Perineuronal nets in subcortical auditory nuclei of four rodent species with differing hearing ranges

Abstract

Perineuronal nets (PNs) are aggregates of extracellular matrix molecules that surround some neurons in the brain. While PNs occur widely across many cortical areas, subcortical PNs are especially associated with motor and auditory systems. The auditory system has recently been suggested as an ideal model system for studying PNs and their functions. However, descriptions of PNs in subcortical auditory areas vary, and it is unclear whether the variation reflects species differences or differences in staining techniques. Here, we used two staining techniques (one lectin stain and one antibody stain) to examine PN distribution in the subcortical auditory system of four different species: guinea pigs (Cavia porcellus), mice (Mus musculus, CBA/CaJ strain), Long-Evans rats (Rattus norvegicus), and naked mole-rats (Heterocephalus glaber). We found that some auditory nuclei exhibit dramatic differences in PN distribution among species while other nuclei have consistent PN distributions. We also found that PNs exhibit molecular heterogeneity, and can stain with either marker individually or with both. PNs within a given nucleus can be heterogeneous or homogenous in their staining patterns. We compared PN staining across the frequency axes of tonotopically organized nuclei and among species with different hearing ranges. PNs were distributed non-uniformly across some nuclei, but only rarely did this appear related to the tonotopic axis. PNs were prominent in all four species; we found no systematic relationship between the hearing range and the number, staining patterns or distribution of PNs in the auditory nuclei.

1 | INTRODUCTION

Perineuronal nets (PNs) are aggregates of extracellular matrix molecules that surround some neurons in the brain (Karetko & Skangiel-Kramska, 2009; Morawski, Brückner, Arendt, & Matthews, 2012, for review). In cortex, PNs are best known for surrounding parvalbumin-positive fast-spiking inhibitory interneurons, but PNs have also been associated with as many as one third of pyramidal cells (Härtig, Brauer, & Brückner, 1992; Hausen et al., 1996; Wegner et al., 2003). In the visual cortex, PNs develop at the close of the visual critical period, and extending the critical period with dark rearing delays the development of PNs (Pizzorusso et al., 2002; Ye & Miao, 2013). Further, abnormal patterns of ocular dominance due to early monocular deprivation can be corrected in adult rats via enzymatic digestion of PNs in the primary visual cortex (Pizzorusso et al., 2006). Similar results have been achieved in other areas: digestion of PNs in basolateral amygdala leads to formation of fear memories that are more vulnerable to erasure (Gogolla, Caroni, Lüthi, & Herry, 2009); digestion of PNs in striatum leads to a wider and more variable gait (Lee, Leamey, & Sawatari, 2012); digestion of PNs in auditory cortex does not affect initial learning of a go/no-go task but facilitates reversal training (Happel et al., 2014). Throughout cortex, digestion of PNs leads to phenomena related to increased plasticity (reviewed by Sorg et al., 2016). Although regulation of plasticity, specifically inhibition of structural plasticity, is one of the main functions associated with PNs, they have also been suggested to affect synaptic plasticity, protect against oxidative stress, and support fast spiking in the cells they surround (Härtig et al., 1999; Corvetti & Rossi, 2005; Beurdeley et al., 2012; Suttkus, Rohn, Jäger, Arendt, & Morawski, 2012; Cabungcal et al., 2013; de Vivo et al., 2013).

