Serotonin, one of the most widely spread neurotransmitters in the central nervous system, has been known to play a critical role in brain morphogenesis and functions (Azmitia, 2001; Gaspar et al., 2004). The defects of serotonergic neurons are related to many psychiatric disorders, including depression, anxiety, and autism spectrum disorder (ASD) (Simpson et al., 2011). Acting as a key regulator of serotonergic activity, the serotonin transporter (SERT) is usually selected as the target of antidepressant treatments. SERT represents a potential mediator for anxiety- and depression-related behaviors. However, chronic exposure to selective serotonin reuptake inhibitors (SSRIs) was reported to elicit hearing disorders, such as tinnitus (Kehrle et al., 2015; Pattyn et al., 2016), auditory hallucinations (Kogoj, 2014), and hearing loss (Blazer and Tucci, 2018).

However, how the functions of SERT affect the auditory system remains unclear. Both increases and decreases in serotonin levels during early development have been found to impair the formation and function of the primary somatosensory cortex of rodents (Cases et al., 1996; Persico et al., 2000; Jennings et al., 2006). Pre- and post-natal exposure to the SSRI disturbed the chemoarchitecture of the mouse auditory cortex (AC) and resulted in ASD-like behavior (Simpson et al., 2011). Our previous study found that both SSRI treatment and SERT knockout (KO) did not change the auditory brain responses but abolished the auditory mismatch negativity in adult animals (Pan et al., 2020). These investigations suggest that SERT may affect the auditory functions by manipulating the serotonin level and serotonergic neurons of AC.

In rodents, 2 weeks after birth, layer IV afferent neurons in the primary visual, auditory, and somatosensory cortex are innervated by aggregates of serotonin-containing processes (D’Amato et al., 1987; Blue et al., 1991). At present, the morphological and especially the functional implications of SERT in AC remain unclear. In the present study, we examined the fine dendritic structure of the neurons in primary AC. We found that although expression of SERT is low in the AC, both pyramidal neurons and interneurons in SERT KO mice showed significant reductions in dendritic length, the number, and density of dendritic spines than did wild-type (WT) mice. The electrophysiological features of AC were also impaired by the SERT deficit. These results suggest that SERT plays a key role in development and functional maintenance of AC. This may explain the observed hearing disorders in patients utilizing drugs that target SERT, such as SSRIs, in the treatment for psychiatric disorders.

All the animal experiments involved in this study were approved by the Institutional Animal Care and Use Committee (IACUC) of Southern Medical University. The principles formulated by the Animal Care Committee of Southern Medical University were followed throughout the experiment.

By using immunostaining, the expression of SERT in the auditory neural system was examined systematically. As shown in Figure 1, no robust SERT expression was detected in the cochlea (spiral ganglion neurons; Figure 1A), auditory brainstem (cochlear nuclei, superior olivary complex, and inferior colliculus; Figures 1B–D), or auditory thalamus (medial geniculate body; Figure 1E). Hence, we examined the SERT expression in the primary AC. Our data from AC exhibit sporadic spread of SERT immunoreactive puncta, which represents the SERT immunoreactive fibers (Figure 1F). However, the SERT immunoreactive puncta was usually detected around the soma of cortical neurons (inset of Figure 1F) with a much lower density than the expression in the raphe nuclear complex (Figure 1G). This result suggests that AC neurons receive sparse serotonergic projections, if any.

To determine the roles of these serotonergic projections, the morphology of AC neurons from SERT KO mice was examined in comparison with that of WT mice. Our data show that for pyramidal cells in layer IV, the number of intersections and the total length of dendrites were significantly reduced in both apical and basal dendrites of the SERT KO AC neurons (Figure 2). Moreover, the reduction of dendrites was also found in layer II/III interneurons of SERT KO AC (Figures 2F,G).

