Retigabine, a potassium channel opener, restores thalamocortical neuron functionality in a murine model of autoimmune encephalomyelitis

Background: Multiple Sclerosis (MS) is an autoimmune neurodegenerative disease, whose primary hallmark is the occurrence of inflammatory lesions in white and grey matter structures. Increasing evidence in MS patients and respective murine models reported an impaired ionic homeostasis driven by inflammatory-demyelination, thereby profoundly affecting signal propagation. However, the impact of a focal inflammatory lesion on single-cell and network functionality has hitherto not been fully elucidated. Objectives: In this study, we sought to determine the consequences of a localized cortical inflammatory lesion on the excitability and firing pattern of thalamic neurons in the auditory system. Moreover, we tested the neuro-protective effect of Retigabine (RTG), a specific K v 7 channel opener, on disease outcome. Methods: To resemble the human disease, we focally administered pro-inflammatory cytokines, TNF-α and IFN-γ , in the primary auditory cortex (A1) of MOG 35-55 immunized mice. Thereafter, we investigated the impact of the induced inflammatory milieu on afferent thalamocortical (TC) neurons, by performing ex vivo recordings. Moreover, we explored the effect of K v 7 channel modulation with RTG on auditory information processing, using in vivo electrophysiological approaches. Results: Our results revealed that a cortical inflammatory lesion profoundly affected the excitability and firing pattern of neighboring TC neurons. Noteworthy, RTG restored control-like values and TC tonotopic mapping. Conclusion: Our results suggest that RTG treatment might robustly mitigate inflammation-induced altered excitability and preserve ascending information processing.


Introduction
Multiple Sclerosis (MS) is a chronic, demyelinating disease of the central nervous system (CNS), characterized by the accumulation of inflammatory lesions in the white-and grey matter of the brain and spinal cord (Filippi et al., 2018).After crossing the blood-brain-barrier (BBB), autoreactive T and B lymphocytes infiltrate into the brain parenchyma, where they secrete soluble mediators, such as proinflammatory cytokines, promoting gliosis and a gradual neurodegeneration, which correlates with clinical disability (Goverman, 2009).Noteworthy, inflammatory demyelination may be observed in several cortical and deep grey matter areas, including the thalamus, from the earliest stages of the disease (Calabrese et al., 2009;Geurts, 2012), thereby suggesting a paramount role played by an altered functionality of the TC system at the basis of MS pathogenesis (Crandall et al., 2015;Groh et al., 2014;Suga, 2012).
Except for the olfactory pathway, all sensory inputs are filtered and processed in the thalamus.Notably, the medial geniculate nucleus (MGN) of the auditory thalamus is endowed with tonotopically organized TC neurons, conveying and shaping incoming stimuli to the primary auditory cortex (A1) by means of two peculiar firing modes, defined as tonic and burst activity (Sherman, 2001;Iavarone et al., 2019).Tonic firing is characterized by a regular pattern of action potentials (APs), encoding and relaying input from the periphery to the cortex.Conversely, rhythmic bursting is characterized by a brief highfrequency succession of APs, emerging on a low-threshold calcium (Ca 2+ ) spike (LTS), and occurs prominently during physiological sleep or specific neurological disorders (Wang et al., 2020;Murray and Anticevic, 2017;Groh et al., 2018).
Functional abnormalities, reshaping in neuronal morphology and silent microstructural alterations in TC circuitry were suggested to be putative underlying mechanisms in the pathogenesis of fatigue and cognitive impairment in MS patients (Deppe et al., 2016;Capone et al., 2020;Rojas et al., 2018;Evangelou et al., 2001).Further electrophysiological investigations revealed a profound impact of myelin loss and regrowth on neuronal excitability and signal propagation (Cerina et al., 2017;Ghaffarian et al., 2016), thereby indicating alterations in the distribution and expression of specific ion channels as a pathological hallmark of the disease (Bittner et al., 2013;Meuth et al., 2008;Jukkola et al., 2012;Bouafia et al., 2014).Previously, we demonstrated that dynamic changes in network and single-cell intrinsic excitability, accompanied by remodeling of the tonotopic organization of the auditory pathway (Cerina et al., 2017), emerged as a result of an aberrant neuron-oligodendrocyte potassium (K + ) shuttling and homeostasis upon inflammatory demyelination (Kapell et al., 2023).Notably, neuronal K v 7 (K v 7.2-K v 7.5) channels are robustly expressed in central, peripheral and sensory neurons (Trimmer, 2015).Here, opening of these channels mediates an outward hyperpolarizing current, termed M− current (I M ), exerting a pivotal role in regulating neuronal excitability and shaping AP discharge (Baculis et al., 2020).Neuronal K v 7 primarily assemble into K v 7.2-7.3heterotetramers, and their activity can be pharmacologically modulated by the small molecule retigabine (RTG).By means of electrophysiological and behavioral studies, we further demonstrated the neuroprotective potential of RTG treatment on disease outcome, thereby indicating that such approach may potentially counteract MSdriven neurodegeneration (Kapell et al., 2023).
However, the underlying mechanisms through which inflammatory mediators characterizing grey matter lesions ultimately promote neurodegeneration require further investigations and animal models specifically tailored to address it.As the TC system seems to have a strategic position in MS pathology and to mimic the occurrence of inflammatory plaques throughout the brain in the human disease, we induced a grey matter lesion in the A1 of MOG 35-55 immunized animals, by means of focal administration of tumor necrosis factor alpha (TNF-α) and interferon gamma (IFN-γ).Then, we assessed the effect of focal experimental autoimmune encephalomyelitis (EAE) induction on TC cell excitability and firing pattern, by performing ex vivo electrophysiological single-cell recordings.Moreover, we investigated the functionality of the auditory TC network, by means of in vivo longitudinal recordings over the course of EAE and tested the neuroprotective effect of RTG.Further, we explored the relative contribution of different K v 7 isoforms, by characterizing single-cell channel activity and network excitability in Kcnq3 deficient mice.

