SCN2A-linked myelination deficits and synaptic plasticity alterations drive auditory processing disorders in ASD

Autism spectrum disorder (ASD) is a neurodevelopmental disorder characterized by complex sensory processing deficits. A key unresolved question is how alterations in neural connectivity and communication translate into the behavioral manifestations seen in ASD. Here, we investigate how oligodendrocyte dysfunction alters myelin plasticity and neuronal activity, leading to auditory processing disorder associated with ASD. We focus on the SCN2A gene, an ASD-risk factor, to understand its role in myelination and neural processing within the auditory nervous system. Through transcriptional profiling, we identified alterations in the expression of myelin-associated genes in Scn2a conditional knockout mice, highlighting the cellular consequences engendered by Scn2a deletion in oligodendrocytes. The results reveal a nuanced interplay between oligodendrocytes and axons, where Scn2a deletion causes alterations in the intricate process of myelination. This disruption instigates changes in axonal properties, presynaptic excitability, and synaptic plasticity at the single cell level. Furthermore, oligodendrocyte-specific Scn2a deletion compromises the integrity of neural circuitry within auditory pathways, leading to auditory hypersensitivity. Our findings reveal a novel pathway linking myelin deficits to synaptic activity and sensory abnormalities in ASD.

Myelination is critical for brain connectivity and temporal processing in the developing brain by coordinating effective axonal conduction and neurotransmission [4][5][6] .Recent studies using neuroimaging and genetic analysis have shown atypical white matter development and abnormal myelination in individuals with ASD and animal models 7,8 .However, the speci c impact of myelin alternations on neural connectivity, communication, and ultimately behavioral manifestations is not fully delineated.Notably, white matter integrity was altered in speci c brain regions of humans with ASD, which are associated with sensory processing 9,10 .Auditory processing abnormalities including auditory hypersensitivity are well-documented in individual with ASD 9 .Auditory hypersensitivity, as an excessive or abnormal response to auditory stimuli, may arise from alterations in sensory gain, neuronal activity, and excitation/inhibition (E/I) balance in the auditory circuitry [11][12][13] .These alterations could be potentially linked to abnormalities in myelination.
ASD has been associated with altered gene expression related to oligodendrocyte (OL) maturation and myelination 8 .OLs, the myelin producing glia, play a more complex role in neural development.Beyond facilitating myelination, immature OLs monitor neuronal activities and in uence their functional development by modulating axonal segments 14,15 .Moreover, OLs engage in dynamic interaction with neurons, contributing to activity-dependent myelination, which in turn supports neural circuit plasticity 16- 18 .Thus, OL dysfunction can disrupt neuron-glia interactions and alter ion channel expression along the axon, ultimately affecting neural plasticity 5,19 .However, the impact of OL dysfunction on neuronal activity and synaptic plasticity at the single-cell level as well as how these alterations contribute to auditory processing de cits in ASD remains largely unexplored.Additionally, the genetic and cellular interactions of OLs underpinning the neurodevelopmental feature of ASD such as auditory processing de cits require further studied.
One gene of particular interest is Scn2a, which encodes the alpha subunit of the voltage-gated Na + channel 1.2 (Na v 1.2), is highly linked to neurodevelopmental disorders including ASD [20][21][22] .Loss-offunction mutations in Scn2a impair dendritic excitability, leading to synaptic dysfunction and behavioral de cits [22][23][24] .Expression of Scn2a in non-neuronal cells, speci cally in oligodendroglia, has been well documented [25][26][27] .Notably, a transcriptional pro le of OL lineage cells from mouse brain showed the highest levels of Scn2a expression are in newly formed OLs, an immature OL population, and a subpopulation of Scn2a-expressing OLs exhibits membrane excitability during early postnatal development 27 .However, the cell autonomous role of Scn2a in myelination and neuron-glia interaction related to ASD remains to be elucidated.In this study, we investigated how Scn2a deletion speci cally in OLs impacts the interplay between myelination and synaptic activity, ultimately leading to auditory processing disorders in ASD.