Most excitatory synaptic connections occur on dendritic spines. Spines can individually detect the temporal coincidence of pre- and post-synaptic activities and thus serve as basic functional units of neuronal integration (Gray, 1959; Yuste and Denk, 1995; Spacek and Harris, 1998; Arellano et al., 2007). To further determine whether synaptic function at neuronal junctions was altered in SERT KO mice, we looked at the morphologic structure of the dendritic spines of AC neurons. We measured the number and density of total dendritic spines in layer IV pyramidal cells and layer II/III interneurons from the AC of WT and SERT KO mice (Figure 3A). In apical dendrites, the total number of dendritic spines of layer IV pyramidal cells was less in SERT KO mice [vs. WT, F(1,159) = 252.5, p < 0.0001, two-way ANOVA]. Significant reduction was observed at a distance of 25–125 μm from cell bodies (vs. WT, p < 0.001, two-tailed t-test; Figure 3B). The mean density of apical dendritic spines reduced significantly in SERT KO mice (vs. WT, p < 0.0001, two-tailed t-test; Figure 3E). The total number of basal dendritic spines of layer IV pyramidal cells was also reduced in SERT KO mice [vs. WT, F(1,356) = 109.6, p < 0.0001, two-way ANOVA] and significantly reduced at a distance of 25–150 μm from cell bodies (vs. WT, p < 0.001, two-tailed t-test; Figure 3C). Meanwhile, the mean density of basal dendritic spines reduced significantly in SERT KO mice (vs. WT, p < 0.0001, two-tailed t-test; Figure 3F). However, the total number of dendritic spines in layer II/III interneurons in SERT KO mice was not significantly different [vs. WT, F(1,292) = 10.93, p = 0.011, two-way ANOVA]. The density was only found reduced at a distance of 25–75 μm from cell bodies (vs. WT, p < 0.05, two-tailed t-tests; Figure 3D). The mean density of dendritic spines was also significantly reduced in SERT KO mice (vs. WT, p < 0.0001, two-tailed t-test; Figure 3G). Together with the results of our Sholl’s analysis, these morphological data implied that synaptic transmission between the subcortical and cortical neurons, as well as the neurons within the AC, might be weakened in the auditory system of SERT KO mice. These defects may affect the functions of SERT KO AC neurons.

Electrophysiological experiments were conducted to investigate the functional alteration of auditory neurons in primary AC (AI) of SERT KO mice. Previous studies have found that neurons at different locations within the same layer of AI respond to different frequencies (Kelly and Sally, 1988; Kaas et al., 1999; Zhang et al., 2001; de Villers-Sidani et al., 2007). Therefore, we measured the “frequency-tonotopic map” of neurons in the AI region of cortex. A total 246 sites and 243 sites were recorded from the auditory cortical area of WT and SERT KO mice, respectively. Normally in WT mice, high-frequency sensitive neurons are located in the rostral sites of the AI region, and low-frequency sensitive neurons are located in caudal side (Stiebler et al., 1997).

As shown in the representative data in Figure 4A, the BFs of AC neurons were distributed regularly from the rostral to caudal sites, forming a compact and ordered “tonotopic map” in WT mice. However, for AC neurons of SERT KO mice, their frequency selectivity did not show a systematic organization, although the neurons that respond to high frequencies were still mostly located to the rostral side. We further analyzed the frequency range and the total area of the AI region. No statistical difference was found between SERT KO and WT mice (p > 0.05, two-tailed t-test; Figures 4B,C). The BFs of all recorded sites were plotted against a normalized AC axis (Figure 4D). The distribution of BFs was quantified with the TI, which was significantly increased in SERT KO mice (p < 0.01, two-tailed t-test). These results showed that a disrupted tonotopic map was found in the AC of SERT KO mice, suggesting that SERT may mainly affect the neuron’s tuning property.

We then measured the BF and the MT of AC neurons recorded from SERT KO and WT mice, by which the frequency selectivity and sound sensitivity of neurons were compared. The BFs and MTs of SERT KO and WT neurons were pooled in Figures 5A,B, respectively. Interestingly, although no change was found in BFs, the MTs of SERT KO neurons were significantly elevated. The MTs of WT neurons were generally below 40 dB SPL with an average of 27.09 ± 12.81 dB SPL. However, the MTs of SERT KO neurons were significantly higher with an average of 53.85 ± 11.33 dB SPL (p < 0.0001, two-tailed t-test; Figure 5C). These data suggest that individual neurons in SERT KO mice were significantly less sensitive to sound stimulation.