Less research has been done involving PNs in subcortical areas. Subcortical PNs are especially associated with motor and auditory nuclei, although not exclusively (Seeger, Brauer, Härtig, & Brückner, 1994; Bertolotto, Manzardo, & Guglielmone, 1996; Sonntag, Blosa, Schmidt, Rübsamen, & Morawski, 2015). Neonatal conductive hearing loss affects PNs in several auditory brainstem nuclei of the superior olivary complex (SOC), and PNs develop in the medial nucleus of the trapezoid body (MNTB) coincident with the maturation of reliable fast spiking (Taschenberger & von Gersdorff, 2000; Myers, Ray, & Kulesza, 2012). Studies examining the function of PNs in the auditory brainstem (specifically the MNTB) have shown that PNs are important to proper spike timing and spike rate, which are in turn essential for encoding sound characteristics such as frequency, intensity, and location (Oertel, 1999; Eggermont, 2001; Blosa et al., 2015; Balmer, 2016). Even though the importance of PNs to auditory function seems clear, there are inconsistencies in the literature regarding PN distribution in the subcortical auditory system. For example, a study in rat found that PNs were more prominent in the cortical areas of the inferior colliculus (IC) and lacking in the central nucleus, while a study in guinea pig found that PNs were densest in the central nucleus (rat: Bertolotto et al., 1996; guinea pig: Foster, Mellott, & Schofield, 2014). This particular discrepancy may be due to a difference in PN distribution between these two species, or it could reflect differences in the methods employed to visualize PNs (an antibody stain in the first study, and a lectin stain in the second). Another point of uncertainty in the literature is whether PN staining varies along the tonotopic axes of auditory nuclei. Some authors have reported gradients of PN staining along a tonotopic axis and others report homogeneous staining, but the analyses have been limited to one or two nuclei rather than an examination across the auditory pathway (Hilbig, Nowack, Boeckler, Bidmon, & Zilles, 2007; Blosa et al., 2013; Sonntag et al., 2015; Fech, Calderón-Garcidueñas, & Kulesza, 2017). Inter-study discrepancies such as these led us to ask two main questions about PN staining in the subcortical auditory system: (a) does apparent variation among species simply reflect the staining method, and, (b) to what extent does PN distribution vary with frequency representation, either among species with different hearing ranges or across the frequency axes of tonotopically organized nuclei?

To answer these questions, we employed double staining of PNs with two markers. One of the main components of PNs are chondroitin sulfate proteoglycans (CSPGs), which are composed of a protein core attached to multiple glycosaminoglycan (GAG) side chains (for review see Galtrey & Fawcett, 2007). Wisteria floribunda agglutinin (WFA) is a common lectin that binds to the GAG side chains of CSPGs, thereby labeling PNs and unaggregated CSPGs in the extracellular matrix. We combined WFA staining with immunohistochemical staining using an antibody to the protein portion of aggrecan (AGG), a specific type of CSPG. Both lectin staining and antibody staining have been used widely to investigate PNs throughout the auditory brainstem (Sonntag et al., 2015, for review). We employed this double staining procedure in tissue from four different species: naked mole-rats (Heterocephalus glaber), guinea pigs (Cavia porcellus), mice (Mus musculus), and rats (Rattus norvegicus). Species were chosen to represent a range of evolutionary development within the order Rodentia (Figure 1a), and because their hearing ranges differ in frequency range enough to allow an investigation of differences in PN distribution based on frequency representation (Figure 1b). The results lead to four main conclusions. First, by comparing staining for PNs in these four species, we conclude that some auditory brainstem nuclei exhibit dramatic differences in PN distribution among species, while other nuclei show consistent PN distributions across species. Second, by comparing the staining patterns of two PN markers, we also conclude that individual PNs can stain with either marker or with both, and PNs within a nucleus can be heterogeneous or homogenous in this respect. Third, we found little evidence for a strong relationship between PN staining and the frequency-specific (i.e., tonotopic) organization of individual nuclei. One exception is the lateral superior olivary nucleus (LSO), where PNs occur through-out the nucleus but PN staining is heavier in low frequency regions. A similar relationship was not apparent in other nuclei with well-known tonotopic organization. Fourth, we found no systematic difference in PN staining or distribution among the species with different hearing ranges: PNs were present in many of the same nuclei in naked mole-rats, guinea pigs, rats, and mice.

FIGURE 1.

The four species examined here belong to the order Rodentia, and represent diversity in their evolutionary relationships and hearing ranges. (a) A phylogenetic tree representing the order Rodentia (adapted from Tree of Life Project, 2006). Mice and rats come from the subgroup Muroidea and are closely related, while naked mole-rats come from the subgroup Bathyergomorpha, and guinea pigs come from the subgroup Caviomorpha. (b) Audiograms of each of the species represented in this study, adapted from Heffner and Heffner (1993), Heffner, Heffner, and Masterton (1971), Heffner, Heffner, Contos, and Ott (1994) and Radziwon et al. (2009). In each study audiograms were obtained behaviorally. Mice and rats hear better at high frequencies, while guinea pigs hear over a broader frequency range. Naked mole-rats have higher hearing thresholds than the other species examined, and hear lower frequencies than mice or rats [Color figure can be viewed at wileyonlinelibrary.com]

