotine dependence (Molas et al., 2017). Thus, MHb neuron burst firing, especially that mediated by T-type channel activation, is a previously unidentified mode of neuronal activity that could regulate diverse behaviors relevant to neuropsychiatric disorders.

Animals
All animal procedures were performed in accordance with the Medical College of Wisconsin animal care committee's regulations. C57BL/6J mice were given ad libitum access to food and water and housed four to five per cage in a temperature (23 6 1°C) and humidity-controlled room (40-60%) with a 14/10 h light/dark cycle. All experiments were performed on adult male or female C57BL/6J mice (8-10 weeks old). C57BL/6J mice were obtained from The Jackson Laboratory. Experiments were performed between zeitgeber time (ZT)8 and ZT14, where ZT0 is lights on.

Slice preparation and electrophysiology
Mice were anesthetized by isoflurane inhalation and decapitated. The brain was trimmed and embedded in 3% low-melting-point agarose, and coronal slices containing the MHb (200 mm thick) were cut using a vibrating slicer (Leica VT1200s). Slices were prepared in a NMDG-based solution containing the following: 92 mM NMDG, 2.5 mM KCl, 1.25 mM NaH 2 PO 4 , 26 mM NaHCO 3 , 20 mM HEPES, 25 mM glucose, 2 mM thiourea, 5 mM Na-ascorbate, 3 mM Na-pyruvate, 0.5 mM CaCl 2 ·2H 2 O, and 7 mM MgSO 4 (pH 7.3-7.4 with HCl). Artificial CSF (ACSF), containing the following: 119 mM NaCl, 3 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 , 1.25 mM NaH 2 PO 4 , 25 mM NaHCO 3 , and 10 mM glucose, was gradually spiked-in to the NMDG-containing solution after slice cutting in 5-min intervals for a total of 20 min at room temperature, similar to a previously described approach (Ting et al., 2018). Some slices were cut in a sucrose-based solution containing the following: 68 mM sucrose, 78 mM NaCl, 3 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 , 1.25 mM NaH 2 PO 4 , 25 mM NaHCO 3 , and 25 mM glucose. These slices were incubated in the same solution for 20 min after cutting. The slices were then allowed to recover for at least 1 h in ACSF. All solutions were saturated with 95% O 2 and 5% CO 2 .
Whole-cell patch-clamp recordings were made using patch-clamp amplifiers (Multiclamp 700B) under infrared differential interference contrast (DIC) microscopy. Data acquisition and analysis were performed using DigiData 1550B digitizers and the analysis software pClamp 10.7 (Molecular Devices). Signals were filtered at 2 kHz and sampled at 10 kHz. Recordings were performed in the presence of the AMPA receptor antagonist 6-cyano-7-nitroquinoxaline-2,3-dione disodium salt (CNQX; 10 mM) and GABA A receptor antagonist picrotoxin (50 mM) unless otherwise stated. For before-and-after comparisons, the selective T-type Ca 21 channel antagonist Z944 (3 mM), the NMDA receptor antagonists (RS)À3-(2-carboxypiperazin-4-yl)-propyl-1-phosphonic acid (CPP; 5 mM) or D-(-)À2amino-5-phosphonopentanoic acid (AP-5; 50 mM), or CNQX (20 mM) were bath perfused in ACSF in recordings where neurons fired apparent bursts, and current protocols that induced bursting before drug perfusion were repeated after drug perfusion. Ca 21 current recordings used two different voltage protocols to evoked Ca 21 currents. For determining the voltage dependence of activation, neurons were first held for 1 s at À90 mV, followed by a 320-ms step to progressively more depolarized holding potentials in 5-mV increments, beginning at À90 mV and finishing at À30 mV, followed by return to À90 mV holding for 500 ms. Neurons were held at À70 mV between sweeps. For determining the voltage dependence of inactivation, neurons were first held for 3.6 s at progressively more depolarized membrane potentials in 5-mV increments, beginning at À120 mV and finishing at À50 mV. After this initial holding potential, neurons were stepped to À50 mV for 320, followed by holding at À90 mV for 1 s. Neurons were held at À70 mV between sweeps. Glass pipettes (3-6 MX) for current clamp recordings were filled with an internal solution containing the following: 140 mM K-gluconate, 5 mM KCl, 10 mM HEPES, 0.2 mM EGTA, 2 mM MgCl 2 , 4 mM Mg-ATP, 0.3 mM Na 2 GTP, and 10 mM Na 2 -phosphocreatine (pH 7.2 with KOH). Glass pipettes for voltage clamp recordings of Ca 21 currents were filled with an internal solution containing the following: 135 mM tetramethyl ammonium (TMA)-OH, 10 mM EGTA, 2 mM MgCl 2 , and 40 mM HEPES, titrated to pH 7.2 with hydrofluoric acid (HF; Todorovic and Lingle, 1998). Series resistance (15-30 MX) was monitored throughout all recordings, and data were discarded if the resistance changed by .20%. Liquid junction potentials were not corrected for. All recordings were performed at 32 6 1°C using an automatic temperature controller (Warner Instruments).

