Resurgent Sodium Current in Neurons of the Cerebral Cortex

In the late ’90, Dr. Indira Raman, at the time a postdoctoral fellow with Dr. Bruce Bean, at Harvard University, identified a new type of sodium current, flowing through the channels that reopens when the membrane is repolarized. This current, called “resurgent Sodium current,” was originally identified in cerebellar Purkinje neurons and has now been confirmed in around 20 different neuronal types. Since moving to Northwestern University in 1999 to establish her own research group, Dr. Raman has dedicated great efforts in identifying the mechanisms supporting the resurgent Sodium current and how its biophysical properties shape the firing of the different cell types. Her work has impacted greatly the field of cellular neurophysiology, from basic research to translation neuroscience. In fact, alterations in the resurgent sodium currents have been observed in several neuropathologies, from Huntington’s disease to epilepsy. In this Perspective we will focus on the current knowledge on the expression and function of the resurgent Sodium current in neurons of the cerebral cortex and hippocampus. We will also briefly highlight the role of Dr. Raman’s as teacher and mentor, not only for her pupils, but for the whole scientific community.


INTRODUCTION
Neuronal voltage-gated Sodium channels (VGSCs) are responsible for the large and transient inward current (I NaT ) underlying the upstroke of the action potential (AP) (Hodgkin and Huxley, 1952a). After opening, VGSCs quickly inactivate and require strong membrane hyperpolarization to become available (Hodgkin and Huxley, 1952b). Many neurons express subthreshold noninactivating Sodium currents: a persistent sodium current (I NaP ) flowing through non-inactivated VGSCs, and resurgent Sodium current (I NaR ) that activates upon membrane repolarization (Stafstrom et al., 1982;Crill, 1996;. Subthreshold sodium currents can powerfully shape spike after-potentials and repetitive firing Khaliq et al., 2003;Yamada-Hanff and Bean, 2013). I NaR has first been described in isolated cerebellar Purkinje neurons, where it contributes to the generation of complex spikes and repetitive spontaneous firing Khaliq et al., 2003). I NaR has been identified in many different cell-types throughout the brain, with conserved biophysical properties (Lewis and Raman, 2014). Thanks to the elegant work of Dr. Raman and collaborators, I NaR , together with I NaT , is the best characterized component of the Sodium current. Indeed, Dr. Raman's original description highlighted the role of I NaR in bursting in Purkinje neurons and developed a model for its generation hypothesizing the existence of a blocking particle that enters the pore at open states and is released upon repolarization Bean, 1997, 2001). Later, Dr. Raman led her laboratory on a quest to decipher the structural and molecular mechanisms of I NaR , discovering the interaction of alpha and beta subunits of VGSCs underlying open channel block and the identity of the blocking particle (Grieco et al., 2005;Raman, 2007, 2010;Aman et al., 2009;Bant and Raman, 2010;Raman, 2011, 2013). Following its description in Purkinje cells, the discovery of I NaR in other cell-types has propelled research on the molecular underpinnings and physiological role of I NaR in the brainstem, basal ganglia and peripheral sensory neurons (Do and Bean, 2003;Cummins et al., 2005;Enomoto et al., 2006;Barbosa et al., 2015), as well as its involvement in disorders such as pain, long QT syndrome and epilepsy (Jarecki et al., 2010;Hargus et al., 2011). In this review we want to focus on I NaR in neurons of the cerebral cortex, where the work of Dr. Raman inspired our own first steps in the field of neurophysiology. The authors started their journey in neuroscience by studying the expression of I NaR in cortical neurons and were deeply inspired by the solidity and elegance of the experimental work of Dr. Raman. One of us (GQ) had the chance to attend Dr. Raman's "Great experiments in Cellular Neurophysiology" course at Northwestern University, experiencing her outstanding teaching.