2 | MATERIALS AND METHODS

All procedures were conducted in accordance with the Northeast Ohio Medical University Institutional Animal Care and Use Committee and NIH guidelines. Results are described from three pigmented guinea pigs of either gender, two male CBA/CaJ mice, two subordinate female naked mole-rats, and two male Long-Evans rats. Guinea pigs ranged in weight from 330 g to 863 g, corresponding to an age range of approximately 5 weeks to approximately 20 weeks. Both mice were approximately 12 months old, rats were approximately 4 and 7 months old, and naked mole-rats were approximately 4–4.5 years old. All animals were adults, but none were considered aged (note that the maximal life span for a naked mole-rat is about 30 years; Edrey, Hanes, Pinto, Mele, & Buffenstein, 2011).

2.1 | Perfusion

Guinea pigs and rats were deeply anesthetized with isoflurane until breathing stopped and corneal and withdrawal reflexes were absent. They were then perfused transcardially with Tyrode’s solution, followed by 4% paraformaldehyde in 0.1M phosphate buffer (PB), pH 7.4, then 4% paraformaldehyde containing 10% sucrose. For mice and naked mole-rats, the perfusate was normal saline, followed by 4% paraformaldehyde. Differences in perfusion methods reflect the different sources of brains; there is no evidence that these small differences would affect the PN staining. Brains were removed and stored in fixative containing 25% sucrose overnight. Brains that were not cut on the following day were moved to 0.1M PB containing 25% sucrose for storage. Brains were frozen and cut into 40 μm sections in the transverse plane on a sliding microtome. Sections were collected in six series (for guinea pigs and rats), or three series (for naked mole-rats and mice). In one guinea pig, one series was stained for cytochrome oxidase to visualize borders between subdivisions of the medial geniculate body (MG). In each animal, one or more series was double stained for PNs.

2.2 | Staining

Initial experiments that stained tissue singly with anti-AGG or with WFA in guinea pig tissue had staining that was similar to double-stained tissue, indicating that double-staining with both markers does not affect the ability of either marker to bind to its target. For double staining for PNs, tissue sections were washed in PBS, then permeabilized for 30 min at room temperature in a solution containing 0.2% Tri-ton X-100 in phosphate buffered saline (PBS; 0.9% NaCl in 0.01M PB, pH 7.4). Non-specific staining was blocked for 1 hr at room temperature in a solution containing 0.1% Triton X-100 and 10% normal goat serum or normal donkey serum in PBS. The type of normal serum used here and throughout the staining protocol was matched to the host of the secondary antibody used to label the anti-AGG antibody. The first PN stain (biotinylated WFA, 1:100, cat# B1355, Vector Labs) was applied for 1 hr at room temperature in a solution containing 0.1% Triton X-100 and 1% normal serum in PBS. Following a wash in PBS, the biotinylated WFA was labeled with either Alexa Fluor 488- or Alexa Fluor 647-conjugated streptavidin (1:100, cat# S-11223 and S-21374, respectively, Life Technologies) in a solution containing 0.1% Triton X-100 and 1% normal serum in PBS for 1 hr at room temperature. The second PN stain (rabbit anti-AGG antibody, AB1031, Millipore, 1:100) was then applied overnight at 4°C in a solution containing 0.1% Triton X-100 and 1% normal serum. The anti-AGG antibody was labeled with either an Alexa Fluor 594-conjugated donkey anti-rabbit antibody (1:100, cat# A-21207, Life Technologies) or an Alexa Fluor 750-conjugated goat anti-rabbit antibody (1:100, cat# A-21039, Life Technologies) in a solution containing 1% normal serum in PBS for 1 hr at room temperature. Tissue was then washed in PBS, mounted from a 0.2% gelatin solution onto gelatin-coated slides, allowed to air-dry, then coverslipped with DPX mounting medium.