RNAscope in situ hybridization
Mice were deeply anesthetized with isoflurane and transcardially perfused with 0.1 M sodium PBS followed by 4% paraformaldehyde in 4% sucrose-PBS (pH 7.4). After perfusion, the brain was removed and postfixed in the same fixative for 4 h at 4°C and was then dehydrated in increasing concentrations of sucrose (20% and 30%) in 0.1 M PBS at 4°C and frozen on dry ice. Coronal MHb sections (10 mm) were cut with a Leica cryostat and mounted on Superfrost Plus microscope slides (Fisher Scientific). Probes targeting the mRNA transcripts of Ca v 3.1 (target region: base pairs 1263-2886 of Mus musculus CACNA1G, NM_009783.3), Ca v 3.2 (target region: base pairs 2287-4243 of M. musculus CACNA1H, NM_021415.4), Ca v 3.3 (target region: base pairs 1259-2600 of M. musculus CACNA1I, NM_001044308.2), ChAT (target region: base pairs 1090-1952 of M. musculus ChAT, NM_009891.2), and Tac1 (target region: base pairs 20-1034 of M. musculus tachykinin 1, NM_009311.2) were designed by and purchased from Advanced Cell Diagnostics Inc. The experiment was conducted as per the manufacturer's instructions for the RNAscope Multiplex Fluorescent V2 Assay. Stained slides were mounted with ProLong Gold Antifade Mountant with DAPI (Invitrogen) and imaged on a Leica TCS SP8 confocal microscope. A probe targeting Bacilus subtilis protein DapB was used as a negative control, whereas probes targeting the ubiquitously expressed Polr2a, PPIB, and UBC served as positive controls.

Statistics and data analysis
Paired t tests were used to analyze bursting before and after drug perfusion in current clamp recordings. Twoway ANOVA with repeated measures was used to analyze the effect of Z944 on evoked Ca 21 currents. Expression of T-type channel mRNA was analyzed by one-way ANOVA on ranks for Ca v 3.1, Ca v 3.2, and Ca v 3.3 from MHb slices because of non-normally distributed data (Shapiro-Wilk p , 0.05), followed by Dunn's post hoc multiple comparison testing. Ca v 3.1 expression in the medial versus lateral MHb was analyzed by unpaired t test. The voltage-dependencies of activation and inactivation were described by the following Boltzmann functions:

Diverse electrophysiological states of MHb neurons
Cell-attached and whole-cell patch-clamp recordings were made to determine the activity state of MHb neurons at baseline. In whole-cell recordings, an assessment was made in the first few seconds to avoid alteration of cell state by the internal solution. MHb neurons exhibited a variety of distinct activity states at baseline. About half (66 of 130, 50.8%) of MHb neurons exhibited spontaneous, tonic AP firing at a frequency of ;2-10 Hz (Fig. 1A,F). About a third (43 of 130, 33.1%) were electrophysiologically silent (Fig. 1B,F). A minority (12 of 130, 9.2%) of MHb neurons exhibited depolarized low-amplitude membrane oscillations (DLAMOs; Fig. 1C,F), characterized by oscillatory fluctuations in membrane potential (V m ; Sakhi et al., 2014). Additionally, we observed a small population of neurons that exhibited high-frequency AP firing. At baseline, most spontaneously high-frequency AP firing MHb neurons fired AP doublets at a frequency of ;75-125 Hz at an average resting V m of À45.2 6 1.5 mV ( Fig. 1D; 8 of 130 total neurons, 6.2%; Fig. 1F), whereas one neuron spontaneously fired long trains of APs on top of a prominent depolarized plateau potential followed by an afterdepolarization ( Fig. 1E; 1 of 130 total neurons, 0.8%; Fig.  1F). The average intraburst frequency was 27.2 6 1.8 Hz, and the average interburst interval was 2.3 6 0.6 s in this neuron. The average resting V m in silent cells was significantly more hyperpolarized relative to that in tonic, silent, DLAMO, and depolarized doublet firing MHb neurons (Fig. 1G). Thus, MHb neurons display a variety of electrophysiological states at baseline, including high-frequency AP firing that has not been previously reported.

