Downregulation of M-channels in lateral habenula mediates hyperalgesia during alcohol withdrawal in rats

Hyperalgesia often occurs in alcoholics, especially during abstinence, yet the underlying mechanisms remain elusive. The lateral habenula (LHb) has been implicated in the pathophysiology of pain and alcohol use disorders. Suppression of m-type potassium channels (M-channels) has been found to contribute to the hyperactivity of LHb neurons of rats withdrawn from chronic alcohol administration. Here, we provided evidence that LHb M-channels may contribute to hyperalgesia. Compared to alcohol naïve counterparts, in male Long-Evans rats at 24-hours withdrawal from alcohol administration under the intermittent access paradigm for eight weeks, hyperalgesia was evident (as measured by paw withdrawal latencies in the Hargreaves Test), which was accompanied with higher basal activities of LHb neurons in brain slices, and lower M-channel protein expression. Inhibition of LHb neurons by chemogenetics, or pharmacological activation of M-channels, as well as overexpression of M-channels’ subunit KCNQ3, relieved hyperalgesia and decreased relapse-like alcohol consumption. In contrast, chemogenetic activation of LHb neurons induced hyperalgesia in alcohol-naive rats. These data reveal a central role for the LHb in hyperalgesia during alcohol withdrawal, which may be due in part to the suppression of M-channels and, thus, highlights M-channels in the LHb as a potential therapeutic target for hyperalgesia in alcoholics.


Results
Hyperalgesia in ethanol-withdrawn rats is accompanied with hyperactivity of LHb neurons in brain slices. Changes in nociceptive sensitivity were measured in rats that have been drinking alcohol for eight weeks in the intermittent access to alcohol in a two-bottle free choice (IA2BC) paradigm. We have previously showed that rats in this paradigm escalate their alcohol consumption and show withdrawal syndrome when alcohol is discontinued 23,24 . The thermal pain sensitivity was significantly higher in rats during 24 h of withdrawal (EtOH-WD) than that of ethanol naive counterparts (CTRL group) (as measured by paw withdrawal latency (PWL) in the Hargreaves test) (Unpaired t-test, t = 3.9, p = 0.0003, Fig. 1A). In parallel, the basal rate of spontaneous spiking was significantly higher in LHb neurons in brain slices of EtOH-WD rats than that from CTRL rats (Unpaired t-test, t = 3.789, p = 0.0004) (Fig. 1B-D).
M-channels in the LHb of ethanol-withdrawn rats are suppressed, and activation of LHb M-channels alleviates hyperalgesia. The above results demonstrate that inhibition of LHb excitatory neurons alleviates hyperalgesia in EtOH-WD rats. To investigate the underlying mechanisms, we examined the m-type potassium channels (M-channels). M-channels are abundantly expressed in the LHb 28 and are involved in numerous physiological and pathological processes 29 , and may contribute to pain 30 . We previously demonstrated that the expression of KCNQ2/3, the m-channel subunits and the sensitivity to M-channel blocker, XE991, in the LHb of ethanol-withdrawn rats, are reduced 22 . To test whether LHb M-channels are involved in hyperalgesia, we first examined the expression of M-channels in the LHb of rats during 24 h withdrawal. We found that the protein levels of both KCNQ2 and KCNQ3 were significantly decreased (Fig. 4A,B; For KCNQ2, Unpaired t-test, t = 2.713, p < 0.0218; For KCNQ3, Unpaired t-test, t = 6.178, p = 0.0001) and the reduction of KCNQ3 was greater (KCNQ2 vs. KCNQ3; Paired t-test, t = 2.703, p = 0.0426). We then examined the effect of the M-cannel activator, retigabine, an agent that could reduce firing of LHb neurons in ethanol withdrawn-rats 22 . Intra-LHb infusion of retigabine at a dose that decreases ethanol consumption in rats 22,31 , significantly increased the PWL in EtOH-WD rats (Two-way ANOVA, for Drug F 2,35 = 7.683, P = 0.0017, for Time F 4,140 = 6.256, P = 0.0001, post-hoc: **p < 0.01 vs. aCSF, N rat = 7-19/group, Fig. 4C,D). This increase was attenuated by intra-LHb infusion of ICA-27243 (10 ng in 200 nl/side), a specific KCNQ2/3 activator 32 , (post-hoc: ### p < 0.001, # p < 0.05 vs. aCSF; p > 0.05, Retigabine vs. ICA-27243), suggesting that the KCNQ2/3 channels in the LHb is the major player in M-channel modulation of nociception.