For cytochrome oxidase staining we used the procedure detailed in Anderson, Wallace, and Palmer (2007). Briefly, tissue sections were washed in PBS, then treated with a solution that combined 20 mg of diaminobenzidine hydrochloride in 10 mL dH₂O with 30 mg of cytochrome c and 3 g sucrose in 30 mL of 0.1M PB. Sections were incubated either at 4°C overnight or at 37°C for 3–5 hr. Sections were mounted from a 0.2% gelatin solution onto gelatin-coated slides, allowed to air-dry, then coverslipped with DPX mounting medium. Care was taken to process tissue series from a given species until staining was consistent between individuals (in general, inconsistent staining appeared attributable to uneven or inadequate fixation).

2.3 | Antibody characterization

Details of antibodies and stains are given in Table 1. The anti-AGG antibody has been previously characterized in mouse (Blosa et al., 2013; manufacturer’s website), and our anti-AGG staining in mouse tissue was morphologically similar to previous staining. Anti-AGG staining was also morphologically similar in the three other species studied here. No peptide is currently available to perform preadsorption controls with this antibody, but given the morphological similarity, the unique extracellular nature of this marker and the complete lack of intracellular staining, we believe the antibody is labeling the intended target.

TABLE 1.

Antibodies and stains

Name	Targeted structure	Manufacturer; RRID; host species; catalog number; lot number; mono/polyclonal	Dilution used
Primary antibody
Anti-AGG	GST of fusion protein: aa1117–1326 mouse AGG	Millipore; AB1031; 2808221; AB_90460; rabbit; polyclonal	1:100
Secondary antibodies
AF594-anti-rabbit	Rabbit gamma-immunoglobulin	Invitrogen; A-21207; 1602780; AB_141637; donkey; polyclonal	1:100
AF750-anti-rabbit	Rabbit gamma-immunoglobulin	Invitrogen; A-21039; 1717037; AB_1500687; goat; polyclonal	1:100
Streptavidin tags
AF488- streptavidin	Biotin	Invitrogen; S-11223; 1733116; AB_2336881	1:100
AF647- streptavidin	Biotin	Invitrogen: S-21374; 1738253; AB_2336066	1:100
Lectin stain
WFA	N-acetylgalactosamine Vector;	B-1355; 2B0312; AB_2336874	1:100

WFA has been used widely as a PN stain across many species (Sonntag et al., 2015, for review). Our WFA staining was morphologically similar to staining in previous studies. We performed a preadsorption control on tissue from each species, where WFA was preadsorbed with an equal volume of N-acetylgalactosamine residue (500 mM; S-9001; Vector) for 1 hr at room temperature before being applied to the tissue. We observed a complete lack of PN and general extracellular matrix staining in tissue from all four species when treated with preadsorbed WFA, indicating that the WFA is binding only to the expected target.

To ensure that the secondary antibodies and streptavidin tags used here do not result in staining in the absence of the anti-AGG anti-body or the WFA, we performed a primary omission control on tissue from each species, where secondary antibodies were applied without the previous application of primary antibodies or stains. All four reagents (AF594-conjugated donkey anti-rabbit, AF750-conjugated goat anti-rabbit, AF488-conjugated streptavidin, and AF647-conjugated streptavidin) were applied in a cocktail in PBS at a concentration of 1:100, following normal permeabilization and blocking steps as described above. No staining resulted from this procedure in any of the four species, indicating that our secondary antibodies and streptavidin tags themselves did not produce any of the staining reported.

2.4 | Data analysis

Each region was assessed through its full rostral to caudal extent in each animal in the study. For guinea pigs and rats, sections were spaced 240 μm apart, while for mice and naked mole-rats the sections were spaced 120 μm apart (to account for the overall smaller brain and smaller nuclei). Differences in PN staining between the two markers, among nuclei or among species were often dramatic; for example, a particular nucleus could have a high density of PNs stained with one marker and no PNs stained with the other marker, and that pattern could be reversed in a different species or a different nucleus in the same species. The prominence of PNs in each nucleus was rated visually as minimal, moderate or high for each PN marker. This qualitative approach reflects the dramatic differences and facilitates comparisons with both PN markers across a large number of nuclei in all four species. Future studies, focused on a smaller number of nuclei, could make use of more quantitative analyses to address variations of PN density within a nucleus, or relationships of PNs to a particular cell type (e.g., fast-spiking cells: Cabungcal et al., 2013; Blosa et al., 2015; GABAergic cells: Foster et al., 2014)