EXPRESSION OF I NaR IN CORTICAL AND HIPPOCAMPAL NEURONS
Soon after  first described it in cerebellar Purkinje cells, the INaR has been identified in several cell types of the cerebellum, brainstem, basal ganglia, and dorsal root ganglia (Lewis and Raman, 2014).  also described the absence of I NaR in acutely isolated pyramidal neurons of the CA3 region. In the cerebral cortex, I NaR was first reported in layer II of the rat perirhinal cortex (Castelli et al., 2007a). This first report was followed shortly by an examination of the expression of I NaR across the hippocampus and parahippocampal region (Castelli et al., 2007b). In the perirhinal and entorhinal cortices 75-100% of excitatory neurons express I NaR depending on layer localization. In these regions the resurgent conductance amounts to 1.5-3% of the conductance of the transient component. As opposed to other types of neurons, the channels responsible for the I NaR in cortical neurons are enriched in the axon initial segment. Indeed, application of TTx to the axon initial segment of perirhinal neurons abolished I NaR to a larger extent than when applied to the soma/proximal apical dendrite (Castelli et al., 2007a). Moreover, patch clamp recordings from acutely isolated cortical neurons rarely show I NaR (Castelli et al., 2007a). In the parahippocampal region, I NaR is expressed most prominently by layer II excitatory neurons of the medial entorhinal cortex (MEC) (Castelli et al., 2007b). In MEC layer II, I NaR was found in all recorded neurons, with an amplitude representing 3.6% that of I NaT , the second largest among neurons of the cerebral cortex (Castelli et al., 2007b;Nigro et al., 2012). MEC layer II neurons express all three components of the Sodium current: I NaT , I NaP , and I NaR . The developmental expression of the I NaR in MEC layer II follows a trajectory that is independent from that of the other components, increasing steadily in amplitude from postnatal day (P) 5 to P10 (Nigro et al., 2012). In parallel, the percent of neurons expressing I NaR also increases in the same developmental window (Nigro et al., 2012).
In the hippocampus I NaR is expressed in subpopulations of excitatory neurons in the dentate gyrus (60%), ventral CA1 (40%), and the majority of subicular neurons (Castelli et al., 2007b;Barker et al., 2017). Interestingly, I NaR does not seem to be expressed by pyramidal neurons of the dorsal hippocampus or in CA3 pyramidal neurons recorded from brain slices (Castelli et al., 2007b). Future studies correlating I NaR expression to transcriptomic cell types might shed light on the molecular identity of neurons expressing I NaR and the molecular mechanisms underlying its expression in the cortex (see below).

CONTRIBUTION OF I NaR TO FIRING PROPERTIES OF CORTICAL NEURONS
Several neuron types expressing I NaR show spontaneous firing with relatively high firing rates and bursting, e.g., Purkinje cells, subthalamic neurons, neurons of the cerebellar nuclei (Lewis and Raman, 2014). I NaR endows these cell-types with the ability to produce repetitive spiking spontaneously (i.e., in absence of synaptic activity) at high frequencies by preventing fast inactivation through open channel block (Raman and Bean, 2001;Khaliq et al., 2003). At depolarized membrane potentials, the open channel block competes with fast inactivation, and during repolarization, the transition from open channel block to open states allows Sodium ions to flow and initiate a new cycle of spiking (Raman and Bean, 2001). Excitatory neurons of the cerebral cortex do not produce spontaneous repetitive firing, nor they reach high firing frequency, but fire trains of action potentials with different degrees of adaptation. I NaR has been shown to contribute to repetitive firing in layer II pyramidal neurons of the rat perirhinal cortex (Castelli et al., 2007a). These neurons produce repetitive firing up to 30 Hz upon depolarizing current injection, a firing frequency much lower than other I NaR expressing neurons outside the cerebral cortex. However, even at those firing frequencies I NaR contributes to most of the Sodium current during the interspike interval promoting depolarization and repetitive firing (Castelli et al., 2007a).
By injecting AP waveforms (recorded in current clamp) in voltage clamp experiments,  demonstrated that I NaR provides a major contribution to the generation of complex spikes in Purkinje neurons. Inspired by these original experiments, Alessi et al. (2016) tested the contribution of different ionic conductances to the generation of the depolarizing afterpotential (DAP) in MEC layer II stellate cells. The authors described two mechanisms generating DAPs in these cells acting at different membrane voltages. At hyperpolarized holding potentials T-type Calcium channels provide most of the depolarization following the fast afterhyperpolarization (fAHP). However, at holding voltages closer to the resting potential, subthreshold Sodium currents (I NaP and I NaR ) are the major contributors to the DAP (Alessi et al., 2016). During spatial navigation, MEC layer II stellate cells show a spatially modulated firing pattern characterized by regularly spaced firing fields arranged in a hexagonal matrix, as described for grid cells (Fyhn et al., 2004;Hafting et al., 2005;Domnisoru et al., 2013;Schmidt-Hieber and Hausser, 2013;Rowland et al., 2018). MEC stellate cells with grid firing patterns also show a higher probability of generating bursts of APs during navigation (Latuske et al., 2015;Bant et al., 2020). Interestingly, bursting probability and amplitude of I NaR show a gradient along the dorso-ventral axis of the MEC that correlates with the gradient in spacing and field size of grid cells along the same axis (Bant et al., 2020). The authors observed that using bursts increased the performance of decoding the animal's position during navigation as compared to isolated spikes. The higher information content of burst points to a cellular mechanism to maximize signal-to-noise ratio in dorsal MEC grid cells (Bant et al., 2020). We would like to highlight that these interesting results were obtained by Dr. Jason Bant, a previous student of Dr. Raman, teaming up with Dr. Lisa Giocomo, who pioneered the study of the topographic organization of biophysical properties in MEC.