Animals
All animal experiments were approved by the responsible local authority (Landesamt für Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen; approval ID: 84-02.04.2015.A585 and 81-02.04.2018.A266) and performed in compliance with the ARRIVE guidelines and the 2010/63/EU of the European Parliament and of the Council of 22 September 2010.Animals were kept in groups in individually ventilated cages (IVC) and maintained on a 12-hour light/dark cycle, with food and water available ad libitum.Constitutive Kcnq3 knock-out (Kcnq3 -/-) mice were obtained from Thomas J. Jentsch (Charité -University Medicine, Berlin) and had been generated as previously described (Gao et al., 2021).C57BL/6J mice were used as controls (termed wild-type, WT, hereinafter).

Surgery − focal injection of cytokines
Cytokine injections in the primary auditory cortex (A1) were performed ten days following EAE induction.Mice were deeply anesthetized and thereafter placed in a stereotactic frame, as previously described (Narayanan et al., 2018).The scalp above the dorsal cortex was excised and the skull was exposed.The following coordinates were measured to target A1: anteroposterior, − 2.18 mm; lateral, +4.2 mm relative to Bregma; dorsoventral, 1 mm from the brain surface (Keith and Franklin, 2019).Craniotomies were performed unilaterally (left hemisphere) above the target area using a dental drill and 1 µl of a solution filled with 150 U/mg TNF-α and 800 U/mg IFN-γ dissolved in PBS was injected in layer IV of A1 (200 nl/min).EAE mice receiving a stereotactic injection of PBS in the same region served as controls.Following surgery, the skin was sutured and animals were allowed to recover for 30 min on a warm pad prior to returning to their home cages.

Tissue preparation for histology and IHC
Mice were transcardially perfused with phosphate buffer saline (PBS) and paraformaldehyde (PFA).Thereafter, tissues were postfixed for 24 h in 4 % PFA prior to being transferred to PBS.Serial coronal brain slices (75 μm thickness) containing the auditory TC system were prepared on a vibratome (7000smz-2, Campden Instruments).

Morphological identification of recorded TC neurons
During recording, neurons were passively filled with 0.5 % biocytin (Sigma Aldrich) and thereafter brain slices were fixed in 4 % PFA overnight for subsequent post hoc visualization.After being washed in PBS, slices were incubated in a blocking solution containing 1 % Triton X-100 and 2 % bovine serum albumin in PBS for 1 h at RT and subsequently processed for immunostaining with 5 µg/ml Streptavidin Cy3 conjugate (Sigma S-6402) for 3 h at RT. Tissues were washed five times in PBS and mounted on glass slides using SlowFade Gold (Invitrogen).

Image acquisition
Images were acquired using a fluorescence microscope (Keyence, BZ-X800) equipped with 4x and 20x Plan Apochromat objectives.Whole hemisphere streptavidin staining tiled images were taken with a 4x objective and processed using Keyence BZ-X800 Analyzer Software version 1.1.2.4.Z-stack images of Iba1-positive cells were acquired with a 40x Plan Apochromat objective and converted into maximum projection images using the full focus sectioning feature of the BZ-X800 Analyzer.

Electrode implantation
For electrode implantation, ten days prior to EAE induction animals were anaesthetized and placed in a stereotactic apparatus as previously described (see Surgeryfocal injection of cytokines).Craniotomies were performed unilaterally (left hemisphere) to accommodate microwire arrays (one array, eight electrodes and one reference/array per brain region; Stablohm 650; California Fine Wire) in the vMGN, as previously reported (Narayanan et al., 2018).Thereafter, mice recovered on a heating pad for 30 min before being housed in IVC racks.Following surgery, the health status of the animals and their recovery were monitored daily for ten days prior to experimentation.

In vivo electrophysiological recordings
WT and Kcnq3 -/-mice were allowed to recover for ten days following electrode implantation prior to be MOG 35-55 immunized.Starting from the day of EAE induction (day 0), WT animals received daily an intraperitoneal injection of either RTG (1 mg/kg) or saline (SAL) and unitary activity recordings were performed during the early (day 12) and chronic (day 35) stage of the disease.Recordings carried out prior to immunization served as a control.The impact of Kcnq3 deletion on neuronal activity was assessed by performing recordings at the same time points in untreated WT and Kcnq3 -/-animals.To determine frequency-related responses of thalamic neurons of the vMGN, recordings were performed in freely moving mice before and during the presentation of an auditory stimulus consisting of six repetitions of either low-or high-frequency tones (2.5 kHz and 10 kHz at 85 dB, respectively), as previously described (Narayanan et al., 2018;Daldrup, 2016) (Supplementary Methods).