Results
Scn2a cKO mice exhibit auditory hypersensitivity without changes in peripheral function.
SCN2A is identi ed as a high-risk gene associated with ASD, wherein abnormal sensory processing including auditory hypersensitivity is frequently observed 9,28 .In our study, mice with Scn2a haploinsu ciency (Scn2a +/-mice) 24,29 exhibited a distinct and pronounced startle response to sudden, loud auditory stimuli (Supplementary Fig. 1A).This heightened startle response is consistent with the hypersensitivity commonly observed in individuals with ASD and in ASD mouse models 24,30 .Intriguingly, we also observed myelin de cits within the auditory nervous system of these Scn2a +/-mice (Supplementary Fig. 1B-C).To investigate the cell-autonomous role of oligodendroglia Scn2a in the interplay between myelination and the auditory processing disorders, we utilized Scn2a conditional knockout (cKO) mice (Pdgfra CreERT ; Scn2a / ) 27 .First, we examined whether oligodendroglia speci c deletion of Scn2a causes auditory processing abnormalities.In an acoustic startle re ex (ASR) test, Scn2a cKO mice displayed larger startle re exes in response to stimuli above 95 dB SPL, comparing to control (n = 15 control vs n = 10 cKO mice, two-way ANOVA test, p = 0.0012, Fig. 1A).Scn2a cKO mice showed stronger startle at loud sound stimulation of 105 dB SPL (p = 0.0214) and 115 dB SPL (p = 0.0077, multiple comparison from two-way ANOVA with Bonferroni correction).The increased amplitude of the startle re ex in Scn2a cKO mice indicates heightened sensory perception, which is considered hypersensitivity.In addition, Scn2a cKO mice exhibited alterations in the pre-pulse inhibition (PPI) test.Typically, a weaker sensory stimulus (or pre-pulse) inhibits the reaction to a subsequent strong sensory event.In control, the inhibition rate drastically increased with the strength of pre-pulse stimulation.In contrast, Scn2a cKO mice demonstrated a diminished capacity to inhibit the startle re ex, particularly in response to the pre-pulse at 81 dB SPL sound stimulation (35.9 ± 3.22% PPI in n = 18 control vs 24.7 ± 3.19% PPI in n = 11 cKO mice, p = 0.0377, multiple comparison from two-way ANOVA with Bonferroni correction, Fig. 1B).The alteration in startle re ex to sound stimulation suggests that Scn2a cKO mice have an impaired sensory gating mechanism, a feature often seen in ASD 13,24 .
To ascertain the extent of auditory processing abnormalities in the sub-cortical regions of Scn2a cKO mice, we utilized in vivo auditory brainstem responses (ABRs) to assess the sum of evoked potential responding sound stimuli along the auditory pathway, particularly in the brainstem (Fig. 1C).There was no signi cant difference in ABR threshold (Fig. 1D) and the peak latencies between control and Scn2a cKO mice (data not shown).However, the amplitude of ABRs in response to 80 dB SPL (click) was signi cantly increased in Scn2a cKO mice compared to control (p < 0.0001, two-way ANOVA).The amplitude of peak II (2.44 ± 0.166 µV, n = 39 control vs 3.30 ± 0.231 µV, n = 26 cKO, p = 0.0002, multiple comparison from two-way ANOVA with Bonferroni correction) and peak III (2.09 ± 0.129 µV in control vs 2.70 ± 0.149 µV in cKO, p = 0.0191) were signi cantly larger in Scn2a cKO mice compared to those of control.There was a trend toward an increase, but no signi cant changes were observed in peak I (p = 0.352), peak IV (p = 0.3248) and peak V (p > 0.9999, Fig. 1E).The increased amplitude of ABRs suggested that Scn2a cKO mice exhibit the hypersensitivity to sound.To evaluate whether there were alterations in central gain, we calculated the ratio between peak amplitudes II to IV relative to peak I. Notably, our analysis indicated a prevalent trend where the ratios of peak II to peak I and peak III to peak I were larger in Scn2a cKO mice, despite the absence of statistical signi cance (p > 0.05, two-way ANOVA, Fig. 1F).We examined the possibility if the hypersensitivity could have originated from alterations in peripheral hearing sensitivity.There was no signi cant change in ABR threshold, peak I amplitude, and distortion product otoacoustic emissions (DPOAE, Fig. 1G), demonstrating intact peripheral hearing.
Collectively, the results demonstrated robust behavioral phenotypes in Scn2a cKO mice, speci cally auditory hypersensitivity and impaired sensory gating, mirroring sensory processing abnormalities in ASD.
Myelin related genes were signi cantly downregulated in mature OL from Scn2a cKO mice To understand the impact of Scn2a deletion on OL development and myelination within the auditory brainstem circuitry, we characterized the single cell transcriptional pro les of mouse auditory brainstem using single nucleus RNA sequencing (snRNA-seq).After the exclusion of low-quality cells, we obtained a total of 3083 reliable nuclei for the analysis: 989 nuclei from the control and 2094 nuclei from Scn2a cKO mice.Shared Nearest Neighbor (SNN) clustering identi ed 16 distinct clusters on Uniform Manifold Approximation and Projection (UMAP) plot, including clusters of neurons (Rbfox3 positive), astrocytes (Aldh1l1 positive), microglia (Itgam positive), and OL lineage cells (Pdgfra and Mog positive, Supplementary Fig. 2).Speci cally, OL lineage cells consisted of ve clusters: Oligodendrocyte precursor cells (OPCs), Differentiating OL1 (Diff.OL1), Differentiating OL2 (Diff.OL2), Premature OL, and Mature OL.To identify the OL lineage, we analyzed OL-speci c genes, such as Pdgfra, Mbp and Mog, expressed in OPCs and mature OLs, respectively (Fig. 2A, Supplementary Fig. 2).Given the dynamic nature of OL differentiation, we utilized trajectory analysis to examine this continuous differentiation.The result of trajectory analysis from those populations displayed a narrow differentiation path connecting from OPCs to mature OLs, that similar to patterns in a previously reported study 25 .The result indicates that OL development follows a clearly de ned trajectory, which was outlined by genetic markers speci c to the OL lineage.
Although there was no statistically signi cant difference in the proportion of OL lineage cells between control and Scn2a cKO mice, a trend towards a reduction in the mature OL population and a slight increase in OPCs and differentiating OLs was observed in the Scn2a cKO mice (Fig. 2B).When OL clusters from each genotype were visualized with different colors in the UMAP scatter plot, a distinct difference was observed in mature OLs (Fig. 2C).To quantify these differences further, we identi ed genes exhibiting signi cant differential expression between control and Scn2a cKO mice (adjusted p with Bonferroni correction < 0.05).Notably, 96 differentially expressed genes (DEGs) were detected in mature OLs between control and Scn2a cKO mice (Fig. 2C, Table S1-5).This result reveals signi cant alterations in gene expression patterns within mature OLs, suggesting the cascading implications of Scn2a deletion in the orchestration of gene expression within mature OLs.Furthermore, to examine the expression level of myelin-associated genes in mature OLs, we employed AmiGo2, a gene ontology data base, to pick genes for comparison.Among the myelin-related genes (total 391 genes), 33 DEGs were identi ed (p < 0.05).Notably, the volcano plot showed that Mobp, encoding myelin associated oligodendrocyte basic protein (MOBP), was signi cantly decreased in Scn2a cKO mice (Fig. 2D).The String analysis revealed that 30 genes out of 33 DEGs have a signi cant interconnection and can be functionally characterized into four main categories: cell skeletal structure, cell junction, membrane, and ion channel (Fig. 2E).In addition, the most signi cant down regulated genes; Mobp 31 , Ncam1 32 , Ptn 33 , Gnao1 34 , Mtmr2 35 , Abca2 36 , Arhgef10 37 and Mpdz 38 are known to be related to myelination, cell growth, and neurological diseases (Fig. 2F).Taken together, the transcriptional pro les highlight a pronounced differential expression of myelin-associated genes in mature OLs from Scn2a cKO mice.
Loss of Scn2a impairs myelination and alters axon caliber size in the auditory brainstem during postnatal development Following the gene expression analysis, we assessed the structural aspects of myelinated axons using transmission electron microscopy (TEM, Fig. 3A).This evaluation focused on determining myelin thickness and axon caliber size within axon bundles located in the medical nucleus of trapezoid body (MNTB) in the auditory brainstem.Ultrastructural analysis revealed signi cant myelin de cits in Scn2a cKO mice, as evidenced by an elevated g-ratio providing critical insights into the intricate alterations occurring at the structural level.The g-ratio is an indicator of myelin thickness measured by the inner radius (r) divided by the outer radius (R) of axon in the corresponding axon bundles.To examine alteration in distribution patter of the g-ratio, linear regression analysis was used in which slope and yintercept were compared between groups.While the regression slopes of control and Scn2a cKO were comparable, y-intercept was signi cantly higher in cKO, indicating a decreased myelin thickness in Scn2a cKO (Fig. 3B).Axon calibers showed a larger inner diameter in Scn2a cKO compared to control (1.39 ± 0.03 µm, n = 1154 axons in 10 controls vs 1.48 ± 0.03 µm, n = 786 axons in 4 Scn2a cKO mice, p < 0.0001, Mann-Whitney U-test, Fig. 3C).The right shift of the cumulative frequency curve indicates an increase in axon caliber size in Scn2a cKO compared to control.Intriguingly, the g-ratio was signi cantly increased in Scn2a cKO, indicating a decrease in myelin thickness of axons in the MNTB area (0.82 ± 0.002 vs 0.84 ± 0.002, p < 0.0001, Mann-Whitney U-test, Fig. 3D).In addition, there was also a right shift in the cumulative frequency distribution of g-ratio in Scn2a cKO, indicating g-ratio was larger in Scn2a cKO across examined axons.The results suggest that the targeted deletion of Scn2a in immature OLs detrimentally affects myelination, concurrently altering the axon caliber size within the auditory brainstem.Notably, the myelin de cits identi ed in Scn2a cKO mice were consistent with signi cant alterations in the expression of myelin-related genes in mature OLs from snRNA-seq data (Fig. 2).For further validation, we tested effect of Scn2a deletion on myelination in another OL-speci c Scn2a cKO mice, in which Scn2a was removed from Sox10 expressing OLs, targeting all OL lineage cells (Sox10 CreER ; Scn2a / ) 39 , also displayed similar alterations in myelin and axon caliber size (Supplementary Fig. 3).The observed myelin de cits in Scn2a cKO mice can be attributed to both impaired OL development and alterations in myelinassociated genes.A reduction in the differentiation of OPCs to OLs, evidenced by a decreased mature OLs (CC1 + OLs) in Scn2a cKO mice, further supports these ndings 27 .Together, these results collectively underscore the pivotal role of Scn2a in myelination and the maintenance of axonal integrity in the auditory brainstem.

