Topographical Organization of M-Current on Dorsal and Median Raphe Serotonergic Neurons

Dorsal and median raphe nuclei (DR and MR, respectively) are members of the reticular activating system and play important role in the regulation of the sleep-wakefulness cycle, movement, and affective states. M-current is a voltage-gated potassium current under the control of neuromodulatory mechanisms setting neuronal excitability. Our goal was to determine the proportion of DR and MR serotonergic neurons possessing M-current and whether they are organized topographically. Electrophysiological parameters of raphe serotonergic neurons influenced by this current were also investigated. We performed slice electrophysiology on genetically identified serotonergic neurons. Neurons with M-current are located rostrally in the DR and dorsally in the MR. M-current determines firing rate, afterhyperpolarization amplitude, and adaptation index (AI) of these neurons, but does not affect input resistance, action potential width, and high threshold oscillations.These findings indicate that M-current has a strong impact on firing properties of certain serotonergic neuronal subpopulations and it might serve as an effective contributor to cholinergic and local serotonergic neuromodulatory actions.


INTRODUCTION
DR and MR are members of the reticular activating system and play important role in the regulation of sleep-wakefulness cycles, movement, and affective states. The activity of DR is increased during wakefulness, meanwhile, it is less active in paradoxical or REM-sleep (Monti, 2011).
M-current is present in several brain areas including the brainstem (Kharkovets et al., 2000;Koyama and Appel, 2006;Hansen et al., 2008). It is a voltage-gated potassium channel modulated by metabotropic receptors; and inhibited by the activation of muscarinic acetylcholine receptor and other G-protein coupled receptors like 5HT 2C receptor (Marrion, 1997;Roepke et al., 2012).
It was shown that DR and MR serotonergic neurons have several subgroups based on expressed markers, morphological characteristics, and in vitro and in vivo electrophysiological properties (Hornung, 2003;Andrade and Haj-Dahmane, 2013). Approximately 60% of DR serotonergic neurons express KCNQ4, the main ion channel subunit for M-current (Hansen et al., 2008;Zhao et al., 2017). However, neither the proportion of MR serotonergic neurons possessing M-current nor the topographical distribution of these neurons in both nuclei, as well as their contribution to distinct electrophysiological parameters of the serotonergic neurons has been determined yet.
It is known from previous publications that the M-current contributes to the action potential after afterhyperpolarization and spike frequency adaptation regulates the resting membrane potential and controls presynaptic functions (Koyama and Appel, 2006;Tzingounis et al., 2010;Huang and Trussell, 2011;Nigro et al., 2014). It also modulates neuronal membrane potential oscillations (Bordás et al., 2015).
Here we provide evidence that serotonergic neurons possessing M-current can be found both rather in the rostral DR and the dorsal MR. M-current of raphe serotonergic neurons determines firing frequency, spike frequency adaptation, and AHP amplitude, but does not affect input resistance, action potential width, delay of the first action potential, and oscillatory activity.
Details of voltage and current clamp protocols are shown in Figure 1 and based on our previous publications (Bordás et al., 2015;Baksa et al., 2019).
A 3-dimensional map of somata with or without M-current was drawn with NeuroLucida software (MBF Bioscience, Williston, VT, USA) by using Paxinos atlas (Paxinos and Franklin, 2013).
Data are presented as mean ± SEM. Statistical significance was determined with a 2-sample Student's t-test. The level of significance was p < 0.05. Pearson's correlation coefficient (R) was applied to assess correlations between parameters by plotting them against each other and by the linear fit of this dataset.