Statistical analysis
All results are presented in the text as mean ± SEM.Statistical significance of data with a Gaussian distribution was determined using unpaired two-tailed Student's t-test.Statistical comparisons across three or more groups were performed using one-way ANOVA or factorial mixed design ANOVAs, followed by Tukey, Dunnett or Šidák post-hoc test.Multiple comparisons were assessed by two-way ANOVA or factorial mixed design ANOVA, complemented by Tukey post-hoc test.Statistical significance differences were reported using the following pvalues: p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), and p < 0.0001 (****).GraphPad Prism (version 9.3.1),Origin, and Adobe Illustrator were used for data analysis and figure preparation.

Immunofluorescence staining reveals dense infiltrates of activated microglia in A1 upon focal cytokine injection
A hallmark of MS is the occurrence of cortical inflammatory lesions, profoundly affecting clinical manifestations and disease progression.To mimic the human pathology and assess the impact of a focal targeted grey matter lesion on TC neuron functionality, 10 days post immunization (dpi), we stereotactically injected TNF-α and IFN-γ into the left A1 of EAE mice (Fig. 1), whilst animals receiving a focal injection of PBS served as controls.First, in order to verify the presence of inflammatory lesions in our model, we immunohistochemically visualized microglia and we explored their activation status using the MHC class I antigen (OX-18) in the perilesional hemisphere of cytokine-and PBSinjected mice.Dense infiltrates of activated microglia were detected in the ipsilateral auditory cortex upon cytokine injection at 12 dpi (Fig. 2A) and persisted in chronic EAE (Fig. 2C).Conversely, the perilesional area of mice which received an intracortical injection of PBS was characterized by less microglia activation, confined to the proximity of the needle track (Fig. 2B, D).Noteworthy, microglia displayed enlarged and more irregular cell bodies upon focal cytokine injection (Fig. 2A'), in comparison to PBS (Fig. 2B'), accompanied by colocalization of Iba1 and OX-18, thereby suggesting a persisting and more robust microglia activation following focal cytokine injection (Fig. 2C'), in comparison to PBS (Fig. 2D').Immunofluorescence staining in the contralateral hemisphere of cytokine-injected mice revealed the higher specificity of inflammatory lesions and microglia activation driven by the injection of inflammatory mediators in A1 (Fig. S1).

TC neuron excitability is modulated by focal injection of cytokines in A1
To determine the impact of focal EAE induction on neuronal excitability, we characterized vMGN TC neuron activity ex vivo in acute brain slices, by performing whole-cell current-clamp recordings in the hemisphere where the cytokines were injected (ipsilateral hemisphere), 12 and 35 days after EAE induction (Fig. 1).Recordings carried out in PBSinjected mice and naïve WT without focal EAE induction served as controls.Tonic firing was evoked by injecting direct current steps of increasing amplitude (from +20 pA to +160 pA, 20 pA increment), while holding the cell at − 55 mV (Fig. 3A).Recordings performed in the ipsilateral hemisphere revealed the emergence of a hyperexcitable phenotype of TC neurons at 12 dpi, as compared to cells recorded in WT (+160 pA, p = 0.0006), thereby indicating that focal injection of cytokines in the A1 profoundly affects the degree of excitability of thalamic relay cells (Fig. 3 A').Further, to explore dynamic changes in TC neuron activity as the disease converts to a neurodegenerative stage, we performed electrophysiological recordings at 35 dpi (Fig. 3B).Interestingly, a robust increase in the number of APs was observed in neurons recorded in the ipsilateral hemisphere compared to both PBS (+160 pA, p = 0.0002) and WT animals (+160 pA, p < 0.0001) (Fig. 3 B').
Next, we investigated temporal aspects of stimulus-evoked TC neuron activity by plotting the instantaneous firing frequency (IFF) of consecutive APs vs. the interspike interval (ISI) number.The firing pattern of TC neurons recorded in the ipsilateral hemisphere at 12 dpi was characterized by a significantly increased initial IFF, in comparison to both TC neurons in the PBS-injected hemisphere (p < 0.0001) and WT (p = 0.0009) (Fig. 3C).The initial AP discharge was followed by a postburst accommodation, characterized by a progressively increasing ISI in all experimental groups.The same experimental approach was employed to monitor dynamic changes in the firing pattern and frequency of AP discharge in the neurodegenerative stage of the EAE course (35 dpi).However, no significant differences were observed between TC neurons recorded from the ipsilateral hemisphere of EAE animals and both PBS and WT controls (Fig. 3D).Morphological and histological features of vMGN TC neurons were identified by inclusion of biocytin in the recording pipette and subsequent streptavidin staining (Fig. 3E).Neurons with typical characteristics were identified revealing large cell bodies (15-25 μm in diameter) giving rise to three to six dendrites (Kanyshkova et al., 2012;Budde et al., 2023).
Taken together, these results suggest that pro-inflammatory cytokines, such as TNF-α and IFN-γ, play a paramount role in driving hyperexcitability of TC relay neurons of the vMGN, thereby highlighting the modulation of thalamic components of the TC system both at the preclinical and progressive neurodegenerative phase of the disease course.