OL-speci c loss of Scn2a alters myelinated segments and Na + channel expression at distal axons in the MNTB
Given the signi cant alterations in myelination, we postulated that myelin changes might manifest in broader structural and functional aberrations within the neural circuitry.During development, immature OLs play a pivotal role by interacting with axons, which in turn determines the length of myelinated axon segments and nodes 5 .To elucidate the consequences of Scn2a deletion in OLs on these interactions, we examined the long-myelinated axon extending from the cochlear nucleus to the MNTB and its terminal, the calyx of Held.Myelinated axon segments, the internode, can be visualized by myelin proteolipid protein (PLP) and caspr (a paranodal marker) anked by two sodium channel clusters (Pan Na v ) denoting nodes of Ranvier (Fig. 4A).Thus, we assessed the length of the internode by measuring the distance between two Na v clusters (Fig. 4B).A structural analysis of the calyx axons showed a shorter internode adjacent to the heminode at the distal axon in Scn2a cKO mice (42.01 ± 2.31 µm, n = 24 axons from 5 control vs 32.46 ± 1.83 µm, n = 46 axons from 5 cKO mice, p = 0.0024, unpaired t-test, Fig. 4C).However, no signi cant difference was observed in the length of subsequent internodes (p = 0.6703, unpaired t-test, Fig. 4C).The reduced length of the last myelinated segment in Scn2a cKO suggests that the myelinated segments near the nerve terminal may be underdeveloped or experience delayed development in these mice 40 .Furthermore, to determine if the thinner and shorter myelin alters ion channel distribution along the distal axon in Scn2a cKO mice, we examined the patterns of voltage-gated sodium channel (Na v ) expression along the calyx of Held axon.Intriguingly, we found a dispersed Na v channel expression at heminodes in Scn2a cKO mice (2.51 ± 0.11 µm, n = 35 axons from 5 control vs 3.29 ± 0.17 µm, n = 67 axons from 7 cKO mice, p = 0.0018, unpaired t-test, Fig. 4D).Dispersed Na v channels at the heminode have also been previously observed in dysmyelinated axon terminals 5 .
Nevertheless, Na v clusters at other nodes did not show any signi cant alteration (p = 0.1209 for node 1 and p = 0.616 for node 2, unpaired t-test, Fig. 4D).This result highlights the pivotal role of Scn2a in OL development, emphasizing its in uence on axon-OL interactions and Na v channel distribution at the distal axons in synapse-rich areas of the auditory brainstem during postnatal development.
The intrinsic excitability at the nerve terminal was altered in Scn2a cKO mice Alterations in ion channel distribution at the heminode in uence the intrinsic properties of nerve terminal and presynaptic ring pattern.We therefore examined intrinsic excitability at the calyx terminal by recording presynaptic APs, evoked by depolarizing current injection.AP waveform analysis revealed that AP threshold was signi cantly higher in Scn2a cKO mice (-49.0 ± 0.76 mV, n = 32 cells from 17 control mice vs -46.9 ± 0.56 mV, n = 43 from 12 cKO mice, p = 0.0217, unpaired t-test, Fig. 5A, B), with no substantial changes in AP amplitude (p = 0.0912, unpaired t-test, Fig. 5B).The maximal dV/dt was lower in Scn2a cKO mice (468.1 ± 28.47 mV/ms, n = 32 control vs 411.8 ± 30.17 mV/ms, n = 43 Scn2a cKO, p = 0.0282, Mann-Whitney U-test), indicating a slower AP rise time in Scn2a cKO mice than control.Furthermore, the minimal dV/dt was signi cantly reduced in Scn2a cKO mice (-359.80 ± 22.60 mV/ms, n = 28 control vs -292.9 ± 21.75 mV/ms, n = 36 Scn2a ckO mice, p = 0.0061, Mann-Whitney U-test), indicating a slower repolarization.Other parameters in Scn2a cKO, despite not reaching statistical signi cance, showed trends of alteration: an increased rheobase current (p = 0.058, Mann-Whitney Utest) and a broadened AP width (p = 0.1437, unpaired t-test, Fig. 5B).An elevated threshold, a slower rising of the spike, and a slower repolarizing at the calyx terminal indicate that the nerve terminal has reduced excitability.To further support these ndings, we also tested the intrinsic excitability by counting the number of APs evoked by incremental current injections.The input-output curves showed a distinct shift, demonstrating fewer APs per current step in Scn2a cKO (p = 0.0176, Two-way repeated measures ANOVA, Fig. 5C).Thus, the result suggests that alterations in Na v channel expression around myelinated segments have a profound impact on presynaptic excitability in Scn2a cKO mice.Intriguingly, another observation is that in ~ 20% of presynaptic recordings from Scn2a cKO (14 of 45 cells), the calyx terminal displayed aberrant spikes or/and spontaneous spikes when the resting membrane potential was around − 65 mV (Supplementary Fig. 4).However, spontaneous ring and aberrant spikes have not been observed before in ex vivo recording from WT with the same experimental setting.The results demonstrated that myelin de cits at the distal axon caused by OL-Scn2a loss can generate asynchronous and abnormal spikes along the distal axon.