Burst firing from hyperpolarized membrane potentials
Although burst firing in MHb neurons has not been previously reported, neurons of the nearby LHb and thalamus display burst firing that resembles the hyperpolarized burst firing we observed, which is mediated by the activation of T-type Ca 21 channels (Wilcox et al., 1988;Perez-Reyes, 2003;Chang and Kim, 2004;Cheong and Shin, 2013;Lambert et al., 2014;Cui et al., 2018;Yang et al., 2018). T-type channels are a member of the low-voltage activated (LVA) family of voltage-gated Ca 21 channels (Perez-Reyes, 2003;Cheong and Shin, 2013). They can be activated by low-level membrane depolarization, though sustained depolarization leads to channel inactivation and can cause a switch to tonic firing (McCormick and Pape, 1990). We thus tested whether depolarization of MHb neurons from a hyperpolarized V m can trigger burst firing and whether prolonged depolarization causes a transition to tonic firing.
In a silent MHb neuron, a depolarization step (10 pA) from the resting V m (À63 mV) triggered high-frequency AP firing (AP frequency = 85 Hz) that resembled the hyperpolarized burst firing observed in Figure 1E, in that it fired a train of APs on top of a depolarized plateau potential and was followed by a large afterdepolarization ( Fig. 2A). When the V m was changed to À58 mV through constant current injection (10 pA), the same depolarization step (10 pA) triggered transient burst firing at 34 Hz, followed by tonic firing at 3.5 Hz. When the V m was further depolarized to À52 mV by 20-pA constant current injection, the 10-pA depolarization step triggered tonic firing at 3.8 Hz. Similarly, ramp depolarization (0 to 150 pA ramp, 5-s duration) from a resting V m of À75 mV triggered initial burst firing in an MHb neuron (Fig. 2B). As the V m became more depolarized during the ramp, burst firing transitioned to tonic firing. Burst firing remained when the V m was more hyperpolarized than approximately À55 mV but transitioned to tonic firing as the V m depolarized above this level. At a V m of À44 mV, only repetitive tonic firing was observed. In a different MHb neuron, initial burst firing to ramp depolarization was similarly observed from a resting V m of À68 mV, but burst firing was not observed with constant current injection to À57 or À53 mV before the ramp (Fig. 2C). Ramp depolarization-induced burst firing from a hyperpolarized V m (,À60 mV before ramp) triggered initial burst firing in 30 of 152 MHb neurons (19.7%). Burst firing from a hyperpolarized V m that is abolished by holding the neuron at depolarized potentials suggests that T-type channel activation contributes to the initial burst firing.
T-type channels are nearly completely inactivated when neurons are held at a V m .À60 mV (Perez-Reyes, 2003;Cheong and Shin, 2013). As the average V m of tonic neurons was À47.0 6 1.5 mV, it is possible that membrane hyperpolarization may convert these neurons to burst firing neurons. In a tonic firing cell with a resting V m of À40 mV, ramp hyperpolarization (0 to À20 pA, 5 s) converted tonic firing to burst firing (Fig. 3A). Tonic firing resumed after the V m spontaneously depolarized above À50 mV. Thus, MHb neuron firing states can be interchangeable depending on their V m .
We evaluated whether the offset of a prolonged hyperpolarizing current step could induce rebound burst firing. A series of hyperpolarizing current steps (0 to À50 pA, D50 ms) triggered rebound burst firing in an MHb neuron (Fig. 3B). Interestingly, as the hyperpolarization duration lengthened, the high-frequency rebound burst was converted to a single AP with a pronounced afterdepolarization, subsequently followed by a spontaneous, high-frequency AP burst. Similarly, progressively greater hyperpolarizing current steps (0 to À80 pA, D20 pA, 100 ms) triggered rebound burst firing (Fig. 3C). Rebound burst firing following the offset of a prolonged hyperpolarizing current step (!1 s, À50 pA) was observed in 19.3% (29 of 150) neurons recorded (Fig. 3D). Burst firing from a hyperpolarized V m could be observed in MHb neurons from all baseline activity states but was most commonly observed in tonic firing neurons (Fig. 3E). Thus, depolarization from a hyperpolarized V m , or hyperpolarization alone, can trigger burst firing in ;20% of MHb neurons.

