Nicotine regulates activity of lateral habenula neurons via presynaptic and postsynaptic mechanisms

There is much interest in brain regions that drive nicotine intake in smokers. Interestingly, both the rewarding and aversive effects of nicotine are probably critical for sustaining nicotine addiction. The medial and lateral habenular (LHb) nuclei play important roles in processing aversion, and recent work has focused on the critical involvement of the LHb in encoding and responding to aversive stimuli. Several neurotransmitter systems are implicated in nicotine’s actions, but very little is known about how nicotinic acetylcholine receptors (nAChRs) regulate LHb activity. Here we report in brain slices that activation of nAChRs depolarizes LHb cells and robustly increases firing, and also potentiates glutamate release in LHb. These effects were blocked by selective antagonists of α6-containing (α6*) nAChRs, and were absent in α6*-nAChR knockout mice. In addition, nicotine activates GABAergic inputs to LHb via α4β2-nAChRs, at lower concentrations but with more rapid desensitization relative to α6*-nAChRs. These results demonstrate the existence of diverse functional nAChR subtypes at presynaptic and postsynaptic sites in LHb, through which nicotine could facilitate or inhibit LHb neuronal activity and thus contribute to nicotine aversion or reward.


Results
Nicotine excites LHb neurons. As reported previously 26,27 , there is a clear border between the LHb and the adjacent MHb (Figs 1A1 and 2). We thus distinguished the MHb and LHb neurons according to the clear border. In addition, consistent with previous reports, we observed that LHb neurons were loosely dispersed 26 , had heterogeneous morphological but similar membrane properties 27 , and were capable of producing burst firing 26,28 .
In the current study, we recorded from ~750 LHb neurons. Nicotine increased firing in both cell-attached ( Fig. 1B-D) and whole-cell modes (Fig. 1F). Nicotine accelerated spontaneous firing of 77/95 LHb neurons in a dose-dependent manner, with an EC 50 of 610 nM (F 6,87 = 15.2, p < 0.001), and reached the plateau at 10 μ M (Fig. 1B-E). Interestingly, upon bath application of 100 nM nicotine, firing first decreased (to 90.4 ± 3.8% of baseline, p < 0.05; Fig. 1D), then sharply accelerated (p < 0.001), returning to baseline level after washout (Fig. 1B-D). To confirm that the site of nicotine action was within the LHb, recordings in a subset of experiments were made from slices with MHb removed (Fig. 1A3). Nicotine acceleration of LHb neuron firings in these slices (F 2,24 = 19.9, p < 0.001; post hoc p < 0.001, Fig. 1E blue triangle) was similar to that with MHb (p > 0.5 with MHb vs without MHb).
Apart from direct effects on LHb neurons, bath-applied nicotine could act indirectly by stimulating cholinergic MHb neurons (~40% of which express α 6*-nAChR mRNA 32 ) that project to the LHb. To minimize indirect effects through MHb, we locally applied ACh (30 μ Μ ) onto the somata of recorded LHb neurons by pressure injection from a second micropipette 33 . In the presence of atropine, ACh induced an inward current (25.6 ± 7.1 pA) in 6/10 neurons tested (p = 0.005) which was significantly depressed by MII [H9A; L15A] (to 33.0 ± 4.0% of ACh, p = 0.016; Fig. 2E). These experiments confirmed that there was a direct, MII [H9A; L15A]-sensitive nAChR (likely α 6*-nAChR)-mediated depolarization of LHb cells. Indirect effects via MHb were also unlikely since nicotine depolarized LHb cells (Figs 1F-H and 2A-D) in the presence of TTX and glutamate and GABA antagonists, which would block the corresponding synaptic inputs 11,34,35 from MHb neurons, and also because nicotine activated LHb neurons in brain slices where MHb was cut off (Fig. 1E).

Nicotine enhances synaptic potentials and currents (IPSCs and EPSCs) in LHb neurons.
We next examined whether nicotine would alter inhibitory and excitatory post-synaptic currents (IPSCs and EPSCs, respectively) in LHb neurons. We first examined how nicotine affects paired-pulse transmission when two PSCs were evoked at a brief inter-stimulus interval (50 msec). Nicotine (10 μ Μ ) selectively enhanced the first IPSC of each pair (IPSC 1 ; p < 0.05; Fig. 3A,C), thus reducing the paired-pulse ratio (PSC 2 /PSC 1 = PPR) (p < 0.01; Fig. 3D). Similarly, nicotine enhanced only EPSC 1 of each pair (p < 0.01; Fig. 3B-D) and reduced the EPSC-PPR (p < 0.05). Since decreased PPR can reflect increased transmitter release at the synapse, these results suggest that nicotine might potentiate synaptic transmission in LHb by increasing presynaptic release of both glutamate and GABA.