MOLECULAR MECHANISM OF I NaR IN CORTICAL NEURONS
Patch clamp experiments from Purkinje neurons obtained from Nav1.6 (Scn8a) null mice showed that this subunit contributes to most of the I NaR in these neurons . However, the alpha subunit on its own cannot generate I NaR . Indeed Nav1.6 is also expressed in CA3 neurons that do not express I NaR . The mechanism generating I NaR involves a blocking particle that interacts with the alpha subunit (Raman and Bean, 2001). In a series of elegant experiments, Grieco et al. (2005) demonstrated that the blocking particle consists of the beta subunit Navβ4 (Grieco et al., 2005). Knockdown expression of the Navβ4 in cerebellar granule cells and peripheral sensory neurons strongly reduced I NaR , further highlighting the role of this beta subunit in generating I NaR (Bant and Raman, 2010;Barbosa et al., 2015). In cortical neurons, Nav1.6 is expressed at high levels in the axon initial segment correlating with the subcellular expression of I NaR in these neurons (Castelli et al., 2007a;Royeck et al., 2008;Buffington and Rasband, 2013). On the other hand, most of the cortical cells in which I NaR has been observed express Navβ4 at very low levels (Buffington and Rasband, 2013) or completely lack scn4b expression (Yu et al., 2003;Lewis and Raman, 2014; Table 1). In situ hybridization (ISH) data from the Allen Brain Institute suggest a low level of expression in a subpopulation of layer II stellate cells in the dorsal MEC, which correlates with the dorso-ventral gradient of I NaR in these neurons (Lein et al., 2007;Bant et al., 2020). In the hippocampus, scn4b is expressed in the dorsal CA1, where I NaR is not expressed (Castelli et al., 2007b). The absence of I NaR in CA1 pyramidal neurons might be explained by a slicing artifact by which the axon initial segment of CA1 neurons was lost. However, the recordings in CA1 were performed from coronal slices that preserve cellular integrity (Dougherty et al., 2012). Additionally, scn4b expression might not be sufficient for I NaR expression in all cell types. A recent single cell transcriptomic analysis of the whole cortex and hippocampus by the Allen Brain Institute (Yao et al., 2021) showed that expression of scn4b is restricted to a handful of layer V neurons, none of whom are known to express I NaR . The discrepancy between ISH and single cell transcriptomic in MEC might be due to the sparseness of scn4b expressing neurons in this region (Figure 1). Additionally, this discrepancy may arise because of sample processing differences between ISH and single cell transcriptomics. In ISH, the tissue remains intact and preserves cytosolic compartments, like distal neuronal processes, that are otherwise destroyed in tissue homogenization for single cell RNA processing. Alternatively, single cell RNA sequencing intrinsically yields low rates of capture (Zheng et al., 2017) and read efficiency which must be accounted for computationally (Galfrè et al., 2021). Tissue homogenization is similar between bulk and single cell RNA sequencing and differences in technique should yield similar results. Bulk transcriptomic analysis comparing the dorsal and ventral MEC from adult mice in Ramsden et al. (2015), however, demonstrates that scn4b is significantly upregulated in the dorsal MEC at P60 (Figure 1B). Moreover, it is expressed at a level which drives distance-based clustering separation between the dorsal and ventral MEC (Ramsden et al., 2015). This discrepancy in the single cell dataset may not be related to weakness in gene detection or homogenization methods. Instead, gene expression differences in the dorsal-ventral axis in the adult single cell dataset may not have been considered and thus cells in ventral layer II may mask scnb4 expression in dorsal layer II. Interestingly, a subgroup of pyramidal tract (PT) projecting layer V neurons showed a significant expression of scn4b, suggesting the possible expression of I NaR in this type of cells. The expression of scn4b in layer V cortical neurons shows a gradient along the rostrocaudal axis of the telencephalon. Strong scn4b expression in layer V is evident in motor areas, sensory cortices, anterior cingulate and retrosplenial cortex. Scn4b seems to be absent in association areas, including the insula, parahippocampal area and medial prefrontal areas (Figures 1C-G). Future electrophysiological experiments might corroborate the presence of I NaR in these cortical neurons. The molecular nature of the blocking particle in cortical neurons also remains enigmatic, particularly in the perirhinal cortex, ventral CA1 and dentate gyrus. Current approaches to characterize the transcriptomic and electrophysiological profile of neurons will shed light on the molecular underpinnings of I NaR in cortical neurons (Cadwell et al., 2016).