Pro-inflammatory cytokines drive enhancement of LTS-mediated firing in vMGN neurons
To explore the effect of pro-inflammatory mediators on LTS generation in TC neurons, we challenged vMGN neurons with depolarizing and hyperpolarizing current pulses (+160 pA and − 180pA, respectively) from resting membrane potential (RMP) and analyzed the number of LTS-mediated APs.In response to a depolarizing current pulse, the number of LTS-mediated APs was not significantly increased upon EAE induction at 12 dpi (Fig. 4A, A').However, in the chronic stage of the disease (Fig. 3B, B'), we observed a higher number of LTS-APs at the onset of the depolarizing current input, thereby indicating that cytokine injection facilitated the emergence of burst firing in TC neurons recorded from the ipsilateral vMGN, as compared to WT controls (p = 0.0446).
Next, we stimulated TC neurons with a hyperpolarizing current injection and measured the number of LTS-mediated spikes characterizing rebound bursts (Fig. 3 C, D).Accordingly, TC neurons displayed a tendency to fire a larger number of APs 12 and 35 days following EAE induction, although without reaching significance threshold (Fig. C', D').
Collectively, these findings suggest that intrinsic hyperexcitability driven by inflammatory mediators is accompanied by dynamic changes in the firing pattern of TC neurons, thereby facilitating the emergence of bursting activity following a focal inflammatory lesion in the A1.Noteworthy, no significant differences emerged when single-cell excitability and the firing pattern of neurons recorded in PBS animals and in the contralateral hemisphere of cytokine-injected mice were compared (Fig. S2).

Effect of Kcnq3 deletion on firing pattern and intrinsic properties of vMGN TC neurons
RTG is a selective K v 7.2-7.5 channel agonist, exerting its action by potentiating I M generation and, consequently, robustly attenuating inflammation-induced hyperactivity of TC neurons (Fig. S3).In the following, we characterized the impact of K v 7.3 subunits on intrinsic excitability and mediating the RTG effect, by performing ex vivo patchclamp recordings in acute brain slices from Kcnq3 deficient mice (Fig. 5A).First, we explored differences in active membrane properties between genotypes by examining TC neuron excitability, using a series of increasing depolarizing current pulses (20-160 pA, 1.5 s duration).TC neurons recorded from Kcnq3 KO mice fired a higher number of APs in response to the same current injection (+160 pA, p = 0.0023) (Fig. 5B).In order to assess the contribution of a specific ion channel modulator in the absence of the K v 7.3 subunits, we further quantified AP reduction in the presence of RTG (Fig. 5C, Fig. S4).Interestingly, RTG application exerted only a partial effect in reducing the number of AP recorded under control conditions in TC neurons from Kcnq3 KO mice, as compared to WT (p = 0.0013).
Next, we explored differences in the passive properties between genotypes (Fig. 5D-H).No significant differences emerged from the analysis of the RMP.Conversely, TC neurons recorded from mutant mice exhibited a higher input resistance (p = 0.0128) and a lower rheobase (p = 0.0051) compared to WT, thereby strengthening the association between lack of K v 7.3 and thalamic hyperexcitability.Similar voltage threshold and AP amplitude were detected in Kcnq3 KO and WT mice, thus suggesting a marginal impact of these subunits in shaping AP waveform.
Furthermore, we sought to determine the impact of K v 7.3 on the firing pattern of thalamic cells (Fig. 5I-K).TC neurons from Kcnq3 KO mice fired bursts consisting of a larger number of spikes (p = 0.040).Moreover, a significantly reduced spike-frequency accommodation (SFA) was observed in mutants (p = 0.0383), thereby suggesting that functional K v 7.3 significantly influence the frequency of AP discharge (Fig. 5L).However, no significant differences were detected in the number of LTS-mediated APs elicited in response to both depolarizing and hyperpolarizing current pulses at RMP.Similarly, temporal aspects of stimulus-evoked TC neuron activity did not differ between WT and mutants (Fig. S4).
The hyperpolarizing I M in neurons is largely mediated by K v 7.2/7.3heteromeric channels.Thus, we explored the consequence of K v 7.3 deletion on current generation and in mediating the RTG effect (Fig. 5M,  N).Analysis of the average basal K v 7-mediated current revealed only a marginal reduction of I M amplitude.RTG application led to a significant current increment in both WT (p = 0.0005) and Kcnq3 KO mice (p = 0.076).However, the agonist-elicited current was much smaller in mutants, as compared to WT (p = 0.0025).Control-like values were restored when RTG and XE991 were simultaneously added to the bath solution.
Collectively, these data clearly indicate that the K v 7.3 subunits strongly influence active and passive membrane properties of TC neurons.Furthermore, RTG-mediated currents largely depend on the presence of functional K v 7.3 channels.