T-type channel expression in the MHb
T-type channels are encoded by three genes: CACNA1G, CACNA1H, and CACNA1I, which encode the pore-forming a1 subunits Ca v 3.1, Ca v 3.2, and Ca v 3.3 (Perez-Reyes, 2003). We used RNAscope fluorescent in situ hybridization (Wang et al., 2012) to determine whether mRNA for T-type channels is expressed in the MHb and to determine the cell populations that express these Step depolarization triggered burst firing from a resting V m of À63 mV. With constant current injection to a V m of À58 mV, step depolarization triggered initial burst firing followed by tonic firing. With constant current injection to a V m of À52 mV, step depolarization triggered solely tonic firing. B, Ramp depolarization from a V m ,À60 mV triggered initial burst firing in 30 of 152 MHb neurons (19.7%). Tonic firing was observed as the V m depolarized above approximately À55 mV. C, Bursting was observed when the V m was À68 mV before ramp depolarization but was abolished with constant current injection to À57 or À53 mV before the ramp. channels. We found that mRNA for Ca v 3.1 was abundantly expressed in the MHb, including in both Tac11 and ChAT1 neurons (Fig. 4). Expression of Ca v 3.2 and Ca v 3.3 was also observed in the MHb in both Tac11 and ChAT1 neurons, albeit at significantly lower levels than Ca v 3.1 (Fig. 5A,B). Ca v 3.1 expression was significantly greater in the lateral MHb relative to the medial MHb (Fig.  5C). Thus, mRNA for T-type channels, in particular Ca v 3.1, are expressed in the MHb in both major neuron populations.

Voltage-gated Ca 21 currents in MHb neurons
Next, we tested whether voltage-gated Ca 21 currents could be evoked in MHb neurons, and we determined the voltage dependence of their activation and inactivation. Ca 21 currents were recorded similarly to previous studies (Joksimovic et al., 2017;Stamenic and Todorovic, 2018), using an internal solution that caused rapid rundown of high-voltage activated (HVA) Ca 21 currents (Todorovic and Lingle, 1998). Ca 21 current activation was studied using a step depolarization protocol where MHb neurons were stepped to progressively more depolarized holding potentials for 320 ms after initially holding neurons at À90 mV for 1 s (Fig. 9A). Rapid inward currents were evoked, and the conductance versus voltage relationship was fit with a Boltzmann function. The average voltage of half-activation (V 50 ) in the absence of Z944 was À47.9 mV. There was a significant effect of Z944 to The voltage dependence of Ca 21 current inactivation was determined by holding MHb neurons for 3.6 s at progressively more depolarized membrane potentials before a 320-ms step to À50 mV (Fig. 9B). Rapid inward currents were evoked when neurons were pre-held at hyperpolarized membrane potentials but were not evoked when preheld at more depolarized potentials. A plot of the current versus initial holding voltage was fit with a Boltzmann function. The average voltage of half-inactivation (V 50 ) was À92.6 mV. These voltages of half-activation and halfinactivation are consistent with those measured for Ttype channels in the rat subiculum using an identical internal solution and voltage protocols (Joksimovic et al., 2017). Collectively, these results indicate that evoked Ca 21 currents in MHb neurons are likely to be mediated by T-type channels.

Depolarized AP doublets
Eight of 130 total neurons (6.2%) spontaneously fired AP doublets and had an average V m of À43.9 6 1.5 mV (Fig.  1D). As T-type channels are likely completely inactivated at this V m , we aimed to further characterize the electrophysiological characteristics of depolarized AP doublet firing. In a depolarized (V m = À44 mV) neuron displaying both spontaneous AP doublets and single spikes with a large afterdepolarization, low-level ramp depolarization (5 pA, 5 s) increased AP doublet firing at a V m of À39 mV (Fig. 10A). In a different MHb neuron that spontaneously fired AP doublets at a V m of Figure 5. T-type channel expression in the MHb. A, RNAscope in situ hybridization for Ca v 3.1, Ca v 3.2, and Ca v 3.3 mRNA in the MHb with Tac1 and ChAT. B, Expression of Ca v 3.1 is significantly greater than Ca v 3.2 and Ca v 3.3 expression, and Ca v 3.3 expression is significantly greater than Ca v 3.2 in the MHb; pppp , 0.001. C, Ca v 3.1 expression is significantly greater in the lateral MHb than the medial MHb; pppp , 0.001.
