Auditory brainstem development and plasticity

Abstract

During development and adulthood, the normal activity of the auditory nerve plays a critical role in the maintenance of both fundamental structural, molecular, and functional parameters of auditory nerve synapses, and the postsynaptic excitatory or inhibitory neurons within the cochlear nucleus (CN). In addition, normal activity within the synaptic circuits of the CN is key to developing and maintaining appropriate synapse connectivity as well as the initiation of binaural sound processing in the superior olivary complex (SOC). Development plays a critical role in the proper neuronal connectivity and establishes a topographic map along the entire auditory pathway. Furthermore, evidence shows that neurons and synaptic circuits in the auditory brainstem are not hard-wired, but instead are plastic in response to hearing deficits. Whether this plasticity in response to hearing loss is compensatory or pathological is still unknown.

Developmental plasticity in the auditory brainstem

Developmental plasticity is important for normal patterns of neuronal connectivity and function. Aberrant auditory brainstem synaptic circuit development leads to impaired neuronal plasticity and hyperexcitability in the auditory cortex (Rais et al., 2018). A key aspect of normal auditory function is the topographic representation of high and low frequency sounds along the entire auditory pathway. This topographic representation starts with the specific connections of myelinated type I spiral ganglion neurons (SGN) innervating cochlear hair cells. Numerous elegant studies have shed light on the developmental molecular mechanisms of SGN topographic connectivity within the cochlea (a topic beyond the focus of this review; see Jahan et al., 2013; Yu and Goodrich, 2014). On the other hand, the mechanisms of topographic connectivity and bifurcation of the central auditory nerve axon projections in the CN are less understood. Studies have shown that synchronous burst of spontaneous activity generated by supporting cells in the cochlea before animals actually start hearing is a critical determinant of the development of normal connectivity in the lower auditory brainstem, particularly in the lateral superior olive (LSO; **Clause et al., 2014; Tritsch et al. 2010). Remarkably, functional studies have shown that this topographic organization exists along the entire central auditory pathway, even in animals in which hair cells never function (e.g. vesicular glutamate transporters-3 (VGLUT-3) knockout mice) (Cao et al, 2008; **Babola et al., 2018). Before hearing onset, and in addition to the spontaneous activity bursts, other types of synaptic plasticity appear to sculpt and refine synapses in the medial superior olive (MSO) of the gerbil SOC. Inhibitory neurons of the MSO exhibit inhibitory long-term potentiation (iLTP) (**Winters and Goldin, 2018). This iLTP reinforces inhibitory inputs coactive with binaural excitation, presumably by weakening distal synapses, thereby establishing the mature soma-biased pattern of inhibition. However, despite these studies, it is still unclear whether the synchronous burst of spontaneous activity in the cochlea, or other novel types of synaptic plasticity mechanisms, play a key role in the branching and the topographic connections of auditory nerve axons within the CN. One study in the CN of mice indicates that the normal branching and the gross topographic organization develop even in the absence of activity before and after the onset of hearing (Wright et al., 2014), but it is likely that activity-dependent mechanisms make these connections more precise with age. Other factors in addition to synaptic activity likely play a role in the proper branching of auditory nerve fibers within the three subdivisions of the CN. Extracellular cues and signaling mechanisms such as bone morphogenic proteins (BMPs) could be involved in the correct axon branch formation and elimination of the auditory nerve within the CN (Xiao et al., 2013; *Kronander et al., 2019). In addition, other studies have shown that neurotrophin and neurotrophin receptor deletion alter normal auditory nerve fibers development and topographic organization within the CN (Fritzsch et al., 2016; Lu et al., 2014; Yang et al., 2017). Recent evidence shows that a conditional deletion of the basic helix-loop-helix (bHLH) gene Neurod1 in the ear leads to a shortened and nearly overlapping cochleotopic projection from SGN to the CN (*Macova et al., 2019). While it leads to disorganized auditory nerve branching, Neurod1 conditional deletion does not seem to affect how auditory nerves synapse on bushy cells of the anteroventral CN, measured by VGLUT-1 immunolabeling. This study shows that Neurod1 conditional deletion from the ear alters the tonotopic maps in the inferior colliculus, affecting the frequency, intensity, and temporal processing of the central auditory system at the physiological and behavioral levels (*Macova et al., 2019). Thus, physiology and molecular specification are two important aspects that fine tune the development properties of the auditory system. However, interactions between these functional and molecular mechanisms remains unclear.