Auditory-stimulus discrimination and burst discharge are affected by Kcnq3 loss in vivo
In order to investigate the impact of Kcnq3 deletion on neuronal network excitability and sensory information processing driven by inflammation, we performed in vivo electrophysiological recordings in vMGN in a longitudinal manner.Here, similarly to previous experimental approaches (Kapell et al., 2023;Narayanan et al., 2018), we presented the mice with two different tones, 2.5 and 10 kHz, and measured changes in stimulus-evoked neuronal activity as compared to baseline in the region representing 10 kHz.Analysis revealed that before EAE induction WT and Kcnq3 KO animals exhibited a similar response to the not relevant stimulus (2.5 kHz), thereby indicating that Kcnq3 loss did not influence tone discrimination.However, in line with our previous ex vivo findings, an increased firing rate was observed in Kcnq3 KO mice elicited at 10 kHz (p = 0.021) (Fig. 6A, A').Twelve days following EAE induction, an increased z-score and stimulus-related firing rates were observed in EAE-Kcnq3 KO mice irrespectively of the frequency of the presented tone, thereby suggesting a state of neuronal hyperexcitability in response to 2.5 kHz (p < 0.0001) and 10 kHz (p < 0.0005), along with an impaired tonotopic organization of the vMGN (Fig. 6B,  B').Furthermore, we detected an increased baseline activity of thalamic neurons recorded from EAE-Kcnq3 KO animals, in the absence of an auditory stimulus (p < 0.0001).Next, we investigated differences in network excitability between EAE-WT and EAE-Kcnq3 KO mice by performing in vivo recordings in the chronic stage of the disease (35 dpi).Neurons recorded from EAE-Kcnq3 KO mice were characterized by an increased intrinsic excitability when their baseline activity was assessed, in the absence of an auditory stimulus (p < 0.0008) (Fig. 6C, C').Strikingly, the stimulus-evoked firing rates did not differ between genotypes, thereby suggesting comparable levels of hyperexcitability at this time point.
We previously showed that Kcnq3 deletion facilitated the emergence of bursting activity of TC neurons recorded in acute brain slices.Hence, we further explored the consequences of inflammation on highfrequency burst discharges in Kcnq3 KO mice and respective WT controls.In line with our ex vivo electrophysiological investigations, an increased burst firing was detected in Kcnq3 KO mice prior to EAE induction (p < 0.0001) (Fig. 6E).This held true when differences in bursting activity between genotypes were assessed at 12 (p = 0.0013) and 35 dpi (p = 0.0111).
Next, we sought to determine whether lack of Kcnq3 affected the response latency.Prior to EAE induction, response latencies to auditory stimuli were significantly reduced in Kcnq3 KO mice, as compared to WT (p = 0.0016) (Fig. 6F).However, no differences between genotypes emerged when response latencies were investigated in the acute (12 dpi) and chronic (35 dpi) stage of the disease.
Collectively, these data indicate that EAE induction is associated with the emergence of hyperexcitability and promotes bursting discharge in vivo (Fig. S5A, B).Furthermore, lack of K v 7.3 is associated with an altered auditory network excitability under control conditions and following EAE induction.Facilitation of bursting activity and altered latency responses might indicate a paramount role played by K v 7.3 subunits in ensuring faithful information processing and transmission.  (/-) (purple) mice, recorded under control conditions (shaded traces) and in the presence of RTG.(B) Number of APs evoked in response to the injection of positive currents with 20 pA increments from RMP. Kcnq3 deletion promoted a significantly increased excitability of TC neurons (n = 8 cells from 3 wt mice; Kcnq3 (-/-) = 10 cells from 5 mice; +120 pA, p = 0.0224; +140 pA, p = 0.0112; + 160 pA, p = 0.0023, two-way ANOVA complemented by Šidák's test for multiple comparison).(C) Bar graph indicating only a partial percentage AP reduction in Kcnq3 (-/-) following RTG application in comparison to controls (n = 8 cells from 3 wt mice; Kcnq3 (-/-) = 5 cells from 3 mice, p = 0.0013, unpaired two-tailed Student's t-test).(D-H) Comparison of RMP, R in (p = 0.0128, unpaired two-tailed Student's t-test), rheobase (p = 0.0051, unpaired two-tailed Student's t-test), AP voltage threshold, and AP amplitude.Normalized LTS-mediated and rebound APs elicited in response to depolarizing (I) and hyperpolarizing (J) current pulses (+160 pA and − 180 pA, respectively) from RMP (n = 8 cells from 3 wt mice; Kcnq3 (-/-) = 5 cells from 3 mice).(K) Box plots comparing the normalized number of spikes in each burst.Kcnq3 loss facilitated the occurrence of a burst discharge characterized by an increased number of spikes as compared to controls (n = 4 cells from 3 wt mice; Kcnq3 (-/-) = 5 cells from 3 mice, p = 0.040, unpaired two-tailed Student's t-test).(L) Bar graph indicating a significant increased SFA index in TC neurons recorded from WT animals as compared to Kcnq3 (-/-) (n = 4 cells from 3 wt mice; Kcnq3 (-/-) = 5 cells from 3 mice, p = 0.0383, unpaired two-tailed Student's t-test).Representative voltage-clamp recordings (M) and bar graphs (N) indicating the impact of Kcnq3 deletion on M− current amplitude and the effect of KCNQ channel modulators.Kcnq3 loss was associated with a significant reduction of the RTG-evoked current in Kcnq3 (-/-) as compared to WT (n = 7 cells from 3 wt mice; Kcnq3 (-/-) = 8 cells from 3 mice; WT control vs. WT+RTG, p = 0.0005; Kcnq3 KO control vs. Kcnq3 KO+RTG, p = 0.0076, mixed model two-way ANOVA complemented by Tukey's test for multiple comparison).Error bars represented as mean ± SEM; p < 0.05 (*), p < 0.01 (**), p < 0.001 (***).See also Supplementary statistical analysis.