RESULTS
The topographical organization of M-current possessing raphe serotonergic neurons was characterized and the impact of this current on other electrophysiological properties was described by patching 56 genetically identified serotonergic neurons (n = 31 from the DR and n = 25 from the MR; Figure 1A).
Next, we sought correlations between the presence of M-current and other electrophysiological parameters. We found that M-current can determine average firing rate and spike frequency adaptation. The average firing rate elicited by 100 pA depolarizing step was 5.3 ± 0.9 Hz for neurons possessing M current (n = 13) and 7.56 ± 0.84 Hz for those lacking M-current (n = 16; p = 0.04). Similarly, spike frequency adaptation, quantified by the adaptation index (AI), was also significantly different between the two groups. In neurons having M-current, the adaptation index was 0.77 ± 0.08 (n = 6), whereas this parameter was 0.44 ± 0.04 in those without M-current (n = 11; p = 0.0006; Figures 1E,F, 2E-H).
The relationship between M-current and AHP was also investigated by comparing M-current to maximal, medium, and slow AHPs. Maximal AHP amplitude was 20.27 ± 1.6 mV in the neurons possessing M-current (n = 13) while 15.37 ± 1.44 mV (n = 16) in neurons lacking M-current (p = 0.015). Medium AHP amplitude recorded 100 ms after the AP peak potential was 17.01 ± 2.95 mV in neurons possessing M-current, while in neurons lacking M-current this parameter was 11.65 ± 1.54 mV (p = 0.046). Slow AHP measured 300 ms after the AP peak was 10.87 ± 4.39 mV in neurons having M-current and 3.77 ± 1.25 mV in those without M-current (p = 0.024; Figures 1C, 2I-J).
Certain electrophysiological parameters seemed to be independent of the M-current. The input resistance of cells having M-current was 966.54 ± 118.14 MΩ, while the same parameter was 841.5 ± 64.74 MΩ for the ones lacking M-current, respectively (p = 0.16; Figure 1E).
Serotonergic raphe neurons possessed TTX-resistant high threshold membrane potential oscillations with wide frequency and amplitude ranges. The presence or absence of M-current did not seem to influence these parameters. The oscillatory frequency at the power maximum was 21.67 ± 2.86 Hz for cells having M-current and 19.69 ± 2.21 Hz for lacking M-current (p = 0.29), whereas the power maximum was 1.95 ± 1.31 mv 2 /Hz for cells with and 3.82 ± 3.39 mv 2 /Hz for ones without M-current (p = 0.33; Figures 1G,H).
The topographical organization of neuronal somata lacking or possessing M-current was also evaluated (Figures 3A-C). Neurons possessing M-current are located rostrally in the DR with a greater likelihood (-4.74 ± 0.05 mm from the bregma for the ones having and -4.9 ± 0.05 mm for the ones lacking Mcurrent, p = 0.02) and rather located dorsally in the MR (3.99 ± 0.07 mm from the top for the ones having and 4.36 ± 0.09 mm for the ones lacking M-current; p = 0.007; Figures 3D-G).
Further correlations between electrophysiological parameters besides M-current were also found. First, the frequency of high threshold membrane potential oscillations was inversely  proportional to the input resistance (R 2 = 0.57), whereas power amplitude was directly proportional to the input resistance (R 2 = 0.21; Figures 4A-F) in those cases when the whole trace was included in the analysis. Delay of the first action potential from the onset of the depolarizing impulse from −80 mV was inversely proportional to the maximal firing rate (R 2 = 0.24; Figures 4G,I).

DISCUSSION
We showed that M-current positive neurons are topographically organized: in the DR, they are located rostrally; whereas M-current possessing somata are dorsal in the MR. M-current determines firing frequency, spike frequency adaptation, and the amplitude of the AHP of serotonergic neurons. High threshold membrane potential oscillations also correlate with the input resistance, and the delay of the first action potential is inversely proportional to the maximal firing rate.
M-current contributes to several electrophysiological parameters, such as neuronal excitability, AHP, and spike frequency adaptation. Although these correlations were presented in various brain structures (Koyama and Appel, 2006;Tzingounis et al., 2010;Nigro et al., 2014;Bordás et al., 2015), these have not been demonstrated on raphe serotonergic neurons. As expected, we were able to show that AHP and spike frequency adaptation is influenced by the M-current on these neurons as well.
High threshold membrane potential oscillations are under the regulation of M-current on mesopontine cholinergic neurons (Bordás et al., 2015). To the best of our knowledge, we presented for the first time that high threshold oscillations are present on raphe serotonergic neurons. In contrast with our findings on the Several studies demonstrated that DR and MR serotonergic neurons are not homogenous but rather form different subgroups in terms of development, gene expression, projection, neurochemical markers, as well as in vitro and in vivo electrophysiological properties.
Developmentally, raphe serotonergic neurons arise from six rhombomeres. Subsequently, neurons with different origins have distinct gene expression patterns. Raphe serotonergic neurons form 11 transcriptomic clusters partially based on differences in expression of genes encoding ion channels (Okaty et al., 2015;Ren et al., 2019). Genetically identified subgroups of serotonergic neurons can be distinguished by different ion channel subunit expression.
MR and DR, as well as subregions of DR (including the lateral wing) project to different brain regions, and these projections have topographical organization along the dorsoventral axis (Hale and Lowry, 2011;Ren et al., 2018). Based on their projections and topography, MR and DR serotonergic neurons differ in firing pattern and action potential morphology (Hajós et al., 2007;Crawford et al., 2010;Fernandez et al., 2016). We revealed a topographical distribution of M-current possessing serotonergic neurons, which might also refer to common developmental origin or potential similarities in their projections.
M-current is inhibited by muscarinic acetylcholine receptor activation (Brown and Passmore, 2009); therefore, one can assume that serotonergic neurons with M-current undergo muscarinic cholinergic neuromodulation. Since it is also known that serotonin is capable of influencing M-current via the 5HT 2C receptor (Roepke et al., 2012), one can hypothesize that M-current possessing raphe neurons are subjects of serotonergic autoregulation. These neuromodulatory actions might be more pronounced in the rostral DR and the dorsal MR. As the rostral DR has a role in movementrelated stress tolerance (Greenwood et al., 2003) and the MR is important for adaptation to chronic psychosocial stress (Graeff et al., 1996), neuromodulatory regulation via M-current might be involved in stress adaptation. These findings potentially provide further support to Zhao et al. (2017) that M-current might serve as a target for anxiolytic therapy.

DATA AVAILABILITY STATEMENT
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

ETHICS STATEMENT
The animal study was reviewed and approved by Committee of Animal Research of the University of Debrecen.

AUTHOR CONTRIBUTIONS
TB, AC and AK performed the experiments. AK, KP and BP analyzed the experiments and wrote the article. All authors contributed to the article and approved the submitted version.