2.5 | Image capture and processing

Photomontage images and high magnification photomicrographs were collected with NeuroLucida systems (MBF Bioscience) attached to either a Zeiss Axioimager Z2 or a Zeiss Axioimager M2 microscope. Photomontages were collected as image stacks at 1 μm depth intervals using a 20X objective lens, and high magnification photomicrographs were collected as image stacks using a 63X oil-immersion lens (NA = 1.4) and structured illumination (Apotome 2, Zeiss) to provide optical sections at 0.2 μm depth intervals. Final images are maximum intensity projections of collected stacks that were exported as tif files. Adobe Photoshop (CS6, Adobe Systems) was used to crop and colorize images, and to position scale bars. Brightness and contrast levels were adjusted globally when necessary. Final figures were assembled using Adobe Illustrator. To the extent possible, the photomicrographs reflect representative observations that were consistent across sections and individuals.

3 | RESULTS

Results are presented for each level of the subcortical auditory system, with emphasis on comparing staining of PNs among species and between markers. In addition to staining PNs, both markers used here stain unaggregated CSPGs in the extracellular matrix. The appearance of this matrix can also vary between markers and areas. We comment on unaggregated CSPG staining when it helps to distinguish nuclei or is notably similar or dissimilar between species.

3.1 | PN staining is similar across species in the ventral cochlear nucleus (VCN), but differs in the dorsal cochlear nucleus (DCN)

PN staining is observed with both markers in the cochlear nucleus (Figure 2a). In the VCN, PNs stain prominently across all four species. PN staining is heavy throughout the VCN in the naked mole-rat, the guinea pig, and the mouse. In the rat, PN staining is heavier around the periphery of the VCN and present but less prominent in the central region of the VCN. Across species, PNs are homogenous and tend to stain well with both the WFA and AGG markers (Figure 2b). The broad distribution as well as the shape of the PNs, which provide a silhouette of the surrounded soma, suggest that PNs are associated with each of the major cell types in the VCN (including spherical and globular bushy cells, stellate/multipolar cells and octopus cells). In none of the species was there evidence for PNs surrounding granule cells. There are no obvious dorso-ventral differences in staining, indicating a lack of staining difference across the VCN tonotopic axis. There is, however, a group of prominently stained PNs at the ventral tip of the VCN that tends to stain more heavily with AGG across species (Figure 2a). This may correspond to the cochlear root neurons, identifiable in rat, mouse, and gerbil as large neurons present at the auditory nerve root (López, Merchán, Bajo, & Saldaña, 1993; Cant & Benson, 2003).

FIGURE 2.

PN staining is similar across species in the VCN, but differs in the dorsal cochlear nucleus (DCN). (a) Low magnification photomontages of the CN of naked mole-rat, guinea pig, rat, and mouse. WFA is shown in green and AGG is shown in magenta; overlap of the two markers appears white. In each species, the VCN and DCN are easily distinguished. The VCN shows prominent PN staining across all four species, while the DCN shows heavy staining for PNs in the guinea pig (especially in the fusiform cell layer), lighter staining in mouse and rat, and PN staining is completely absent in naked mole-rat. Scale bars = 250 μm. DCN = dorsal cochlear nucleus, VCN = ventral cochlear nucleus. Orientation: D = dorsal, M = medial. (b) High magnification examples of PNs in the VCN of guinea pig (GP) and naked mole-rat (NMR). VCN PNs stain equally well with both markers, WFA and AGG, and are relatively homogenous through the VCN. Scale bars = 20 μm. (c) PN staining is heterogeneous in the DCN as compared to the VCN, and differs based on species. Arrows indicate high magnification examples of PNs in the DCN of guinea pig (left) and mouse (middle). A high magnification image from naked mole-rat CN is shown on the right. The white dotted line represents the border between the VCN and the DCN, and emphasizes the lack of PN staining in the naked mole-rat DCN. Scale bars = 20 μm [Color figure can be viewed at wileyonlinelibrary.com]