To better understand the impact of nicotine on GABA and glutamate signaling in the LHb, we then tested the effects of nicotine on spontaneous synaptic currents. Nicotine induced a transient shift to higher frequencies and larger amplitudes of spontaneous IPSCs (sIPSCs) (Fig. 3E-G), with an EC 50 of 40 nM for frequency (F 6,98 = 9.2, p < 0.001; Fig. 3N, black line) and 10 nM for amplitude (F 6,98 = 3.9, p = 0.002; Fig. 3O, black line). By contrast, nicotine induced a slowly developing increase in spontaneous EPSC (sEPSC) frequency (Fig. 3H,I) with an EC 50 of 140 nM (F 6,105 = 8.9, p < 0.001; Fig. 3N, blue line), and a small but significant increase in sEPSC amplitude (p < 0.001, Kolmogorov-Smirnov test; Fig. 3J) with an EC 50 of 210 nM (F 6,105 = 4.9, p < 0.001; Fig. 3O, blue line). At a holding potential of − 40 mV, sEPSCs and sIPSCs appeared as distinct inward and outward currents, and nicotine (1 μ Μ ) induced a transient outward current followed by a sustained inward current and increased frequency of sIPSCs and sEPSCs (Fig. 3K,M). Similar changes were seen under current clamp, with an initial hyperpolarization followed by depolarization and increased frequency of sIPSPs and sEPSPs (Fig. 3L,M). Thus, nicotine has We also examined the impact of nicotine on miniature synaptic currents, where spontaneous release was determined in the presence of TTX, Ca 2+ -free ACSF (where CaCl 2 was replaced by MgCl 2 ) or in presence of 100 μ M  cadmium chloride (CdCl 2 , a non-selective blocker of voltage-gated calcium channels). All of these treatments allow determination of the effects of nicotine on synaptic release in the absence of possible effects on presynaptic action potential generation (TTX) or activity-dependent effects on release (calcium-free or cadmium-containing media). Nicotine caused a smaller but still significant increase in the frequency of spontaneous miniature EPSCs (F 3,77 = 5.8, p = 0.001; post hoc p = 0.004) and IPSCs (F 3,66 = 10.2, p < 0.001; post hoc p < 0.001) in presence of TTX (Fig. 4A1,B1,C), in Ca 2+ -free ACSF (EPSC: p = 0.042; IPSC: p < 0.001; Fig. 4A2,B2,C) or in presence of CdCl 2, EPSC: p = 0.014; IPSC: p = 0.005; Fig. 4A3,B3,C), when compared with the effects of nicotine on spontaneous release when presynaptic activity is allowed to occur. In addition, both miniature EPSC (F 3,77 = 5.3, p = 0.002; Fig. 4D) or IPSC (F 3,66 = 7.6, p < 0.001; Fig. 4D) amplitudes were significantly reduced in the presence of TTX (EPSC: p = 0.003; IPSC: p < 0.001) or Ca 2+ -free ACSF (EPSC: p = 0.046; IPSC: p = 0.018). These results are consistent with both presynaptic and postsynaptic actions of nicotine at excitatory and inhibitory synapses, increasing glutamatergic and GABAergic release, and that these actions depended in part on calcium channels.
Mechanism of nicotine's biphasic action on LHb neurons. As described in Fig. 1, upon the application of 100 nM nicotine, LHb firing was first decreased; and then sharply accelerated, returning to baseline after washout (Fig. 1A-C). We proposed that the decrease of firing was mediated by nicotinic potentiation of GABAergic transmission, and the enhancement of firing was mediated in part by nicotinic potentiation of glutamatergic transmission. To test this hypothesis, we compared the effects of 100 nM nicotine in the absence and presence of a GABA A receptor antagonist (20 μ M gabazine), glutamate receptor antagonists (20 μ M DNQX plus 50 μ M AP5), or a cocktail of both GABA A and glutamate receptor blockers (gabazine + DNQX + AP5).
As shown in Table 1, nicotine-induced inhibition of firing was completely abolished (F 3,28 = 7.7, p < 0.001) by gabazine (p = 0.042) or by the cocktail (p = 0.011), but not by DNQX + AP5 (p > 0.5). In contrast, nicotine-induced enhancement of firing was partly attenuated (F 3,34 = 5.9, p = 0.002) by DNQX + AP5 (p = 0.013) or by the cocktail (p = 0.04), but not by gabazine alone (p > 0.5). Thus, the ability of nicotine to increase and decrease firing was mediated by nicotine enhancement of IPSCs and EPSCs, respectively, although the excitatory effect of nicotine also involved mechanisms other than glutamate receptors.