Research Article: New Research À48 mV, membrane hyperpolarization (À50 pA, 300 ms) briefly converted depolarized AP doublet firing to tonic firing at a V m of À53 mV, followed by a return to AP doublet firing with spontaneous depolarization to a V m of À48 mV (Fig.  10B). Spontaneous AP doublets at depolarized membrane potentials were not blocked by perfusion with Z944 ( Fig.  10C; n = 5). Thus, AP doublet firing at sustained depolarized potentials does not require T-type channel activation.

Discussion
In this study, we found that a subpopulation of MHb neurons are capable of firing high-frequency AP bursts from a hyperpolarized V m that strongly resemble burst firing mediated by T-type Ca 21 channel activation (Llinás and Jahnsen, 1982;McCormick and Pape, 1990;Cheong and Shin, 2013;Lambert et al., 2014). This burst firing could be induced by rebounding from membrane hyperpolarization or depolarization from a hyperpolarized V m , whereas depolarization from a V m . À60 mV led to tonic AP firing or depolarized AP doublet firing. Using RNAscope in situ hybridization, we identified robust mRNA expression of the Ttype Ca 21 channel Ca v 3.1 primarily in the lateral MHb, where it co-localized with both ChAT and Tac1, markers which define the ventral and dorsal MHb, respectively (Molas et al., 2017). Expression of Ca v 3.2 and Ca v 3.3 was also observed in the MHb but was comparatively sparse. Hyperpolarization-induced and depolarization-induced burst firing from a hyperpolarized V m , and evoked Ca 21 currents, were blocked by bath perfusion of the selective Ca v 3 antagonist Z944. Thus, a subpopulation of MHb neurons are capable of high-frequency burst firing, in contrast to previous reports of exclusive tonic firing by MHb neurons, with burst firing from a hyperpolarized V m sensitive to T-type channel blockade.
Interestingly, single-cell RNA sequencing of the mouse habenula revealed that transcriptomically defined clusters of MHb neurons heterogeneously express Ca v 3.1 (Hashikawa et al., 2019). The highest Ca v 3.1 expression was observed in neuron clusters located in the lateral aspect of both the dorsal and ventral MHb and in clusters which express Tac1 or ChAT, consistent with our results. This study also identified mRNA for Ca v 3.3 in some MHb clusters. A previous autoradiographic ISH study concluded that mRNA for Ca v 3.1 was expressed in the LHb but not the MHb (Talley et al., 1999), and hyperpolarization-induced rebound burst firing mediated by T-type channels has been observed in the LHb Yang et al., 2018).
Voltage-gated Ca 21 channels fall into two broad categories: HVA and LVA channels (Perez-Reyes, 2003;Cheong and Shin, 2013). T-type Ca 21 channels are LVA channels, whereas the other families of voltage-gated Ca 21 channels, including L-type, P/Q-type, N-type, and R-type channels, are HVA channels. Unique properties of T-type Ca 21 channels differentiate them from the HVA family. In contrast to HVA channels, T-type Ca 21 channels Figure 6. Location of neurons that bursted from a hyperpolarized membrane potential. Neurons that bursted from a V m ,À60 mV were predominantly located in the lateral and central MHb. Map includes neurons that bursted to ramp depolarization and/or to step hyperpolarization. Figure 7. The selective T-type channel antagonist Z944 blocked ramp depolarization-induced burst firing from hyperpolarized membrane potentials. A, Representative traces demonstrating the effect of Z944 to block ramp-induced burst firing from a hyperpolarized V m . B, The number of bursts per ramp was significantly reduced following Z944 perfusion; ppp , 0.01. require less membrane depolarization to open and rapidly inactivate with prolonged depolarization, leading to lowthreshold Ca 21 spikes (LTS) that can trigger a burst of Na 1 and K 1 -mediated APs on top of the Ca 21 -mediated LTS. In neurons with a resting membrane potential of À60 mV, the majority of T-type Ca 21 channels are inactivated, such that membrane hyperpolarization alleviates inactivation and leads to rebound burst firing on subsequent depolarization (Perez-Reyes, 2003;Cheong and Shin, 2013). In more hyperpolarized neurons, depolarization can trigger burst firing, but sustained and/or stronger depolarization inactivates T-type Ca 21 channels and prevents burst firing (Perez-Reyes, 2003;Cheong and Shin, 2013).