In the dorsal cochlear nucleus (DCN), the guinea pig shows extensive staining for PNs in the fusiform cell layer (stained more heavily with WFA), and in the deep layer (stained more heavily with AGG; Figure 2a). Rat and mouse show some PN staining in the DCN but, especially in mouse, staining is more prominent in the general extracellular matrix rather than aggregated into PNs. In naked mole-rat, PN staining is completely absent in the DCN, despite well-stained PNs being present in the VCN of the same section (Figure 2a). This implies an absence of PNs rather than a failure to stain. The differences among species in the DCN are emphasized in Figure 2c. The guinea pig (first column) displays staining of prominent PNs around large cells in the fusiform cell layer (arrows). PNs are also present in mouse DCN (arrows, second column). However, the PN-surrounded cells are smaller, and the PNs do not tend to stain as brightly as those in guinea pig, making the general extracellular matrix staining appear more prominent by comparison. In naked mole-rat DCN (third column), PNs are completely absent, as is extracellular matrix staining with WFA or AGG. The dotted white line in the third column represents the VCN–DCN border, and emphasizes the extreme difference in PN staining between the VCN and DCN in the naked mole-rat.

3.2 | PN staining is similar across species in the SOC; some nuclei have heterogeneous PNs and some nuclei have homogenous PNs

In all four species examined here, PN staining in the SOC can be used to distinguish many of the nuclei of the SOC (Figure 3a). Although the arrangement of SOC nuclei varies among species, PN staining within a given nucleus is generally consistent across species. For example, PNs in the LSO stain with either WFA (Figure 3b, green arrows) or with AGG (magenta arrows), but rarely with both markers. In rat, naked mole-rat, and, to a lesser extent, guinea pig, the staining within the LSO appears to vary across its tonotopic axis: the lateral LSO, representative of lower frequencies, exhibits heavier staining with both markers (Figure 3a). Whether this represents a difference in the number of PNs, or a difference in staining for unaggregated CSPGs is unclear. Further, this gradient of staining is either much subtler or absent in the mouse LSO. The MNTB contains particularly prominent PNs in each of these species. Notably, PNs in the MNTB stain homogenously with both WFA and AGG (white arrows, Figure 3c). This differentiates PNs in the MNTB from PNs in the LSO, and may indicate functional differences between PNs in these areas (compare Figure 3b and c). Unlike the LSO, the MNTB lacks obvious staining differences across its tonotopic axis (medio-lateral). The medial superior olivary nucleus (MSO) is minimal in mouse and small in rat; neither species has substantial PN staining in the nucleus. The MSO is prominent in guinea pig and stains heavily with AGG. This, coupled with previous work in naked mole-rat (Heffner & Heffner, 1993; Gessele, Garcia-Pino, Omerbašić, Park, & Koch, 2016) led us to label the heavily AGG-stained area between the LSO and MNTB in the naked mole-rat as the MSO. Similarities across species are less robust for the periolivary nuclei, which likely reflects the well-known variation of these nuclei across species (Schwartz, 1992; Thompson & Schofield, 2000). Discrete periolivary nuclei have not been described in the naked mole-rat, and we observed no periolivary staining for PNs in naked mole-rat in this study (Heffner & Heffner, 1993; Gessele et al., 2016). In guinea pig and mouse, the lateral nucleus of the trapezoid body (LNTB) stains prominently for PNs, with some PN staining in rat LNTB (Figure 3a). In guinea pig, rat, and mouse, the ventral nucleus of the trapezoid body (VNTB) stains prominently with WFA, which reveals a mixture of PN and extracellular matrix staining (Figure 3a). The superior paraolivary nucleus (SPN) is characterized by less prominent, WFA-stained PNs in guinea pig, rat, and mouse. In the naked mole-rat, a notable collection of PNs just dorsal to the MNTB may represent an SPN in this species (Figure 3a, asterisk).

FIGURE 3.