Discussion
Converging evidence indicates that the LHb is activated by aversive stimuli 36 . Nicotine can have strong aversive effects 37 , which can play an important role in promoting nicotine addiction 1,2 . In keeping with this, we found that nicotine strongly activated LHb neurons in vitro, as has been observed in vivo 38 ; in vivo, LHb activation of RMTg neurons would in turn inhibit midbrain dopamine (DA) neurons and thus contribute to aversion 12,39 . In addition, we found that the actions of nicotine in LHb brain slice were mediated by distinct nAChR subtypes. Pharmacology and α 6*-knockout experiments suggested that α 6*-nAChRs played a prominent role in the sustained enhancement of LHb firing and glutamate release, while DHβ E-sensitive nAChRs, likely reflecting α 4β 2-nAChRs, were important for the transient increase in GABAergic transmission. Our results show that different nAChR subtypes played a significant role in the regulation of LHb activity, which may contribute to both nicotine aversion and reward.
At different concentrations found in the blood of human smokers (25-444 nM) 31 , nicotine acted at diverse nAChR subtypes to modulate LHb activity in brain slices from rats. Low concentrations (EC 50 20-50 nM) potentiated IPSCs via DHβ E sensitive (α 4β 2) nAChRs, leading to rapid but only brief hyperpolarization, consistent with α 4β 2-nAChR's propensity to desensitization [40][41][42] . Higher nicotine concentrations (EC 50 ~200 nM) elicited a more sustained increase in glutamate release via MII-sensitive (α 6* ± α 3β 2)-and MLA-sensitive (α 7)-containing nAChRs; pharmacology and knockout experiments combined suggest a particular role for α 6*-nAChRs in nicotine excitation of LHb neurons. Nicotine's actions on synaptic transmissions depended in part on calcium and TTX-sensitive sodium channels. At higher concentrations (EC 50 ~400 nM), nicotine directly depolarized LHb neurons via MII-and DHβ E-sensitive nAChRs but not by MLA-sensitive nAChRs. Similarly, the nicotine-induced increase in firing was sensitive to Μ Ι Ι , MLA and DHβ E. The EC 50 of nicotine for enhancing firing (~600 nM) was about 1.5 fold greater than the EC 50 of the nicotine-induced current responses. The difference between the EC 50 for nicotine enhancement of firing and depolarization may reflect the presence of different nAChRs at presynaptic terminals versus postsynaptic neurons in LHb.
Genetic variations in the CHRNA6-CHRNB3 gene cluster increase vulnerability to tobacco smoking [43][44][45] . α -CTx-MII infusion into α 6*-nAChRs-expressing mesolimbic regions (midbrain and nucleus accumbens) [46][47][48] decreases nicotine self-administration by rats 21,25 . Our findings show that the LHb is another critical brain region for α 6*-nAChR-mediated actions for nicotine. First, global disruption of α 6-nAChRs increase basal mEPSC frequency as well as firing rate, indicating α 6-nAChRs is a critical regulator of LHb neuronal activity. Second, although the MHb also expresses many nAChRs 32 and projects to LHb 26 , the effects we observed probably originated mainly in the LHb, since α 6*-mediated LHb currents were elicited by local applications of ACh or by bath applications of nicotine in the presence of TTX, and persisted when the MHb was cut away from the LHb. Moreover, nicotinic excitation of MHb neurons in vitro 49 does not require α 6*-nAChRs 8 . Previous immunoreactivity studies 28 suggest that α 6-nAChRs are mainly expressed in the ventral inferior MHb, with little expression in the LHb. In the current study, using the patch-clamp electrophysiology technique, one of the most sensitive ways to detect functional activity of receptors, we were able to detect that nicotine modulated physiological activity of LHb neurons through action of nAChRs containing the α 6 subunit.