A switch in neurotransmitter receptor subunits at the synapse in response to hearing onset could be another key mechanism contributing to proper synaptic refinement necessary for normal function during development. Studies have shown that hearing onset triggers molecular remodeling at excitatory and inhibitory synapses in the lower auditory brainstem (Lujan et al., 2019; Noh et al., 2010; Pilati et al., 2016). Highly specialized glutamatergic synapses such as the calyx of Held in the medial nucleus of the trapezoid body (MNTB) of the SOC mature rapidly to achieve temporal and high-fidelity transmission of sound localization information (Futai et al., 2001; Yang et al., 2011). Around hearing onset, a developmental switch of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) glutamate receptor subunit composition occurs, shifting from slow-gating GluA1–dominant to fast-gating GluA4-dominant occurs at both the calyx of Held synapses (Koike-Tani et al., 2005; Yang et al., 2011) and at the LSO glutamatergic synapses (Pilati et al., 2016). While the mechanisms driving this switch are still largely unknown, patterned stimuli mimicking spontaneous burst of synaptic activity in early postnatal days can initiate this gating switch at calyx synapses via a novel form of plasticity trough N-methyl D-aspartate (NMDA) and metabotropic glutamate receptors (*Lesperance et al., 2020). Further functional studies, as well as anatomical-molecular studies, are needed to determine whether this novel plasticity and the switch in receptor subunit expression exists at other highly specialized synapses such as the auditory nerve synapse in the CN.

In recent years, evidence shows that in addition to neurons, glial cells play an important role in the formation of neural circuits. In the CN and MNTB, astrocytes and oligodendrocytes are present as early as postnatal day 0 (P0) in rats and mice, while microglial cells only appear after P6 (Dihn et al., 2014; Saliu et al., 2014). Remarkably, glial progenitor cells (astrocytes and $\mathrm{\text{N}}{\mathrm{\text{G}}}_2{}^{+}$ oligodendrocyte precursors: OPC) begin to divide before birth, and rapidly increase their rate of division by hearing onset, while later declining (Saliu et al., 2014). In the MNTB, the calyx of Held develops remarkably rapidly (Holcomb et al., 2013); during the first four postnatal days in rats, MNTB neurons go from being multi-innervated to mono-innervation. The developmental time of the calyx of Held coincides with glial cells proliferation, and reports suggest that glial cells coordinate these processes. In the MNTB, glial progenitor cells could play a key role via at least four mechanisms: 1) direct physical role in which processes of progenitor glial cells surrounding the cell body of MNTB neurons compete with developing calices (Holcomb et al., 2013); 2) by taking up excess glutamate via glia glutamate transporters, preventing saturation of glutamate receptors in immature calices (Renden et al., 2005; Uwechue et al., 2012); 3) through chemical signaling, e.g. releasing signaling factors such as glutamate and D-serine that activate glutamate receptors on principal neurons (Reyes-Haro et al., 2010), or BDNF, which regulates presynaptic neurotransmitter release from calyx of Held terminals (Jang et al., 2019); and 4) by direct synaptic connections mediated by AMPA receptors from immature calices on NG2⁺ glial cells (Müller et al., 2009). Unlike in the MNTB, how astrocytes and OPC participate in synapse formation and maintenance in the CN, LSO or MSO is still unknown. The fidelity and reliability of neurotransmission of auditory nerve and calyx of Held synapses require compact axonal myelination (Sinclair et al., 2017, *Stange-Marten et al., 2017). In the MNTB, the function and maturation of oligodendrocytes to promote myelination require Nav1.2-driven spiking of immature oligodendocytes. It is suggested that AMPA receptor-dependent glutamate-mediated currents could depolarize excitable immature oligodendrocytes enough to fire spikes in the MNTB (**Beret et al., 2017). Removal of Nav1.2 altered the morphology of immature oligodendrocytes and lead to a decrease in the expression of myelin binding proteins (MBP). Therefore, oligodendroglial Nav1.2 function has been primarily linked to oligodendrocytes maturation and myelination in the MNTB (**Beret et al., 2017). Thus, it is attractive to speculate that the excitability of oligodendrocytes could contribute to the temporal fidelity of auditory signals in the lower auditory brainstem. In summary, correct patterns of neuronal connectivity, and therefore synaptic function during development, appear to depend on supporting cells in the cochlea, as well as glial cells within the auditory brainstem (progenitor glial cells). Further studies are needed to shed light on the mechanisms of neuron-glia communication in central sound processing during development, adulthood, and aging, as well as in response to hearing loss.