We observed that ;20% of MHb neurons burst fired to ramp depolarization from a hyperpolarized V m or Figure 8. The selective T-type channel antagonist Z944 blocked rebound burst firing. Representative traces demonstrating the effect of Z944 to block rebound burst firing following 1.5-s (A) or 100-ms (D) hyperpolarization. The number of bursts and rebound AP frequency were significantly reduced following 1.5-s (B, C) or 100-ms (E, F) hyperpolarization after Z944 perfusion; pp , 0.05, ppp , 0.01. Figure 9. Voltage-gated Ca 21 currents in MHb neurons. A, top, Representative currents induced by step depolarization from À90 mV to progressively more depolarized test potentials in the absence of Z944. Bottom, Voltage dependence of activation for voltage-gated Ca 21 currents before and after Z944 perfusion. B, top, Representative currents evoked by step depolarization to À50 mV after different initial test potentials. Bottom, Voltage dependence of inactivation for voltage-gated Ca 21 currents. displayed rebound burst firing. This bursting was observed in neurons that originally displayed all types of activity states at baseline, indicating that although T-type channel-mediated bursting is rare at resting conditions, a larger population of MHb neurons are capable of burst firing. Interestingly, the LHb similarly has a low percentage of spontaneously bursting neurons at baseline in Sprague Dawley rats and C57BL/6 mice, whereas congenitally learned helpless rats and mice exposed to chronic restraint stress have a substantially greater percentage of LHb neurons that burst at baseline . It is possible that the percentage of MHb neurons displaying different activity states may be altered in behavioral states regulated by the MHb, such as elevated anxiety-like states or nicotine withdrawal.
We also found that some neurons burst fired during ramp depolarization but did not exhibit rebound bursting following 100-ms or 1.5-s hyperpolarization. This is likely because depolarization following hyperpolarization is necessary for Ttype Ca 21 channel activation. Hyperpolarization-activated cyclic nucleotide-gated (HCN) channels, in particular HCN3 and HCN4, are expressed in the MHb in rodents (Monteggia et al., 2000;Santoro et al., 2000;Notomi and Shigemoto, 2004;Görlich et al., 2013). Consistent with this, depolarizing current sag was frequently observed with 1.5-s hyperpolarization (Fig. 3D), which is a hallmark of HCN channel activation (Biel et al., 2009). It is thus likely that HCN channel activation can provide the rebound depolarization necessary to activate T-type Ca 21 channels in some MHb neurons.
In addition to burst firing from a hyperpolarized V m , some neurons displayed high-frequency AP doublet firing at depolarized potentials (average V m = À45.2 6 1.5 mV). This AP doublet firing was not affected by T-type channel blockade. This is unsurprising, as T-type channels are likely completely inactivated at these depolarized potentials. The ionic conductances mediating this depolarized burst firing remain unknown. In neurons in other brain regions, persistent Na 1 current can contribute to afterdepolarization and promote burst firing (Azouz et al., 1996;Brumberg et al., 2000;Su et al., 2001). It is possible that persistent Na 1 current or other ionic conductances mediate this depolarized burst firing.
In summary, we demonstrated that a subpopulation of MHb neurons are capable of high-frequency AP firing, in contrast to previous reports that MHb neurons solely fire tonic APs at a frequency of ;2-10 Hz. Two distinct modes of high-frequency AP firing were observed, including T-type channel-mediated bursting and T-type channel-independent AP doublets. T-type channel-mediated bursting occurred from a previously hyperpolarized V m , whereas T-type channel-independent bursting remained with sustained depolarization and was not blocked by Z944. Ca v 3 mRNA was expressed in the MHb, most notably Ca v 3.1 in the lateral MHb, consistent with the predominant location of bursting MHb neurons. Thus, T-type Ca 21 channels may contribute to the regulation of behaviors relevant to neuropsychiatric disease by shaping the frequency and pattern of MHb neuron AP firing.