Discussion
In this study, we demonstrated that manipulation of LHb neuronal activity altered thermal pain sensation in rats withdrawn from intermittent voluntary alcohol drinking. Specifically, silencing LHb excitatory neurons attenuated thermal hyperalgesia in EtOH-WD rats, whereas chemogenetic activation of these neurons provoked hyperalgesia in ethanol-naïve rats. We also confirmed our published finding of reduced expression of M-channel subunits in the LHb of EtOH-WD rats. We further showed that pharmacological activation or genetic upregulation of M-channels in the LHb alleviated hyperalgesia in EtOH-WD rats. These results suggest that suppression  , rats infected with hM3Dq showed a decreased paw withdrawal latency than those infected with eGFP. Two-way ANOVA, for group F 2,30 = 3.583, P = 0.0402, for Time F 4,120 = 4.534, P = 0.0019, post-hoc: *p < 0.05, **p < 0.01 vs GFP + CNO; # p < 0.05, ## p < 0.01 vs. hm3Dq+Veh; N rat = 9-12/group. www.nature.com/scientificreports www.nature.com/scientificreports/ of M-channels in the LHb contributes to the hyperactivity of LHb neurons and hyperalgesia of EtOH-WD rats. Thus, the LHb plays a key role in both pain and alcohol-reward processing.
We also identified molecular adaptations in LHb neurons during ethanol withdrawal. Ion channels are chief modulators of neuronal excitability. Among them, potassium channels act as a break for overflow firing. Consistent with previous findings that expression of M-channel subunits, KCNQ2/3, is reduced in some brain regions of rodents that chronically exposed to ethanol 22,35 , the current study showed that the KCNQ2/3, which are abundantly expressed in the LHb 22,28 , were reduced in EtOH-WD rats. The hyperalgesia observed in EtOH-WD rats was attenuated by intra-LHb infusion of M-channel activators or by overexpression of KCNQ3 in the LHb. Although M-channels have been identified as a modulator of pain transmission in the peripheral neuronal pathway 30,36 , our study shows a role of M-channels in the central nervous system (CNS) in pain sensation. Human brain imaging and animal studies have shown that the LHb undergoes structural and functional changes with pain and long-lasting hypersensitivity 18,20,37 . Chronic pain is thought to be caused by aberrant neuronal responses along the pain transmission pathway from the dorsal root ganglion to the spinal cord, thalamus, and cortex 38,39 . The sensitization of peripheral nociceptors leads to hyperalgesia. However, the functional and morphological changes in the CNS also significantly contribute to the maintenance of chronic pain and its comorbidities. Thus, the crosstalk between sensory and affective components has been suggested to contribute importantly to the relationship of pain and psychiatric conditions. The LHb has been implicated in promoting behavioral avoidance due to its role in aversion 40,41 .
The major contribution of this study is the identification of LHb excitatory neurons as regulators of thermal hypersensitivity. We took advantage of chemogenetic approaches to selectively regulate LHb excitatory neurons under the control of a CaMKIIa promoter. Compared with traditional excitotoxic lesions and pharmacologic interventions, chemogenetic approaches have the advantage of greater ability to transiently regulate a specific population of neurons 42 . Our observation that silencing LHb excitatory neurons reduced thermal hypersensitivity is consistent with previous pharmacologic intervention studies 9,21 .
Additionally, we showed apparent hyperalgesia in naïve rats in which the LHb was activated chemogenetically, as shown by the reduced PWL in the Hargreaves test. This finding indicates that the LHb has an important role in nociception. Recently, the LHb has been gaining more attention because of its role in psychiatric disorders such as anxiety and depression. Evidence has indicated an association between the LHb and depressive-like behaviors www.nature.com/scientificreports www.nature.com/scientificreports/ in a mouse model of neuropathic pain 18 and a rat model of alcohol drinking 43 . Moreover, patients suffering from chronic pain also tend to suffer from depression, and importantly, patients who are diagnosed with depression also commonly suffer from chronic pain. Thus, depression is considered a positive predictor of the development of chronic pain 44 . In this view, LHb may act as a general hub for the connection between pain and depression.
How does the LHb regulate thermal sensitivity? Although the detailed nociceptive circuitry, which allows the LHb to signal to the dorsal horn of the spinal cord to regulate thermal sensitivity, has not yet been established, it may involve several potential signal cascades. Central nociceptive regulation can be achieved through a descending pathway. This pathway starts from the periaqueductal gray (PAG) to the rostral ventromedial medulla (RVM) and reaches to the superficial laminae of the spinal dorsal horn (SDH) (referred to as PAG-RMV-SDH pathway) 45,46 . This pathway comprises an essential neural circuit that exerts powerful modulatory influences on pain 47 . PAG has been shown to mediate thermal hyperalgesia in alcohol-dependent rats 48 . The LHb may regulate the thermal sensitivity 20 through its projections to the PAG 49 .