Plasticity in mature auditory brainstem synaptic circuits

Mature neurons and synaptic circuits in the lower auditory brainstem are not hard-wired but are instead plastic in response to hearing deficits. In vitro electrophysiological studies report that bilateral ear plugging and broadband noise exposure can spur a novel type of presynaptic plasticity at the auditory nerve synapse (endbulb of Held) on CN (**Ngodup et al., 2015; **Zhuang et al., 2017). In particular, these studies show that sound-driving activity regulates the probability of neurotransmitter release (Pr) as well as the number of release sites (N) at the endbulb synapse. In contrast, none of those studies reported a change in the mEPSC amplitude or decay, which could suggest compensatory changes of AMPA glutamate receptors at the PSD. On the other hand, structural studies in response to monaural conductive hearing loss have shown that endbulb synapses had larger postsynaptic densities as well as a remodeling of postsynaptic AMPA receptors, suggesting that a postsynaptic homeostatic mechanism leads to plasticity in response to sound reduction (**Clarksonet al., 2016; Whiting et al., 2009). While interpreting these effects is challenging, they show that there are multiple ways in which auditory brainstem neurons respond to changes in their inputs. Thus, further studies and more coordinated efforts among different research groups are needed to determine whether the different responses relate to the etiology of hearing loss, experimental procedure, age, animal species, and/or genetic strain. Importantly, the field will benefit from studies addressing whether the plasticity observed in response to hearing loss is compensatory or pathological.

Modulators of synaptic activity including nitric oxide (NO), serotonin (SE), and noradrenaline (NE) could play a key role in our understanding of the plasticity of auditory brainstem synapses in the normal and in the hearing impaired (see for review Schofield and Hurley, 2018). Recently, a novel form of synaptic plasticity mediated by NO has been reported between T-stellate cells of the ventral CN (**Cao et al., 2019). T-stellate cells are key glutamatergic neurons that convey spectral information from the cochlea (via the auditory nerve) to brainstem nuclei, the inferior colliculus, and the thalamus (Mellott et al., 2014). Donata Oertel and colleagues (**Cao et al., 2019) found that T-stellate cells are functionally interconnected, and that NO signaling potentiates this interconnection, which generates a short-term central gain change at the level of the CN. This novel gain control mechanism could account for the ability of T-stellate cells to enhance the encoding of spectral cues. Neuronal NO expression increases in animals who experience noise exposure or tinnitus (Coomber et al., 2014, 2015); thus, Cao et al. (2019) raised the possibility that the central gain control mechanism could be plastic. Understanding the role of modulators in central sound processing in normal hearing individuals will shed light on compensatory mechanisms of mature synaptic circuits in the hearing impaired.

Understanding the plasticity and cellular mechanisms that occur during development is critical for comprehending normal patterns of neuronal connectivity and function. Similarly, it is essential to reveal the cellular and plastic mechanisms at mature central auditory synapses in the hearing impaired. The development of new tools including genetically engineered mouse lines or labeling methods for specific neuronal populations and synaptic circuitries in the lower auditory brainstem, are needed. Progress of the field would profit from closer interactions between anatomists, electrophysiologists, and molecular biologists.