Auditory neurons respond not only to their most sensitive frequency (i.e., BF) but also to other frequencies of sound. One of the most important functions of auditory neurons is that they can selectively respond to sound within a range of frequencies. In WT mice, individual neurons respond to sound frequencies in addition to the BF at sound levels above MT. These frequency-threshold intensity points formed a “V-shaped” frequency-tuning curve (Figure 6A). In SERT KO mice, the frequency-tuning curves of cortical neurons have much wider bandwidth (Figure 6B). To evaluate the frequency selectivity of a single neuron, we used “Q30” value to analyze the changes in frequency selectivity of AC neurons. The higher the value of Q30, the sharper the frequency-tuning curve of the neuron, and the better the frequency selectivity. For AC neurons of WT mice, their Q30 values mostly ranged from 1 to 2.5, with an average of 1.78 ± 0.91 (Figures 6C,D). But in SERT KO mice, Q30 values of most AC neurons were below 1 (0.65 ± 0.27, mean ± SD; Figures 6C,D). This result suggested that AC neurons in SERT KO mice have less selectivity to sound frequency.

Our results of immunohistochemical study in the whole auditory system showed that SERT was rarely expressed in the auditory neurons, including those in the AC, of adult WT mice (Figure 1). These data suggest that the morphological and functional defects observed in AC neurons of SERT KO mice may not due to the disruption of serotonergic innervations directly but more likely related to the effects of SERT deficit on cortical circuit development.

As the staining results showed, the length of basal dendrites of pyramidal cells was significantly shortened (Figure 2), and the number and density of dendritic spines were significantly reduced in SERT KO mice (Figure 3), suggesting that the ability of pyramidal cells to receive input information was decreased, which might cause less sensitivity to sound response (Figure 5). Moreover, the reduction of the length, the number, and density of dendritic spines in inhibitory interneurons of SERT KO mice suggests that the reduced inhibitory of the interneurons might cause less frequency selectivity of neurons (Figure 6). However, deficient SERT may change the neural morphology in different ways in different brain regions. The length of dendrite and the density of dendritic spines were increased in the infralimbic cortex of SERT KO mice. Meanwhile, SERT KO has little effects on the morphology of basolateral amygdala neurons (Wellman et al., 2007). These different results suggest that the morphology of neurons was also specifically determined by their circuits; even the SERT was abolished in the early stage of the development. Interestingly, many evidences have suggested that SERT is expressed not only on serotonin neurons but also on other neurons such glutamatergic neurons during development. The loss of SERT in these neurons may affect the neural circuit in the cortex during development (Chen et al., 2015). This may explain why the morphology of AC neurons was altered by SERT KO, although few direct serotonergic projections were detected in AC of WT animals.

SERT is a target of SSRIs treatments and regulates the activity of serotonergic system (Kalueff et al., 2009). In the present study, we examined the role of SERT function in primary AC by using SERT KO mice. In WT mice, the primary AC formed a compact and ordered “tonotopic map”; the optimal response frequencies from the rostral to caudal side was regularly distributed. But in SERT KO mice, the primary AC formed a distorted tonotopic organization (Figure 4). However, in SERT KO mice, the disordered frequency topology did not show a systematic decrease in frequency, although the neurons that respond to high frequencies were still mostly medial to the rostral side, and the neurons that respond to low frequencies were still mostly lateral to the caudal. Next, we further analyzed the total area of the tonotopic map in the AI region and the frequency range of the response to sound (Figure 4B). There was no statistical difference between SERT KO mice and WT mice (Figure 4B). It suggests that SERT may mainly affect the orderliness of the neuron’s frequency response. Similarly, manipulations of rodent brain serotonin levels during early development, through either increases (produced in SERT or monoamine oxidase KO mice) or decreases (produced by parachlorophenylalanine or other treatments), alter the formation of the whisker (barrel) representation in the primary somatosensory cortex and promote aggressive and/or anxiety-related behaviors (Cases et al., 1995, 1996; Persico et al., 2000; Persico et al., 2001; Holmes et al., 2003; Jennings et al., 2006). In our previous study, auditory mismatch negativity response was found abolished by SERT KO, indicating the low ability in frequency discrimination (Pan et al., 2020). This finding was supported by the functional changes of AC observed in the present study. Under the administration of SSRIs, the animal showed anxiety-related and ASD-like behavior (Simpson et al., 2011; Pan et al., 2020). Although the abnormal behavior and auditory function were quite consistent in SERT KO and SSRI-administered animals, the direct evidence is still very limited for us to understand the effects of SSRIs on development and function of cortical circuits. However, the SERT KO mouse model used in the present study is constitutive and global. To investigate the role of SERT and SSRI in the development of ASD and anxiety, a knock-in SERT mouse model, such as SERT Ala56, should be expected in future studies (Siemann et al., 2017).