Conduction failures impaired the reliability and delity of high frequency spikes at the nerve terminal in Scn2a cKO mice
To understand the implications of thinner myelin sheath, larger axon, and ion channel expression alterations at the distal axon observed in Scn2a cKO, we evaluated AP propagation along the distal axon and the delity of APs at the nerve terminal (Fig. 6A).In control, the calyx terminal e ciently followed high-frequency stimulation and demonstrated spikes without failures at both 50 Hz and 100 Hz.
Conversely, the calyx terminal in Scn2a cKO mice exhibited AP failures at 100 Hz and had drastically more failures at higher frequencies (Fig. 6B, C).Among total recorded axons in response to 200 Hz stimulation, 28% of axons in the Scn2a cKO displayed failures compared to only 18% in control mice (data not shown).Furthermore, violin plots revealed a larger proportion of axons with over 80% failure at higher frequencies (300-500 Hz) in Scn2a cKO mice while no axons in control mice showed failure rate above 80% in the same frequency range (Fig. 6C).Thus, myelin defects and disruption of Na v channel clustering caused by Scn2a loss detrimentally affect the reliability and temporal delity of spikes at the nerve terminal.
To discern whether AP failures at the nerve terminal are caused by conduction failures throughout the distal axon or failure to evoke AP, we tested if increasing the stimulus intensity could recover AP failures.First, we determined the minimum stimulating intensity (threshold intensity, TI) necessary to trigger a single AP for each axon.We found that threshold intensity was signi cantly higher in Scn2a cKO than in control (0.55 ± 0.09 V, n = 11 from 3 control mice vs 1.43 ± 0.30 V, n = 9 from 4 cKO mice, p = 0.0048, Mann-Whitney U-test, Fig. 6D).Next, we quanti ed failure rate in AP train (200 Hz stimulation) with incremental intensities (Fig. 6E).The calyx terminal from Scn2a cKO showed signi cantly higher failure rates at the TI (10 ± 6.49%, n = 11 from 3 control mice vs 47.78 ± 7.95%, n = 9 from 4 cKO mice, p = 0.0117, Two-way repeated measures ANOVA with Sidak correction, Fig. 6F) which was recovered by increased stimulating intensity.Interestingly, all axons in control with failures at TI were able to recover by TI + 20% intensity, whereas such intensity could only recover ~ 56% of axons in Scn2a cKO.Most of the axons with failures in Scn2a cKO required at least TI + 40% to recover (p = 0.0012, Two-way repeated measures ANOVA, Fig. 6F).These ndings were similarly observed in another OL-speci c cKO mice (Sox10 CreER ; Scn2a / , Supplementary Fig. 5).Taken together with EM and immunostaining analysis, these results suggest that a shortened distal internode and thinner myelin critically impair axon conduction throughout the distal axon and cause AP failures at the nerve terminal.
Alterations in structural and functional properties of the nerve terminal impact synaptic transmission and short-term plasticity.
Alterations in myelination and axonal properties at the nerve terminal critically impact synaptic plasticity 4,41 .To investigate the effects of myelin plasticity changes on synaptic activities at distal axons during the critical period, we evaluated synaptic transmission and short-term plasticity at the calyx of Held-MNTB synapse using whole-cell patch-clamp recordings.In the analysis of miniature excitatory postsynaptic currents (mEPSCs), we found a signi cant reduction in the frequency of mEPSCs in Scn2a cKO mice (p = 0.0004 Mann-Whitney U-test) without changes in amplitude (p = 0.1989, Mann-Whitney Utest, Fig. 7A).The reduction in mEPSC frequency might be associated with changes in presynaptic properties.Furthermore, we recorded evoked EPSCs triggered by afferent ber stimulation at the midline.There was no signi cant difference in the amplitude of a single EPSC between groups (6.56 ± 0.340 nA, n = 25 cells from 9 control, 6.45 ± 0.612 nA, n = 14 cells from 8 cKO mice, p = 0.9174, unpaired t-test, Fig. 7A).However, the paired-pulse ratio was signi cantly reduced in Scn2a cKO mice (0.852 ± 0.019, n = 25 vs 0.678 ± 0.060, n = 14, p = 0.0157, Mann-Whitney U-test, Fig. 7B), indicating an alteration in release probability (Pr).To further investigate this, we analyzed EPSC trains induced by 100 Hz ber stimulation (50 stimuli, Fig. 7C), employing three established analysis methods for the calyx synapse 42 .Although trends were observed in Pr and RRP size, statistical signi cance varied (Supplementary Fig. 6).Across the train method analyses 43 , the vesicle number of RRPs was signi cantly reduced in Scn2a cKO mice (p = 0.0241, Mann-Whitney U-test), and Pr tended to be higher, albeit not signi cantly (p = 0.0737, Fig. 7D).Additionally, the replenishment rate was signi cantly lower in Scn2a cKO mice (p = 0.0107, Mann-Whitney U-test, Fig. 7E).These changes in the number of vesicles around the active zone, including the size of the readily releasable pool, and reduced replenishment could lead to enhanced short-term depression.Taken together, OL dysfunction led to myelin alterations at the distal axon, which in turn impacted synaptic plasticity at local synapses within the auditory brainstem circuitry.