PN staining is similar across species in the SOC; some nuclei have heterogeneous PNs, and some nuclei have homogenous PNs. (a) Low magnification photomontages of the SOC of guinea pig, rat, mouse and naked mole-rat. WFA is green, AGG is magenta and overlap is white. Presence and arrangement of SOC nuclei differ among species, however staining within a given nucleus is relatively consistent across species. Scale bars = 250 μm. LNTB = lateral nucleus of the trapezoid body, LSO = lateral superior olivary nucleus, MNTB = medial nucleus of the trapezoid body, MSO = medial superior olivary nucleus, SPN = superior paraolivary nucleus, VNTB = ventral nucleus of the trapezoid body. Orientation: D = dorsal, M = medial. * = collection of PNs in naked mole-rat that may represent an SPN in this species. (b) High magnification examples of PNs in the LSO of naked mole-rat (left) and mouse (right). LSO PNs tend to stain with either WFA (green arrows) or AGG (magenta arrows), but not both markers. Scale bars = 20 μm. NMR- naked mole-rat. (c) White arrows indicate examples of PNs in the MNTB of guinea pig (left) and rat (right). MNTB PNs stain with both WFA and AGG. Scale bars = 20 μm. GP = guinea pig [Color figure can be viewed at wileyonlinelibrary.com]

3.3 | PNs are present in the nuclei of the lateral lemniscus (NLL) across species, but the prominence of the two markers differs among nuclei and among species

In each species examined here, three subdivisions of the nuclei of the lateral lemniscus can be distinguished: the dorsal, intermediate, and ventral nuclei of the lateral lemniscus (DNLL, INLL, and VNLL, respectively). The NLL generally stain more heavily than the surrounding tissue for PNs, however the prominence of the two markers differed among the nuclei and among species. In the naked mole-rat (Figure 4a, first two panels) the DNLL stains heavily with AGG, while the INLL and VNLL stain with both markers. In contrast, the DNLL in the guinea pig, rat, and mouse stains more heavily with WFA. In the guinea pig and the mouse, the INLL and the VNLL stain with both markers. This was also true in the rat, however AGG staining was relatively more prominent in the INLL and VNLL as compared to the other species. Examples of PNs from the DNLL of each species are shown in Figure 4b. A PN in naked mole-rat DNLL stains with AGG, but WFA staining is minimal (magenta arrow; general extracellular matrix staining with WFA is still present). This contrasts with PNs from guinea pig, rat, and mouse DNLL, where PNs stain with both markers, although the WFA staining tends to be more prominent (white arrows).

FIGURE 4.

PNs are present in the nuclei of the lateral lemniscus (NLL) across species, but the prominence of the two markers differs among nuclei and among species. (a) Low magnification photomontages of the NLL in naked mole-rat, guinea pig, rat, and mouse. WFA is shown in green and AGG is shown in magenta. The NLL are divided into three subdivisions: dorsal NLL (DNLL), intermediate NLL (INLL), and ventral NLL (VNLL). In naked mole-rat and mouse, two tissue sections are shown so that all three subdivisions can be visualized. PNs are present across the NLL, however the marker that most prominently stains them differs among nuclei and among species. For example, the DNLL stains more heavily with AGG in naked mole-rat, and more heavily with WFA in guinea pig, rat, and mouse. Scale bars = 250 μm. D = dorsal, M = medial. (b) High magnification examples of PNs in the DNLL of naked mole-rat, guinea pig, rat, and mouse. PNs in the DNLL of naked mole-rat stain almost exclusively with AGG (magenta arrow), while PNs in other species stain with both markers, usually more heavily with WFA (white arrows). Scale bars = 10 μm. GP = guinea pig, NMR = naked mole-rat [Color figure can be viewed at wileyonlinelibrary.com]

3.4 | PN distribution in the IC varies based on species, and PNs in the IC are heterogeneous

Although both markers stain PNs in the IC of all four species, the IC shows the greatest heterogeneity of PN staining of all the subcortical auditory regions analyzed here (Figure 5a). While there are differences among the classic IC subdivisions (i.e., central nucleus, lateral cortex, dorsal cortex), the variations generally did not follow closely the subdivision borders (see Foster et al., 2014; Beebe, Young, Mellott, & Schofield, 2016 for detailed quantitative descriptions in guinea pigs). Rather, some of the prominent variations in PNs occurred in regions that include a small part of a classical subdivision and may or may not cross subdivision borders. The central part of the IC has prominent staining with both markers in naked mole-rat and guinea pig, whereas rat and mouse have relatively little staining centrally. In addition, naked mole-rat IC has areas of increased staining dorso-laterally and ventro-medially, with areas of decreased staining dorso-medially and ventro-laterally. In guinea pig, staining is most prominent in the ventro-medial IC and throughout central regions of the IC, with areas of decreased staining dorso-medially and ventro-laterally, similar to the staining pattern in naked mole-rat. Rat and mouse each have areas of increased staining ventro-laterally and ventro-medially, and mouse exhibits increased staining dorso-medially as well. Although some variations in staining patterns roughly follow the dorso-lateral to ventro-medial tonotopic axis in the IC, there is no clear association between high or low frequencies and increased or decreased PN staining across species. Throughout the IC, individual PNs stain with WFA (Figure 5b, green arrows), with AGG (magenta arrows), or with both markers (white arrows). This was true across species (examples shown from mouse, top, and rat, bottom, Figure 5b).

