The noise-induced or age-related loss of synaptic connections between auditory-nerve fibers (ANFs) and inner hair cells (IHCs) has been documented in several mammalian species, including mouse (Kujawa and Liberman, 2009; Sergeyenko et al., 2013), guinea pig (Hickman et al., 2020; Shi et al., 2013), gerbil (Gleich et al., 2016; Jeffers et al., 2021), monkey (Valero et al., 2017) and human (Wu et al., 2019; Wu et al., 2021). Noise exposures causing only transient threshold elevation and no loss of sensory cells can nevertheless cause immediate loss of more than half the IHC synapses and ultimate degeneration of large numbers of ANFs (Kujawa and Liberman, 2009). In the aging mouse or gerbil ear, these IHC/ANF synaptic connections disappear before hair cells die (Gleich et al., 2016; Sergeyenko et al., 2013). In aging humans, IHCs have lost, on average, half of their ANF innervation by age 60 yrs (Wu et al., 2019) and the degree of primary neural degeneration is greater in those with a history of acoustic overexposure (Wu et al., 2021).

Cochlear synaptopathy in animal work is typically quantified by immunostaining the organ of Corti for a key protein in the pre-synaptic ribbon (CtBP2) and for a glutamate-receptor subunit, e.g. GluA2, (Khimich et al., 2005; Liberman et al., 2015) or another post-synaptic component of glutamatergic synapses, e.g. PSD-95 (Shi et al., 2015). In the normal cochlea, each ANF typically contacts a single IHC by a single unmyelinated terminal, forming a single synapse at which a single pre-synaptic ribbon in the IHC is apposed to a single patch of specialized membrane in the ANF terminal, where the post-synaptic glutamate receptors are localized (Hua et al., 2021; Liberman, 1980). In a normal mouse cochlea, each IHC contains up to 20 pairs of closely apposed pre- and post-synaptic puncta, depending on cochlear location (Kujawa and Liberman, 2009; Meyer et al., 2009; Stamataki et al., 2006), and the decrease in such counts after noise exposure, or in the aging ear, is a highly quantitative metric of synaptopathy (Liberman et al., 2015; Sergeyenko et al., 2013).

Although a massive and rapid post-exposure reduction in synaptic counts has been documented by many labs, in many animal models, reports differ as to the degree to which these synapses spontaneously re-appear with increasing post-exposure time. The original mouse studies of synaptopathy concluded that there was no significant recovery based on ribbon counts carried out at 1 day, 3 days and 8 wks post-exposure, as well as a loss of ANF cell bodies in the spiral ganglion of roughly equivalent degree (50%) at 2 years post-exposure (Kujawa and Liberman, 2009). A later study in the guinea pig reported that a massive decrease in ribbon counts (to 35% of normal) at one day post exposure was followed by almost complete recovery (to 90% of normal) at one month (Shi et al., 2013). This dramatic post-exposure recovery in guinea pig was corroborated in follow-up studies in our lab (Hickman et al., 2020). More surprising than this apparent species difference, several studies have suggested that there is also significant post-exposure synaptic recovery in a different strain of mouse, i.e. C57BL/6J as compared to the CBA/CaJ strain used in the early studies from our laboratory (Kaur et al., 2019; Kim et al., 2019; Shi et al., 2015; Song et al., 2016) .

An inter-strain difference in the reversibility of noise-induced synaptopathy would be important, because of the powerful insights it could provide into the underlying recovery mechanisms. However, it is difficult to compare synaptic counts across laboratories, because many synaptic puncta at short post-exposure times are faintly stained (see Figure 7 in (Hickman et al., 2018)), and most studies provide no detail as to how they divide this continuum of staining intensity into a binary decision of present vs. absent. Thus, recovery of synaptic counts reported in one study could be seen simply as an increase in staining intensity in another, depending on the signal-to-noise ratio of the immunostaining, the laser and detector gains during confocal acquisition, the gamma adjustment of the resultant images and the thresholding criteria for defining when faint fluorescent signals are categorized as synaptic puncta.

The purpose of the present study was to provide a more definitive test of possible interstrain differences between CBA/CaJ and C57BL/6J with respect to synaptic recovery, by direct comparison in the same lab using the same techniques, and by careful quantification of immunostaining intensity as a function of post-exposure time. Although present results are somewhat different depending on exactly which frequency region is evaluated, we find that ribbon counts in both strains show a non-monotonic relation to post-exposure time, decreasing from 0 hrs to 1 day post exposure and then either recovering back to the 0 hr values, or worsening, at 8 wks post exposure. Synaptic counts in both strains tend to be worst at 0 hrs and 1 day post-exposure, and both strains show significant recovery at some frequency regions by 2 wks and 8 wks. In both strains the decrease in synaptic counts is mirrored by an increase in the numbers of “orphan” ribbons and a selective decrease of staining intensity of the GluA2 puncta. On the other hand, the two strains do show a significant difference with respect to the response of ANF terminals to the noise. In C57Bl/6J, there is a striking extension of neurites towards the apical poles of the IHCs that is selective to the synaptopathic region at 1 day post-exposure and is followed by retraction to their normal positions at longer post-exposure survivals. No such neural plasticity is seen in CBA/CaJ.