RTG treatment reverses the impaired auditory network functionality emerging upon EAE induction
Testing the translational potential of a neuroprotective compound, such as RTG, requires the monitoring of neuronal network dynamics in the living animal over the course of the disease.To address this, singleunit activity was recorded in the vMGN, before and following EAE induction, in freely moving mice treated with either RTG (1 mg/kg) or saline (SAL), by employing the experimental approach described above (Fig. 7).First, we confirmed that the animals were able to discriminate between the two tones (2.5 and 10 kHz) and the absence of intrinsic preimmunization differences between the two experimental groups by analyzing Z-scores and stimulus-related firing rates in naïve mice (Fig. 7A, A').Twelve days after immunization, the mice treated with SAL lost their ability to discriminate between the two tones, as reflected by their Z-score increment in response to both 2.5 and 10 kHz (Fig. 7 B).Moreover, the magnitude of the vMGN neuron response was higher as compared to the one observed before immunization.Conversely, treatment with RTG for 12 days following EAE induction indicated a robust reduction in the Z-score in comparison to SAL-treated mice.Analysis of the stimulus-related firing rates at this time point revealed that while the response of the mice treated with RTG did not exceed pre-immunization levels, SAL-treated animals exhibited an increased rate in response to both 2.5 kHz (p < 0.0001) and 10 kHz (p < 0.0001), thus indicating hyperexcitability (Fig. 6B').This held true when differences in baseline activity were investigated (p < 0.0001).Treatment with RTG significantly reduced thalamic hyperactivity in comparison to SAL-treated mice at baseline (p < 0.0001), in response to 2.5 kHz (p < 0.0001), and 10 kHz (p < 0.0001).Next, we explored the neuroprotective effect of a prophylactic RTG treatment in preventing chronic neuronal hyperexcitability, by measuring single-unit activity 35 dpi (Fig. 6C, C').Interestingly, similarly to what observed in the acute stage of the disease, we detected a robust reduction in the stimulus-evoked responses to either 2.5 kHz (p = 0.0160) and 10 kHz (p < 0.0001) in RTG-treated mice as compared to the SAL-control group.
Next, to assess the impact of RTG treatment on TC neuron firing pattern in vivo, we measured dynamic changes in burst discharge prior and following EAE induction.In line with our ex vivo findings, vMGN neurons exhibited a robust increase in bursting activity upon EAE induction (Fig. 6E).However, RTG treatment resulted in a robust attenuation of burst discharge both 12 (p < 0.0001) and 35 dpi (p < 0.0001).
No further differences emerged from the analysis of latency responses following auditory stimulation between the two experimental groups (Fig. 6F).
Taken together, our data indicated that RTG treatment reverts immunization-dependent hyperexcitability in our in vivo model, thus corroborating the ex vivo findings.