Discussion
Understanding the fundamental mechanisms of neuronal activity and their interplay with genetic factors provide invaluable insights for neurodevelopmental disorders, including ASD.We investigated the role of OL-Scn2a in the developing auditory brain from the genetic pro ling to the systemic analysis.Scn2a cKO mice showed hypersensitivity to sound stimulation.Single-nucleus RNA sequencing revealed that the loss of OL-Scn2a impacts myelin-related gene expression in the auditory brainstem.Structural analysis of myelinated axons demonstrated thinner myelin and axonal alterations in Scn2a cKO mice.Electrophysiology and immunostaining showed that alterations in myelinated segments and nodal structure along the distal axon in uenced the excitability of the nerve terminal, causing a reduced delity and reliability of spikes in Scn2a cKO mice.Additionally, those changes in the interaction between immature OLs and nerve terminal impact synaptic functions.Our ndings highlight the temporal progression of myelination in early development and its potential links to the onset or severity of ASD symptoms, offering potential aids for diagnosis, prognosis, and therapeutic interventions.
Abnormal development in white matter and myelination has been found in ASD animal models and humans with autism [44][45][46][47] .However, how myelin changes are associated with compromised neural circuit function in ASD remains unclear.Disruptions to myelin can in uence axonal integrity, altering ion channel distribution, and consequently affect neuronal excitability 5,19 .Speci cally, the spacing and periodicity of myelin segments determine Na v channel distribution at nodes and heminodes, enabling saltatory conduction and precise neural signaling 5,40 .In MBP deleted rats, hypomyelination led to disorganized Na v channel clustering and shorter internode, which are associated with a delayed AP onset, a longer AP half width, and AP failures at the nerve terminal 5,48 .In the cortical gray matter neuronal circuitry, cuprizone-induced demyelination caused hyperexcitability in pyramidal neurons by altering the axon initial segment (AIS) position and reducing the e cacy of AP generation 19 .The current study expands the dimensions of what have been previously reported by examining the effects of OL-Scn2a loss on myelin and axonal integrity speci cally at distal axons near nerve terminals.We demonstrate that OL-Scn2a loss leads to alterations in ion channel redistribution and aberrant excitability at nerve terminals such as conduction failure and spontaneous spiking.This is particularly pertinent as Scn2a-expressing immature OLs are abundant in synapse-rich regions like the MNTB.Interestingly, we observed that OL-Scn2a loss had more pronounced effects at nerve terminals than on nodes, which differs from prior models like MBP-deleted or cuprizone-induced demyelination.In these models, mature OLs were either unable to myelinate properly or underwent cell death 5,19 .In contrast, Scn2a cKO mice may have impaired interactions between immature OLs and distal axons, signi cantly affecting nerve terminal function including neurotransmitter release and short-term synaptic plasticity during postnatal development.
Therefore, our study underscores the region-speci c relationship between myelin integrity and ion channel distribution in the developing brain.We emphasize that any disturbances in myelin structure can trigger cascading effects on neuronal excitability and synaptic function in the CNS, especially at nerve terminals in the auditory nervous system.
While OL dysfunctions are highly associated with the pathophysiological process of ASD, few studies have focused on ion channels of OLs.Ion channels, including voltage-gated calcium channels (Ca v ), Na v , K v , and inward-recti er potassium channels (K ir ), are expressed in OLs [49][50][51] .While the functions of Ca v and K ir channels in OLs are relatively well-de ned, the roles of Na v remain unclear.Our snRNA-seq data demonstrated a decline in the mature OL population coupled with an increase in OPCs and differentiating OLs in Scn2a cKO mice.In addition, OL-Scn2a deletion was found to down-regulate myelin-related genes in mature OLs.How are Na v 1.2 channels, encoded by Scn2a, involved in OL maturation and myelination?One possible explanation is that the activation of Na v 1.2 may be pivotal for triggering Ca v channel activation, leading to a Ca 2+ ux within OLs, which is involved in OL proliferation, migration, and differentiation 50 .Speci cally, Ca 2+ signaling facilitated by R-type Ca v in myelin sheaths at paranodal regions, might in uence the growth of myelin sheaths 49,52 .To activate high-voltage activated calcium channels such as L-and R-Type e ciently, the activation of Na v 1.2 channels should be required for depolarizing OL membrane to around − 30 mV 53 .Consequently, the synergic interplay between Na v 1.2 and Ca v channels could amplify calcium signaling in OLs, initiating the differentiation and maturation processes 50 .Another possibility is that Na v 1.2-mediated spiking in immature OLs could facilitate the release of neurotrophic factors such as BDNF, which impacts myelination via autocrine signaling.OLs are signi cant providers of BDNF and express the BDNF receptor TrkB 54 .Therefore, Na v 1.2-mediated excitability of immature OLs may enhance myelination through BDNF-TrkB autocrine signaling in response to neuronal activity.
One prevalent characteristic of ASD is sensory processing disorders, notably auditory hypersensitivity.
For example, Fmrp1 KO mice showed hypersensitivity in hearing perception attributed to aberrant activity in the auditory cortex 55,56 .Similarly, Shank3 KO mice displayed auditory hypersensitivity, as evidenced by ampli ed ABR responses and a heightened startle re ex in the ASR 57 .Here, Scn2a cKO mice displayed augmented ABR amplitudes and a stronger startle re ex, despite no peripheral changes.
Notably, these mutant mice showed myelin defects in the CNS 58,59 .This raises the question: How are alterations in myelin and neuronal properties associated with auditory processing disorder?Defects in myelination within auditory pathways can lead to synchronization issues among neurons responsible for sound processing.Increased asynchronized signals can compromise the precise timing required for sound localization and the discernment of complex auditory patterns.In Scn2a cKO mice, a loss of temporal delity at presynaptic terminals and inconsistent spike conduction along myelinated axons was observed.Notably, ~ 20% of presynaptic recordings exhibited aberrant and asynchronous spikes at the nerve terminal in Scn2a cKO.This atypical ring of auditory neurons can contribute to auditory hypersensitivity.Neuronal ring rates were abnormally increased following demyelination, resulting in hyperexcitability of the neural circuitry 19,60 .This hyperexcitability may be caused by alterations in sodium channel distribution and a rise in extracellular potassium along the myelinated axon 61 .Another possibility is that hypersensitivity may also emerge from changes in the intricate balance between excitatory and inhibitory regulation within the auditory brainstem.A decline in inhibitory neurons or disruptions in inhibitory inputs can induce hypersensitivity 62,63 .In Shank3 KO mice, hypersensitivity might be caused by diminished inhibitory regulation in the auditory circuit 57 .In addition, our snRNA-seq data suggested a downward trend in genes associated with neuronal migration in interneuron populations from Scn2a cKO mice (data not shown).Interestingly, the MNTB serves as a major inhibitory source, releasing GABA and glycine to various nuclei within the superior olivary complex.Consequently, abnormalities in temporal delity and altered neurotransmission in the MNTB may disrupt the excitatory and inhibitory balance in this subcortical circuitry, potentially leading to auditory hypersensitivity.
In conclusion, the integrity of the myelin sheath plays a pivotal role in regulating neuronal excitability and ensuring proper auditory function.Defects in myelination can create a spectrum of auditory dysfunctions, including hypersensitivity.Our results demonstrated how OL-Scn2a is involved in the relationship between myelin defects, neuronal excitability, and auditory pathology in ASD, potentially paving the way for targeted therapeutic interventions.