One of previous findings related to serotonin and early cortical organization found that the primary sensory cortex of mice, especially the layer IV neurons of visual, auditory, and somatosensory cortex, is transiently innervated with dense 5-HT-containing projection (D’Amato et al., 1987; Blue et al., 1991; Bennett-Clarke et al., 1994). Moreover, in early brain development, 5-HT is taken up into glutamatergic thalamocortical terminals (Bennett-Clarke et al., 1996; Lebrand, 1996) and used in combination with the 5-HT 1B receptor on layer IV afferents (Salichon et al., 2001). Serotonin plays a trophic role in brain morphogenesis, including cell proliferation, migration, and differentiation. It is also one of the first neurotransmitters to appear in the central nervous system (Azmitia, 2001; Gaspar et al., 2004). We found that the typical dendritic morphology was shortened (Figure 2) and that the density of dendritic spines was lower in SERT KO mice compared with WT mice (Figure 3). In SERT KO mice, the deficit of SERT may alter neural morphology and distort tonotopic organization in primary AC.

The structure and arborization of dendrites have a profound impact on the processing of neuronal information because they determine the extent of a neuron’s synaptic field. Thousands of spines stud the pyramidal cell’s apical and basal dendritic branches, increasing the neuron’s receptive surface area and allowing for integration of thousands of excitatory signals to influence the output (Spruston, 2008). Obviously, dendritic spines receive most of the excitatory impulses of a pyramidal cell, consistent with the information processing capacity of the neuron. A cell with spiny processes in homologous nuclei has more spines, the higher level of the subject in the animal series. Thus, as an example in vertebrates, the Purkinje cell of birds shows fewer spines than that of mammals.

Since neural circuits are defined by inter-neuronal communications, output precision in individual cells becomes essential to network function. Cortical interneurons use rhythmic inhibition to create narrow windows for effective excitation, entraining excitatory pyramidal cells to fire certain oscillatory patterns (Beierlein et al., 2000; Uhlhaas and Singer, 2010). Moreover, optogenetic manipulation of specific excitatory and inhibitory circuits directly caused changes in social and cognitive behaviors in mice (Yizhar et al., 2011). Thus, circuits tend to be malleable but vulnerable during post-natal development. Notably, several neurodevelopmental disorders, such as ASD and schizophrenia (SCZ), manifest during this plasticity period (Zoghbi, 2003; Hossein and Folsom, 2009).

Interestingly, in our results, the neuron functions altered; especially, the MT for a BF of SERT KO mice was significantly higher than 40 dB (Figure 5). It suggested that absent SERT reduced the neurons’ intensity sensitivity to response. Moreover, in SERT KO mice, Q30 values were mostly below 1 (Figure 6). Meanwhile, the frequency-tuning curve was dull, and the frequency selectivity of neurons was significantly worse (Figure 6). It suggested that neurons in SERT KO mice showed less frequency selectivity to response.

All of these results may be related to shortened typical dendritic morphology of neurons (Figure 2) and the lower density of dendritic spines of primary AC in SERT KO mice (Figure 3), which may result in weakened synaptic information transmission between the cortical pyramidal cells and the interneurons of AC in SERT KO mice. The altered function of the primary AC suggests that SERT plays a critical role in circuit development and function.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors.

The animal study was reviewed and approved by Institutional Animal Care and Use Committee (IACUC) of Southern Medical University.

HZ and JT designed the experiments. WP, JP, and YZ performed the experiments. All authors wrote, reviewed, edited, and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.