I NaR IN EPILEPSY
The contribution of I NaR to repetitive spiking, DAP and burst generation suggests that it might play a role in disorders where neuronal excitability is altered, such as epilepsy. This component of the Sodium current is expressed in cortical areas that are strongly affected in temporal lobe epilepsy (TLE), such as MEC, perirhinal cortex, hippocampus and subiculum. I NaR expressing MEC layer II neurons are spared in temporal lobe epilepsy, but they show an increased excitability (Bear et al., 1996). The increased excitability is in part due to a reorganization of the synaptic network (Kuman et al., 2007), however intrinsic mechanisms are also at play. Indeed, in the absence of synaptic inputs, layer II MEC neurons show a higher firing rate in response to current injection in a rat model of TLE (Hargus et al., 2011). These changes in intrinsic excitability correlate with an increased amplitude of all components of the Sodium current (Hargus et al., 2011). The molecular underpinnings of the increased excitability reside in an increased expression of the Nav1.6 subunit. Indeed, pharmacological blockade of Nav1.6 channels with 4,9-anidro-tetrodotoxin rescued the excitability of TLE MEC layer II neurons to levels like control neurons (Hargus et al., 2013). Moreover, similar alterations in intrinsic excitability and Sodium currents are recapitulated in mice carrying a mutated Nav1.6 isoform Pan and Cummins, 2020). An increased excitability and augmented I NaR and I NaP have recently been described in human excitatory cortical neurons differentiated from pluripotent stem-cells obtained from patients affected by early infantile epileptic encephalopathy type 13 (Scn8a-related epilepsy) (Tidball et al., 2020).