Discussion
In this study, we aimed to shed light on the contribution of proinflammatory mediators to the altered excitability arising upon EAE induction, focusing on the role played by the K v 7 channels and their pharmacological modulation as a neuroprotective strategy to restore neuronal network functionality.
The effect of cytokine release on ion channel function and, thus, neuronal excitability, is an intriguing, but hitherto partially understood component of MS and EAE pathogenesis.Recent studies in the cuprizone model of general de-and remyelination explored changes in the expression and regulation of HCN channel activity driven by inflammatory mediators (Oniani et al., 2022;Chaudhary et al., 2022).Notably, while IFN-⍺ application significantly reduced the number and duration of bursts, presumably by attenuating HCN-mediated currents, an opposite effect was detected in the presence of IL-1β.Divergence and redundancy of pro-inflammatory cytokines, complex signaling cascades, and dynamic changes in their expression throughout the disease, along with intrinsic limitations and variability that experimental models entail, make a translation of such findings into the clinic challenging (Gobel et al., 2018;Bierhansl et a., 2022;Rychlik et al., 2023).In the attempt to mimic the appearance of brain lesions observed in the human disease, here we employed the focal EAE model, by inducing an inflammatory lesion in A1, and investigated its impact on vMGN TC neuron functionality.We found that TC neurons projecting to the hemisphere where the inflammatory lesion was induced (ipsilateral site) exhibited hyperexcitability already at 12 dpi, as compared to recordings performed in both EAE sham and WT control mice.These observations corroborate our recent findings indicating the emergence of a cortical hyperactive phenotype driven by inflammatory demyelination in the acute phase of the disease (Kapell et al., 2023).Our data further suggests that the plethora of inflammatory mediators released by CNS-infiltrated cells, notably TNF-⍺ and IFN-γ, might contribute to the altered excitability observed in MS models.Inflammation-induced maladaptive mechanisms, such as an increased frequency of glutamate-mediated spontaneous and TTX-resistant miniature excitatory postsynaptic currents, were reported to emerge early in the disease and strengthen the link between cytokine release and synaptic dysfunction (Acharjee et al., 2018;Centonze et al., 2009).Accordingly, a recent work by Ellwardt et al. (Ellwardt et al., 2018) hypothesized that cortical hyperactivity in EAE might be mediated by TNF-⍺ secretion.Interestingly, a study by Shim et al. (Shim et al., 2018) revealed that this cytokine promoted an increased excitability of cerebellar neurons by fostering glial glutamate release.
Alongside an increased excitability, we detected an enhanced burst activity in the perilesional vMGN.This firing mode arises as a result of the interplay of specific ionic conductances, such as I h and I T (Datunashvili et al., 2018;Cain and Snutch, 2013;Bosch-Bouju et al., 2013) and has been suggested to be a hallmark of pathological Fig. 6.Auditory-stimulus discrimination and burst discharge are affected by Kcnq3 loss in-vivo.(A, B, C) Z-scores of unitary activity recordings performed during the early (day 12) and chronic (day 35) stage of the disease.Recordings obtained before EAE induction served as a control.Responses of vMGN neurons from WT (black) and Kcnq3 (-/-) (purple) mice were assessed following the presentation of auditory stimuli of either low (2.5 kHz, blue insets) or high (10 kHz, yellow insets) frequencies.(A', B', C') Bar graphs indicating baseline activity and frequency-specific responsiveness of vMGN neurons.(A') Comparison of the stimulusevoked neuronal activity (10 kHz) revealed an increased firing rate in mutants (n = 34 cells from 6 wt mice before EAE induction; Kcnq3 (-/-) = 34 cells from 9 mice before EAE induction; 10 kHz, p = 0.021, two-way ANOVA complemented by Šidák's test for multiple comparison).(B') 12 days following EAE induction TC neurons recorded from Kcnq3 KO mice displayed an increased firing-rate (n = 34 cells from 6 wt-EAE day 12; EAE-Kcnq3 (-/-) day 12 = 34 cells from 9 mice; baseline, p < 0.0001; 2.5 kHz, p < 0.0001; 10 kHz, p < 0.0005, two-way ANOVA complemented by Šidák's test for multiple comparison).(C') Thalamic neurons recorded from mutants were characterized by an increased baseline activity 35 days following EAE induction (n = 34 cells from 6 wt-EAE day 35; EAE-Kcnq3 (-/-) day 35 = 34 cells from 9 mice; baseline, p < 0.0008, two-way ANOVA complemented by Šidák's test for multiple comparison).(D) Pie charts representing the percentage of RNs and n-RNs from WT and Kcnq3 (-/-) mice in the presence of an auditory stimulus.(E) Analysis of the firing pattern of vMGN neurons before and after EAE induction revealed that absence of Kcnq3 facilitated burst discharge (day 0, p < 0.0001; day 12, p = 0.0013; day 35, p = 0.0111, two-way ANOVA complemented by Šidák's test for multiple comparison).(F) Bar graphs indicating the latency of vMGN neurons responses following the presentation of an auditory stimulus.Neurons recorded from Kcnq3 (-/-) animals are characterized by a significantly reduced response latency before EAE induction (p < 0.0001; two-way ANOVA complemented by Šidák's test for multiple comparison).Error bars represented as mean ± SEM; p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), p < 0.0001 (****).See also Supplementary statistical analysis.conditions, such as epilepsy.However, more ionic conductances shape the bursting properties of TC neurons, including I M , a fast transient K + current (I A ), and a Ca 2+ -activated K + current (I KCa ).The original view of bursts as all-or-none events and its functional relevance have been questioned by recent studies aimed at unraveling the information encoded in burst length and frequency (Elijah et al., 2015).In comparison to a previous study aimed at resembling the appearance of a focal demyelinating lesion (Narayanan et al., 2018), we detected an increased percentage and number of bursts, both ex vivo and in vivo.Divergences in this regard might be due to intrinsic differences in the experimental models employed (EAE vs. cuprizone, focal inflammatory vs. focal demyelinating lesion).Therefore, further investigations are required and might be complemented by behavioral readouts.In this respect, other ion channels and additional thalamic and prethalamic cell types shaping the generation of bursts in TC neurons may be influenced in the present experimental model.
Alterations in TC neuron firing pattern might reflect an impaired transmission of ascending information to the cortex.Accordingly, several recent studies shed light on the effect of general de-and remyelination processes on the activity of the thalamocortical system in experimental models of MS (Cerina et al., 2017;Narayanan et al., 2018).Notably, irreversible changes in the tonotopic organization of the auditory pathway were detected in EAE (Kapell et al., 2023) and even following complete remyelination in the cuprizone model (Cerina et al., 2018), thereby strengthening the link between altered excitability and impaired cytoarchitecture and network functionality.Noteworthy, recent studies elucidated the behavioral correlate of inflammation-and demyelination-induced altered excitability, thereby providing evidence that neural hyperactivity is accompanied by increased anxiety and reduced exploratory behavior, along with long-term memory impairment (Kapell et al., 2023;Ellwardt et al., 2018;Cerina et al., 2018).By combining ex vivo and in vivo electrophysiological approaches, we gained novel insights into an exacerbation of neuronal hyperexcitability in the neurodegenerative phase of the disease (35 dpi).Evidences indicated early neuroaxonal damage as a hallmark of MS, eventually becoming more prominent in the chronic phase of the disease (Schirmer et al., 2013).Alterations in ion channel expression and regulation robustly contribute to this phenomenon and were reported to be involved in neuroaxonal injury.We previously described impaired K + homeostasis at and around Nodes of Ranvier as a key mechanism to drive altered excitability in EAE (Kapell et al., 2023).Furthermore, axonal injury was attributed to diffuse distribution of several subtypes of ion channels, such as Kv1.1, Kv7, Cav2.2, Nav1.2 and Nav1.6 along chronically demyelinated neurons (Beraud et al., 2006;Kornek et al., 2001;Craner et al., 2004;Alrashdi et al., 2019;Bianchi et al., 2018;Schattling et al., 2012).
Among these players, we investigated the contribution of the K v 7 channels, with a focus on the K v 7.3 subunits, whose involvement in the pathogenesis of MS and respective experimental models was previously explored (Kapell et al., 2023;Hamada and Kole, 2015).Here, we precisely characterized K v 7 channel function under homeostatic and pathological conditions and explored the effect of selective pharmacological modulators, such as RTG and XE991, on TC neuron functionality by taking advantage of Kcnq3 KO animals.We found that lack of K v 7.3 dramatically affected active and passive properties of TC neurons.These observations are in accordance with previous studies, indicating increased excitability and disruption of spike-frequency adaptation in mutants (Kapell et al., 2023;Gao et al., 2021;Peters et al., 2005).This feature is mediated by the activity of slow hyperpolarizing currents, such as I M and I KS (Otto et al., 2006;Schwarz et al., 2006), whose activation provides a fundamental inhibitory sink in response to trains of APs.Therefore, an aberrant spike-frequency adaptation might affect neuronal firing pattern and, hence, lead to detrimental consequences for neuronal coding.For instance, it was suggested that this biophysical property is of fundamental importance in localizing moving sounds (Goodman and Brette, 2010).Moreover, our in vivo investigations revealed that lack of K v 7.3 subunits further exacerbates vMGN TC neuron hyperexcitability, in accordance with our previous findings in A1 (Kapell et al., 2023).
I M is mediated by different combinations of K v 7.2, K v 7.3 and K v 7.5 subunits (Barrese et al., 2018;Brown and Passmore, 2009), with RTG being described to have a higher affinity for heteromeric channels (Kim et al., 2015).Accordingly, significant differences between genotypes emerged when RTG-evoked currents were compared, thereby suggesting that K v 7.3 channels play a paramount role in mediating the RTG effect.In line with our observations, previous studies reported that mutants having Kcnq3 deletion did not compensate by upregulating Kv7.2 channels (Kapell et al., 2023;Gao et al., 2021).