Nicotine has both negative and positive motivational effects which contribute to its abuse potential 1 . Here we demonstrate that nicotine, acting through α 6*-nAChRs, depolarizes LHb neurons and enhances glutamatergic signaling. Manipulating LHb activity, especially through α 6*-nAChRs, may therefore be of great value in treating nicotine addiction. In addition, we have shown that, over a wide range of concentrations (10 nM-100 μ M), nicotine robustly activates LHb neurons. In contrast, nicotine at < 100 nM has no significant effects on DA neurons in the ventral tegmental area (VTA); only at ≥ 250 nM does nicotine increase DA neuronal firing and burst activity 50,51 . Interestingly, once sufficient nicotine concentrations are reached, nicotine has a stronger effect on VTA-DA cells than on LHb neurons. For example, we found that nicotine at 1 μ M, a concentration within the range of peak blood levels in smokers, increased VTA-DA neuronal discharges by an average of 2.32 ± 0.4 (n = 7) fold in vitro (see also ref. 52). This is in agreement with other studies 53 , and was significantly greater than the 1.52 ± 0.07 (n = 13) fold increase in firing of LHb neurons (p = 0.017, unpaired t-test). Overall, the finding that nicotine at low concentrations activates LHb but not DA cells indicates that the LHb is exquisitely sensitive to the aversive effects of nicotine.
In addition to the LHb directly encoding aversive stimuli, activation of LHb neurons may contribute to aversion by inhibiting DA neurons indirectly via the RMTg 54,55 . Like the LHb, the RMTg is activated by aversive stimuli. Also, our findings that nicotine excites LHb glutamatergic neurons agree with recent studies showing that nicotine increases presynaptic glutamate release onto RMTg neurons 12,39 . Furthermore, a previous electrophysiological study in anesthetized rats showed that systemic administration of nicotine led to VTA-DA neuron activation which was preceded by a short-lasting inhibition 56 . We speculate that this inhibition of DA neurons could be caused by nicotine-induced excitation of LHb neurons, which activates the RMTg to inhibit DA cells. Emerging data also demonstrates that activation of the LHb-to-RMTg pathway induces acute avoidance 9 , and that activation of the Hb-IPN pathway reduces nicotine intake 57 . The Hb and its targets may also be important for symptoms of nicotine withdrawal, since inhibition of nAChRs in the Hb attenuates withdrawal symptoms 6 . In agreement, chronic nicotine exposure causes a hypodopaminergic state 58 , which may mediate the effects of nicotine withdrawal. Thus, reducing Hb-mediated inhibition of VTA DA neurons could be a useful strategy to relieve these withdrawal symptoms. These findings indicate that the LHb likely plays a critical role in different aspects of nicotine addiction.
Taken together, our data suggest that different subtypes of functional nAChRs are situated both presynaptically on local afferents innervating LHb neurons and postsynaptically on the somatodendritic regions of these neurons. By activating these receptors, nicotine likely enhance the output of LHb neurons and, in this way, contribute to the aversive properties of nicotine. Thus, direct nicotine action within the LHb via the pathways we have identified may represent a mechanism through which nicotine-related aversion could contribute to nicotine addiction.  [16][17][18][19][20][21][22][23]. As data obtained from either sex did not differ significantly, they were pooled.

Methods
Brain slice preparation and electrophysiology. Brain slices were prepared as described earlier 59 .
Spontaneous firing was recorded by the loose-patch cell-attached technique, allowing long-lasting recordings without perturbing the cytoplasmic contents, and in whole-cell mode to measure membrane potential and input resistance. Currents-induced by nicotine or acetylcholine (ACh) were recorded in TTX, AP5, DNQX, gabazine, and strychnine. Compounds were applied by bath perfusion, except acetylcholine (30 μ M, with 0.5 μ M atropine to block muscarinic receptors) which was ejected by pressurized air from a micropipette placed near the recorded cell 33 .
Data Analysis and Statistics. Baseline electrophysiological data were recorded for 5-10 min, before drug superfusion and during the washout. To calculate the percent change in EPSCs/IPSCs/firing frequency or amplitude for a given cell, recordings during the initial control period were averaged and normalized to 100%. eIPSC and eEPSC amplitudes were calculated by averaging the peak current from six sweeps during baseline and during each drug application. For the measurement of inward currents induced by bath application of nicotine, traces were filtered at 10 Hz, and the means of 30-s baseline before nicotine application and 30 s during the maximal effects of nicotine were calculated and subtracted to give the magnitude of the nicotine current. The peak of nicotinic currents induced by puffing ACh onto neuronal somata was measured after the traces were filtered at 500 Hz. All values given in the text and figures indicate mean ± S.E.M. Statistical significance was assessed by a two-tailed Student's t test, a one-way ANOVA with a Tukey's post hoc test for multiple group comparisons, or a Kolmogorov-Smirnov (K-S) test. 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 nicotine, α the slope parameter, and x o the nicotine concentration which induces a half-maximal change.