Animals
All procedures were approved in advance by the Institutional Animal Care and Use Committee of University of Michigan.Conventional Scn2a mutant mice (Scn2a +/+ and Scn2a +/-mice) 29 and conditional Scn2a knockout mice 27 were used.To create OL-speci c Scn2a knockout mice, we crossed a Scn2a / mice with two OL speci c Cre-recombinase expressing mouse lines, Pdgfra CreERT mice and Sox10 CreERT , generating double transgenic mice (Pdgfra CreERT ; Scn2a / and Sox10 CreERT ; Scn2a / ) as described in a previous study 27 .Littermates without a Cre-recombinase but with Scn2a / were used as control.70 mg/kg of tamoxifen was administered via i.p. injection at postnatal days (P) 4, 6, and 8.Both male and female mice aged P15-P21 were used for immunohistochemistry, EM, and ex vivo electrophysiology.Mice of both sexes aged P21-P27 were used for snRNA-seq, ABR, ASR, and DPOAE.
Single nuclei RNA sequencing (snRNA-seq) A snRNA-seq was conducted using brainstem tissues from control (Scn2a / ) and Scn2a cKO mice (Pdgfra CreERT ; Scn2a / , P22).The brainstem was dissected and mechanically homogenized using dounce homogenizer in homogenizing buffer (250 mM sucrose, 25 mM KCl, 5 mM MgCl 2 , 10 mM Tris, 1 uM DTT and 0.1% Triton-X100) supplemented with enzymatic RNase Inhibitor (400 U/ml).700 µl of homogenization solution was added.Homogenization involved 5 strokes of a loose pestle and 10-15 strokes of a tight one.The resultant solution was completed to 1 ml and passed through a 40µm strainer.Post centrifugation, nuclei were rinsed with PBS containing RNAse inhibitor.A subsequent ltration through a 20 µm strainer was done before resuspending in 1 ml of PBS forti ed with RNAse inhibitor.For xation, 3 ml of 1.33% PFA was added to the nuclei for 10 minutes, followed by permeabilization using 160 µL of 5% Triton X-100 for 3 minutes.After washing out the PFA, nuclei were suspended in PBS with RNAse inhibitor and quanti ed manually via a hemocytometer.The xed nuclei were barcoded and prepared for library using Evercode WT Mini kit (Parse Biosciences) following manufacturer's guidelines.The nal library underwent sequencing on Novaseq 500 at Novegen (Sacramento, CA).
The raw reads were mapped and quanti ed using split-pipe v1.0.3 (Parse Biosciences).The count data was analyzed using Seurat 64 in R. To minimize arti cial error, the cells with more than 1% mitochondria DNA and 5000 transcripts were considered dead cells and doublet cell, and those were excluded from further analysis.All nuclei were clustered using shared nearest neighbor (SNN) and plotted using Seurat in R. Each cluster was manually identi ed by examining expression of cell markers (Pdgfra/Cspg4 for OPC, Mog/Mbp/Plp1 for mature OL, Aldh1l1 for astrocyte, Itgam/Csf1r for microglia, Pecam1 for vascular cells, Rbfox3 for neuronal population).To examine oligodendrocyte differentiation, the oligodendrocyte lineage clusters were isolated from the data set and trajectory analysis was conducted using Monocle3 in R 65 .The population of each differentiation stage of oligodendrocytes was calculated from [number of nuclei from each cluster/total number of nuclei from all of OL lineage].To examine gene expression pattern in each cluster, the differentially expressed genes between control and cKO were obtained with DESeq2 66 method with Bonferroni multiple comparison correction in Seurat.For further analysis, we isolated matured OL cluster, where showed signi cant gene expression difference between groups, and examined expression level of myelin related genes.To list myelin related genes, Amigo2 67 was used.391 genes were searched on Amigo2 using the keyword, "myelin".The expression level of the genes was compared between groups using DESeq2 methods without multiple correction.The DEGs from the analysis were used gene network analysis using String 68 .The network was manually categorized based on their function.All of data visualization was done using ggplot2 69 and String.
Recordings were not corrected for the predicted liquid junction potential of 11 mV.Patch electrodes had resistances of 4-5 MΩ.Current-clamp recordings were continued only if the initial uncompensated series resistance was < 20 MΩ 4,5 .Lucifer Yellow (1 mM, Invitrogen) was added to the pipette solution to visualize the calyx of Held terminal.Presynaptic action potentials (APs) from the calyx of Held terminal were evoked by stimulation with a bipolar platinum-iridium electrode (Frederick Haer, Bowdoinham, ME) placed near the midline spanning the afferent ber tract of the MNTB.An Iso-Flex stimulator driven by a Master 10 pulse generator (A.M.P.I., Jerusalem, Israel) delivered 100-µs pulses at 1.2 times threshold (< 15 V constant voltage).Signals were ltered at 2.9 kHz and acquired at a sampling rate of 10-50 µs.AP waveform parameters were analyzed from the rst AP induced by minimum current injection (rheobase current) and the subsequent AP phase plot, where membrane potential slop (dV/dt) is plotted against the membrane potential 70 .In both and Scn2a cKO mice, all cells displayed a single in ection in the rising phase of the AP, indicating APs were generated at the heminode adjacent to the presynaptic terminal.Presynaptic AP trains were obtained by averaging three sweeps ( ve for a single AP) in each experiment.Data were analyzed o ine and presented using Igor Pro (Wavemetrics, Lake Oswego, OR).