Also, morphological, behavioral and electrophysiological evidence suggests the involvement of several neurotransmitters in the analgesia pathway under both physiological and pathological conditions [50][51][52] . In the central analgesic system descending serotonergic pathways, serotonergic neurons in the raphe nuclei have been considered an important component 53 . Its descending projections, either directly or via the nucleus raphe magnus, modulate the responses triggered by noxious stimulation of the spinal dorsal horn neurons 54 . The LHb has a robust projection to the dorsal raphe (DR) and modulates the activity of DR neurons 14,55,56 . Therefore, withdrawal-induced changes in LHb-raphe circuits may also have an important role in the modulation of thermal responses to noxious stimuli. Further studies are needed to test these possibilities.
Whether pain sensitivity alters alcohol drinking or alcohol drinking alters pain sensitivity through the habenula is still an open question. Alcohol intake by the rats in the IA2BC paradigm was increased after one week in the paradigm (three sessions) but the increase in pain sensation was more apparent after four weeks of alcohol exposure (data not shown), which suggest that, at least in the IA2BC paradigm, repeated alcohol drinking and abstinence seem to alter pain sensitivity. To understand this better, in the future study, we will activate habenula that has been shown to increase pain sensitivity and check whether this will lead to increase alcohol consumption.
We manipulated the activity of LHb neurons in free moving rats by a combination of AAV-driven overexpression of hM3Dq or hM4Di DREADDs in the LHb and systemic administration of CNO. Notably, CNO www.nature.com/scientificreports www.nature.com/scientificreports/ induced a significant increase in c-fos expression in the LHb in ethanol-naïve rats that expressed hM3Dq in the LHb, suggesting that systemic administration of CNO activates LHb neurons. However, we should interpret these results with caution, since a recent study indicates that clozapine, a metabolite of CNO, mediates the effects of CNO on DREADD receptors in the brain 57 , although CNO is the most well-characterized and advisable ligand for DREADDs until other selective ligands are fully characterized 58 . To avoid possible off-target effects of CNO-driven clozapine via non-DREADDs endogenous receptors, we checked CNO effects in rats without DREADDs expression. Further control validations such as dose-dependent responses and more control groups such as low-dose clozapine treatment should be performed in future studies.
To modulate neuronal activity, we employed the DREADDs strategy "in general". Although DREADDs GPCRs are genetically modified based on muscarinic receptors, one of which can suppress the M-current via Gq signaling pathway 59 , the DREADDs may also have different cellular signal cascades via various potassium channels to change neuronal activities. Given that M-channels are slowly closing at close-to-resting membrane potentials (RMP) and modulating afterhyperpolarization (AHP) [60][61][62] and that repeated alcohol exposure reduces the AHP of the LHb 22 , M-channels of LHb excitatory neurons may play a key role in the adaptation induced by ethanol withdrawal. However, the adaptation may also involve other potassium channels. For example, it has been demonstrated that Kir4.1, one of the inwardly rectifying potassium channels that is predominantly localized in non-neuronal glial cells 63 and has been known to be reduced in reward-related brain regions after repeated alcohol exposure 35 , modulates the pathological firing patterns of LHb neurons 64 . Therefore, it will be of interest to further investigate the changes in other potassium channels including other KCNQ subunits and their co-factors in the LHb during alcohol withdrawal.
We provided several lines of evidence about the impact of the LHb on hyperalgesia in rats withdrawn from chronic voluntary ethanol drinking. The increased sensitivity to thermal stimuli was concomitant with the increased basal firing rate of LHb neurons and the inhibition of M-channels. Pharmacological activation of LHb M-channels or chemogenetic inhibition of LHb neurons reduced hyperalgesia and ethanol intake. Conversely, chemogenetic activation of LHb neurons increased pain sensitivity in ethanol-naïve rats. Collectively, these findings suggest that M-channels play an important role in the hyperactivity of LHb neurons of ethanol-withdrawn rats, and LHb activity plays a crucial role in hyperalgesia and relapse-like drinking. Manipulating LHb activity through M-channels may be of therapeutic value in the treatment of hyperalgesia and relapse drinking.