DISCUSSION
In the current perspective we aimed at reviewing the state-ofthe-art of the research on the I NaR in neurons of the cerebral cortex. The expression of I NaR has been reported in nine cortical cell-types, and we propose its expression in a population of layer V PT neurons based on their expression of Navβ4. The contribution of I NaR to the firing behavior of cortical neurons has been well demonstrated in the perirhinal and entorhinal cortices. In these areas, I NaR supports repetitive firing, DAPs and bursting. Bursting in LII MEC neurons has been proposed to maximize signal-to-noise ratio in grid cells and might represent a cellular mechanism for a reliable transmission of spatial information to the hippocampus (Bant et al., 2020).
The molecular mechanisms underlying I NaR in cortical neurons are yet to be described. The molecular identity of the blocking particle in many cortical cell types remains unknown and future studies describing the correlation of trascriptomics and electrophysiology will allow to uncover potential candidates. These studies will also provide molecular targets for pharmacological treatments of epileptic encephalopaties involving I NaR .
With this perspective, we wished to emphasize the pivotal influence of Dr. Raman's work on the mechanism of openchannel block, molecular identity of the blocking particle, and physiological role of I NaR in cerebellar Purkinje neurons. Her findings sparked the quest for I NaR in the cerebral cortex and provided the foundations for our current understanding of the role of I NaR in the firing properties of cortical neurons.
In addition to her scientific contribution, Dr. Raman has also always been interested in forging new generations of scientists. Anyone who had the privilege of attending a class or a lecture given by Dr. Raman knows it will not be boring. Her scientific knowledge will engage you and her ability to introduce interesting anecdotes on the "behind the scenes" will enchant you. Her outstanding teaching abilities have been recognized by multiple awards received at Northwestern University, but her drive to good mentoring and good science has not stopped in Chicago. Dr. Raman has contributed to a series of feature articles titled "Living science" published in eLIFE from 2015 to 2019 (Raman, 2015a(Raman, ,b, 2016(Raman, , 2017(Raman, , 2019. In addition, her piece on good mentoring (Raman, 2014) faithfully describes the challenges faced by both the trainee and the advisor.  Ramsden et al. (2015). Gene expression driven separation between ventral and dorsal medial entorhinal cortex from P60 mice. Scn4b is highlighted by a black rectangle and appears as one of 8 genes significantly enriched in the dorsal adult MEC. Heatmap displays 34 genes significantly up or downregulated in either ventral or dorsal MEC samples (p-val. < 0.05, log 2 FC threshold ± 1.5, Benjamini-Hochberg corrected; hierarchical clustering performed by pairwise-Spearman correlation matrix). (C-G) ISH experiments from the Allen Institute showing the expression of scn4b across the cerebral cortex. (C) Scn4b is expressed in the dorsal MEC. (D-F) Expression of scn4b in the isocortex decreases from rostral (C) to caudal (F) regions. (G) Magnification of the somatosensory (SS) area highlighted by a dotted box in (C). CA, Cornu Ammonis; DG, dentate gyrus; MEC, medial entorhinal cortex; IL, infralimbic; PrL, prelimbic; ACA, anterior cingulate area; MO, motor area; SS, somatosensory area; Ins, insula; AUD, auditory area; TeA, temporal association area; PER, perirhinal cortex; PIR, piriform cortex; RSP, retrosplenial area; VIS, visual area; LEC, lateral entorhinal cortex. Image credit: For (A): adapted from Nigro et al. (2012) with permission from Elsevier. For (C-G) Allen Institute. 2004 Allen Institute for Brain Science. Allen Mouse Brain Atlas. Available from: https://mouse.brain-map.org/.
Dr. Raman contribution to science and her effort to make the scientific world a better place for young trainees and women has left a mark on the authors as well as, we are sure, on all the people that had trained with her and have known her.

AUTHOR CONTRIBUTIONS
GQ and MN conceived the study and analyzed the data. KD analyzed the data. All authors wrote the manuscript.

FUNDING
This study has been supported by the Centre of Excellence scheme of the Research Council of Norway-Centre for Neural Computation (Grant No. 223262), and the Kavli Foundation. This project has received funding from the European Union's Horizon 2020 Research and Innovation Programme under the Marie Skłodowska-Curie Grant Nos. 885955 to MN and 839988 to GQ.