Auditory brainstem responses (ABRs)
To record ABR, the mice (P21-P27) were anesthetized with 3.5% iso urane and maintained with 2.3% iso urane during recording (1 l/min O 2 ow rate).ABR recordings were performed in a sound attenuation chamber (Med Associates, Albans, VT).Subdermal needle electrodes (Rochester Electro-Medical, Lutz, FL) were placed on the top of the head, ipsilateral mastoid, and contralateral mastoid as the active, reference, and ground electrode, respectively.The signal differences in the ABRs between the vertex and the mastoid electrodes were ampli ed and ltered (100-5000 Hz).Acoustic stimuli were generated by an Auditory Evoked Potentials Workstation (Tucker-Davis Technologies [TDT], Alachua, FL).Closed eld click stimuli were presented to the left ear.The signals consisted of a series of amplitude-modulated square waves (0.1 ms duration, 16/s) through TDT Multi-Field Magnetic Speakers.The sound stimuli were delivered through a 10-cm plastic tube (Tygon; 3.2-mm outer diameter) at a repeat rate of 16/s.Sound intensities ranged from 90 to 20 dB, with 5-dB decrements, and responses to 512 sweeps were averaged.
Distortion Product Otoacoustic Emissions (DPOAEs) Mice were anesthetized with 3.5% iso urane and maintained 2.3% urane recording.DPOAE recordings were performed in a sound attenuation chamber (Med Associates, Albans, VT).Acoustic stimuli were generated by an Auditory Evoked Potentials Workstation (TDT, Alachua, FLO).The ER-10B + recording microphone (Etymotic Research, Elk Grove Village, IL) with ear-tip was inserted into the ear canal.The sound stimuli were delivered through two TDT Multi-Field Magnetic Speakers connected to the recording microphone by 10-cm coupling tubes (Tygon; 3.2 mm outer diameter).Pure tones were presented at 20% frequency separation between f1 and f2 at 8, 12, 16, and 32 kHz.Sound intensities ranged from 80 to 20 dB, with 10-dB decrements, and responses to 512 sweeps were averaged.Distortion products were calculated as 2f1-f2 minus the noise oor that were detected by the recording microphone and ampli ed by RZ6 processor (TDT).