Materials and Methods
Animals. All studies were conducted on adult male Long Evans rats (n = 157). All procedures and methods were performed by the National Institutes of Health guidelines and regulations with the approval of the Institutional Animal Care and Use Committee of Rutgers, the State University of New Jersey, New Jersey Medical School. The rats were individually housed in ventilated Plexiglas cages in climate-controlled rooms (20-22 °C). The rats had unlimited access to food and water (or as otherwise indicated) and acclimatized to the housing conditions and handling before the start of the experiments. They were kept on a 12 h light/dark cycle.

Intermittent access to 20% ethanol two-bottle free choice drinking (IA2BC). Male Long Evans
rats (2-month-old, 280-320 g at the start of the experiments) were trained to drink ethanol in the intermittent access in two-bottle free choice (IA2BC) paradigm as previously described 23,24,65 . Briefly, rats were given 24 h concurrent access to one bottle of 20% (v/v) ethanol in tap water and one bottle of plain water, starting at 15:00 on Mondays. After 24 h, the ethanol bottle was replaced by a second water bottle that was available for the next 24 h. This pattern was repeated on Wednesdays and Fridays. On all other days, the animals had unlimited access to two bottles of water. In each ethanol drinking session, the placement of the ethanol bottle was alternated to control for any side preferences. The amount of ethanol or water consumed was determined by weighing the bottles before and after 24 h of access. Ethanol intake was determined by calculating the grams of alcohol consumed per kilogram of body weight.
Stereotaxic surgery and microinjections. Stereotaxic surgery for pharmacological or genetic modulation of the LHb and its histological verification were performed as described 22,66,67

Intra-LHb virus injection and clozapine-n-oxide (CNO) treatment.
We expressed the engineered human muscarinic receptor (DREADDs, designer receptors exclusively activated by designer drugs) to inhibit or excite neuronal activities. Either of AAV5-CaMKIIa-hm4Di-mCherry, AAV5-CaMKIIa-hm3Dq-mCherry, or control AAV5-CaMKIIa-eGFP (titers of 10 12 −10 13 vg/ml, UNC Vector Core, Chapel Hill, NC) were injected bilaterally into the LHb (AP: −3.83 mm, ML: ± 0.63 mm, DV: −5.45 mm) of rats. A volume of 420 nl per side was delivered at a rate of 70 nl min −1 . CNO (1 mg/kg, dissolved in 0.5% DMSO v/v saline, intraperitoneal injection, i.p.) was given 10 minutes before the behavioral tests, 24 h after the last ethanol session. After five weeks of IA2BC paradigm, AAV virus injections were performed. One week after recovery, rats were returned to the IA2BC paradigm and exposed to the paradigm three weeks further. After four to five weeks of AAV virus injection, the behavioral tests were performed.
For overexpression of KCNQ3 in the LHb, we infused into the LHb Cre dependent HSV-LS1-KCNQ3-eYFP virus (gift of Ming-Hu Han in Mount Sinai 33 ) and HSV-IE4/5-CRE virus (from Dr. Rachael Neve, Viral Gene Transfer Core, MIT). HSV-LS1-KCNQ3-eYFP or HSV-LS1-eYFP were mixed in a 2:1 ratio with HSV-IE4/5-CRE www.nature.com/scientificreports www.nature.com/scientificreports/ before injection. The HSV virus were injected into rats that were in the IA2BC paradigm for seven weeks. Three days after injection, rats were returned to the same IA2BC paradigm for an additional one week before the tests were conducted.

pain threshold measurement (Hargreaves test).
Hargrave's test has been used to assess hyperalgesia in alcohol-depedent rats 48 . Paw withdrawal latencies (PWL) were measured with an Analgesia Meter (Model 336; IITC Life Science Instruments, Woodland Hills, CA) as described 8,9,69 . Briefly, each rat was placed in a plexiglass chamber on a glass plate located above a light box. The temperature of glass was set to 25 °C. After an acclimation time to the environment, these rats were subjected to radiant heat, which was applied by aiming a light beam to the middle of the plantar surface of the hind paw. When the rat lifted its paw in response to the heat, the light beam was turned off, and the PWL was recorded. A cutoff time of 20 seconds was used to prevent paw tissue damage. We measured the PWL once a week, at both 2 h and 24 h after the last ethanol drinking session. The two measurements were conducted consecutively, the same PWL changes were used for Fig. 2A,B to determine the correlation between alcohol drinking and pain sensitivity.

Drugs.
We purchased common chemicals from Sigma Aldrich (St. Louis, MO, USA) except retigabine from Alomone (Jerusalem, Israel) and ethanol from Pharmco Products Inc. (Brookfield, CT). Clozapine-N-oxide (CNO) was provided from NIH by the NIDA Drug supply program (NIH, Bethesda, MD).

Data analysis and statistics.
We decided the number of N in the experiments according to the previous our studies and literatures. We compared the mean frequency of spontaneous firing over the last 3-min of 5-min periods of recording and calculated the drug-induced changes after normalizing the data to the preceding 3-min of baseline firing. To determine a correlation between alcohol drinking and pain sensitivity in rats, we calculated the changes in paw withdrawal latency (delta PWL) by subtraction of the response after CNO to the response after SAL at the time point of 1 hour after drug injections (i.p.). Prism (GraphPad, La Jolla, CA) was used for statistical analyses. All compiled data were shown as mean ± SEM and the statistical significance was assessed using paired or unpaired t-tests, and one-or two-way ANOVA with post hoc multiple-comparisons, when appropriate. Linear regression was used to determine the relationship between change in paw withdrawal latency and change in ethanol intake. The values were considered significant when P < 0.05.