Conclusions
We previously showed that RTG treatment significantly mitigates disease burden in EAE animals, thereby slowing neurodegenerative processes in the chronic phase of the disease (Kapell et al., 2023).Here, we further provided evidence that a chronic, prophylactic treatment with this compound significantly reduces EAE-driven hyperexcitability, restores pre-immunization-like values and prevents alterations in thalamocortical network functionality, thereby further suggesting the use of K v 7 channel modulators as a novel neuroprotective strategy in MS and neurodegenerative disorders characterized by hyperexcitability.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 1 .
Fig. 1.Graphic representation of the ex-vivo experimental design.WT and Kcnq3 (-/-) mice were administered myelin oligodendrocyte glycoprotein (MOG 35-55 ) and Micobacterium tuberculosis emulsified in complete Freund's adjuvant, and treated with Bordetella pertussis toxin (PTX) (day 0).A focal EAE was induced in WT mice at day 10, by injecting the proinflammatory cytokines TNF-α and IFN-γ into the left A1 (EAE ipsi).Electrophysiological ex-vivo recordings were performed in the vMGN of the thalamus in the early (day 12) and chronic (day 35) phase of the EAE course.

Fig. 2 .
Fig. 2. Immunofluorescence staining reveals dense infiltrates of activated microglia upon focal cytokine injection.Representative photomicrographs and close-ups of the ipsilateral auditory cortex of cytokine-and PBS-injected EAE mice immunostained with DAPI (nuclear marker), OX-18 (MHC class I antigen), and Iba1 (microglia marker), in the acute (day 12) and chronic stage (day 35) of the EAE course.The perilesional area of cytokine-injected mice is characterized by a higher infiltration of activated microglia 12 (A, B) and 35 dpi (C, D), as compared to animals which received a PBS injection.Scale bars: 500 μm.Close-ups of the Iba1-positive cell morphology revealing an enlarged and more irregular soma upon focal cytokine injection (A'), in comparison to PBS (B').Colocalization of Iba1 (green) and OX-18 (red) was observed in the perilesional cytokine-injected area (C', indicated by arrows), but not in PBS (D').Scale bars: 20 μm.

Fig. 7 .
Fig. 7. Retigabine treatment reverses the impaired auditory network functionality emerging upon EAE induction.(A, B, C) Z-scores of auditory stimulusevoked responses of vMGN neurons.Recordings were performed before and during the presentation of 2.5 kHz (blue insets) and 10 kHz (yellow insets) tones in EAE mice undergoing a prophylactic treatment with either saline (black) or RTG (1 mg/kg, red).Recordings obtained prior to EAE induction served as a control.(A', B', C') Bar graphs indicating stimulus-related firing rates measured as baseline activity and in response to auditory stimuli of low (2.5 kHz) and high (10 kHz) frequencies (n = 34 cells from 7 SAL mice; n = 34 cells from 7 RTG mice).(B') Prophylactic treatment with RTG prevented EAE-induced increased excitability at 12 (baseline, p < 0.0001; 2.5 kHz, p < 0.0001; 10 kHz, p < 0.0001, two-way ANOVA complemented by Šidák's test for multiple comparison).(C') RTG treatment led to a significant reduction of stimulus-evoked responses 35 dpi (2.5 kHz, p = 0.0160; 10 kHz, p < 0.0001, two-way ANOVA complemented by Šidák's test for multiple comparison) (D) Pie charts displaying the percentage of responsive neurons (RNs) and not responsive neurons (n-RNs) to the presentation of an auditory stimulus in animals treated with either RTG or SAL.(E) Bar graphs showing the tendency of vMGN neurons to fire in burst mode in response to an auditory stimulus (day 12, p < 0.0001; day 35, p < 0.0001, two-way ANOVA complemented by Šidák's test for multiple comparison).(F) Latency of neuronal responses in the vMGN to the presentation of an auditory stimulus before (day 0) and at different stages of the EAE course (12 and 35 dpi).Data are presented as mean ± SEM; p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), p < 0.0001 (****).See also Supplementary statistical analysis.