Acoustic startle responses (ASRs)
Mice either sex between ages were put in a Plexiglas holding cylinder located in soundattenuated chamber using the SR-LAB startle response system (San Diego Instruments, San Diego, CA).Weights of mice were recorded to account for sex or size differences.Sound levels from the chamber speakers were calibrated with a digital sound level meter (Part Number MS-M80A, Mengshen).Each trial consisted of an initial acclimation period of 5 minutes of background level noise (at 65 dB) followed by 5 rounds of randomized noise stimuli playing for 40 milliseconds with 15 seconds of background between each noise stimuli at 65 dB (background), 95 dB, 105 dB, 115 dB, and 120 dB.Stimuli was produced by a digital signal processing-controlled system ampli ed and emitted by a loudspeaker.Startle responses were measured inside the sound-attenuated chamber by a movement-sensitive piezo-accelerometer platform.The maximum was used as the startle amplitude at each tone with millivolts (mV) as the unit of measurement and averaged per mouse.For each strength of stimulus, startle amplitudes were averaged across 10 trials.Startle responses to the three initial stimuli were excluded from statistical analyses.
Pre-pulse inhibition of acoustic startle response PPI of ASR was examined 2 days after the ASR assessment.The apparatus and basic experimental conditions were identical to that described above.Test sessions started with the acclimatization period which included three startling stimuli (120 dB) to accustom the mice to the experimental procedure.The initial stimuli were followed by 100 trials (10 × 10 trials) presented in a random order.The PPI session involved: 10 trials with a sham stimulus (65 dB, 40 ms), three types (3 × 10) of pre-pulse trials (PP) which included only 20 ms PP stimuli (69, 73, and 81 dB), 10 pulse trials (P) which included only a pulse (startling) stimulus (120 dB, 40 ms), three types (3 × 10) of pre-pulse-and-pulse trials (PP-P) which included a 20 ms PP (69, 73, and 81 dB) followed 100 ms later by a 120-dB P stimulus.Startle responses were measured for 100 ms after the onset of the last stimulus each trial.For type of trial, startle amplitudes were averaged across 10 trials.The magnitude of PPI was calculated as a percent inhibition of the startle amplitude in the P trial (treated as 100%) according to the formula: [startle amplitude in P trials -startle amplitude in PP-P trials)/startle amplitude in P trials] × 100%.Startle responses to the three initial stimuli were excluded from statistical analyses.

Immunohistochemistry
Mouse (200 µm) stained with tetramethylrhodamine dextran (Invitrogen) were incubated in normal aCSF bubbled with carbogen at 37 o C for 30 min.All slices were xed with 4% (w/v) paraformaldehyde in PBS for 10 min.Free-oating sections were blocked in 3% goat serum and 0.3% Triton X-100 in PBS for 30 min and incubated with the primary antibody overnight at room temperature.The following primary antibodies were used: anti-PanNa (mouse IgG1; 1:400; Sigma, Cat.#S8809), anti-Caspr (guinea pig IgG; 1:200; gifted from Dr. Manzoor Bhat, UTHSCSA), anti-PLP1 (mouse IgG2a, 1:500; invitrogen, Cat.#MA1-80034).Antibody labeling was visualized by incubation of appropriate Alexa uorconjugated secondary antibodies (1:500; Invitrogen) for 2 h at room temperature.Stained slices were viewed with laser lines at 488 nm, 568 nm, and 647 nm using a 40x/1.40 or 63×/1.40oil-immersion objective on a confocal laser-scanning microscope (LSM-710; Zeiss).Stack images were acquired at a digital size of 1024 × 1024 pixels with optical section separation (z-interval) of 0.5 µm and were later crop to the relevant part of the eld without changing the resolution.The confocal image stacks were analyzed using ImageJ software.

Transmission electron microscopy
The was removed and immersed in ice-cold low-calcium aCSF (mentioned previously in slice preparation) and the MNTB from the brainstem was dissected at 200 µm-thick section using a microtome (VT1200s, Leica), then xed with xative consisting of 4% formaldehyde and 1% glutaraldehyde and stored at 4 o C. Further processing was performed by the UTHSCSA Electron Microscopy Lab as previously described 41 .Axon bundle images were analyzed at a nal magni cation of 8000X, with the longest length of the axon measured as the inner diameter and the inner radius divided by the outer radius as the g-ratio.Three images of axon bundles per mouse were analyzed and averaged.

Statistical analysis
Immunostaining data were based on analyses from at least six cells in six slices from three to eleven animals.Experimental data were analyzed and presented using Igor Pro and Prism (GraphPad Software, San Diego, CA).For statistical signi cance, we tested the normality of the data distribution with the  olivary complex(III), inferior colliculus (IV), and medial geniculate nucleus (V) in the auditory brainstem circuitry.D. Summary of ABR thresholds of control (n=46) and Scn2a cKO mice (n=29) in response to click stimuli (from 30 dB to 90 dB SPL).There was no signi cant difference (p = 0.1842, unpaired t-test).E.ABR amplitudes (µV) of wave I to V responding to 80 dB SPL clicks.The amplitude of peak II (p = 0.0002) and peak III (p = 0.0191) were signi cantly larger in Scn2a cKO mice (n=26) compared to control (n=39, multiple comparison from two-way ANOVA with Bonferroni correction).F.The ratio for the amplitude of wave I over the amplitude of each wave (II, III, and IV), evaluating central gain changes.G. Summary of distortion products (dB SPL) at 80 dB at various pure tone stimuli (4, 8, 12, 16, and 32 kHz).All values plotted are means per mouse ± s.e.m., * p < 0.05, ** p < 0.01, *** p < 0.001.identi ed by trajectory analysis from OPCs to mature OLs (gray line, left), and the percentage (%) of each cluster (from OPC to Mature OL) in total OLs from individual controls (n=2) and cKO mice (n=2).C. UMAP plots of snRNA-seq for OL lineage cells from control (blue dots) and cKO (red dots, left).The number of DEGs for each OL lineage (right).D. Volcano plot of differentially expressed myelin related genes (391 genes) in control and cKO mice.E.33 genes showed signi cant differential expression (p < 0.05) and interconnection between those genes was detected in Sting analysis.Those genes were related to myelin structure (green line), cell junction (red), membrane protein signaling (blue), and ion channel (yellow).F. Violin plots of top eight genes being downregulated in Scn2a cKO mice; Mobp, Ncam1, Ptn, Gnao1, Mtmr2, Abca2, Arhgef10 and Mpdz.

Figures
Figures

Figure 2 Single
Figure 2

Figure 3 Loss
Figure 3