FIGURE 5.

PN distribution in the IC varies based on species, and PNs in the IC are heterogeneous. (a) Low magnification photomontages of the IC of naked mole-rat, guinea pig, rat, and mouse. WFA is shown in green and AGG is shown in magenta. Naked mole-rat and guinea pig have heavy staining centrally, where rat and mouse have much less staining. All four species show heavy staining for PNs laterally, although the area of increased staining is dorsolateral in naked mole-rat and ventro-lateral in rat and mouse. Naked mole-rat, guinea pig, and mouse show increased staining ventro-medially. A ventro-lateral area (*) contains distinctive PNs in all four species (see panel c). Scale bars = 250 μm. Aq = cerebral aqueduct. D = dorsal, M = medial. (b) High magnification example of PNs in mouse (top) and rat (bottom) IC. PNs in the IC are heterogeneous in regards to staining patterns. Some PNs stain with both markers (white arrows), while some stain only with WFA (green arrows), or only with AGG (magenta arrows). Scale bars = 20 μm. (c) High magnification examples of PNs on the ventro-lateral IC border (in area marked by * in panel a). PNs in this area extend unusually long distances onto the proximal processes of the cell (small white arrows). Scale bars = 20 μm. GP = guinea pig, NMR = naked mole-rat [Color figure can be viewed at wileyonlinelibrary.com]

In all four species, we noted an area along the ventro-lateral border of the IC (outlined separately from the rest of the IC and marked with an asterisk in Figure 5a) that contains PNs surrounding relatively large cells. This region was characterized by especially prominent PNs, and decreased staining of unaggregated CSPGs in the extracellular matrix as compared to the rest of the IC. In some instances, the PNs in this region were clearly within the traditional borders of the IC, but in other cases (especially in naked mole-rat, see Figure 5a) this collection of PNs clearly extended ventrally beyond the usual IC border. These PNs differ from most other subcortical auditory PNs in that they extend a notable distance from the cell body onto the proximal processes of the cell. Examples from each species are shown in Figure 5c, and small white arrows indicate PN staining continuing onto proximal processes. These PNs are similar to PNs throughout the IC in that they stain with WFA, AGG, or with both markers. For simplicity, only the marker that best stained each PN is shown in Figure 5c.

3.5 | PNs are largely absent in the MG, except for a pocket of PN-surrounded cells in the medial subdivision of the MG (MGm)

Previous studies examining PN distribution in the MG have reported very little or no PN staining (Sonntag et al., 2015, for review). The four species examined here lack PN staining in the MG for the most part, in agreement with previous studies; however, we noted an area on the medial border of the MG that stains with both PN markers in guinea pig, rat, and mouse (Figure 6a, large white arrows). In guinea pig, an adjacent section was stained for cytochrome oxidase, and used to elucidate the borders between MG subdivisions (Anderson et al., 2007). Based on these borders, it appears that the area of PN staining is located in the MGm rather than medial to the MGm border. There was no evidence for a similar group of PNs along medial MG in the naked mole-rat (Figure 6a). Given the presence of WFA-stained PNs ventral to the MG in naked mole-rat, we attribute the lack of medial staining in this species to an actual lack of PNs rather than a failure of staining. Individual PNs in the MGm stain with both WFA and AGG, and these areas also exhibit a high level of general extracellular matrix staining (Figure 6b).

FIGURE 6.

PNs are largely absent in the MG, except for a pocket of PN-surrounded cells in the medial subdivision (MGm). a. Low magnification photomontages of the MG of naked mole-rat, guinea pig, mouse, and rat. WFA is shown in green and AGG is shown in magenta. In guinea pig, MG subdivisions were determined with an adjacent section stained for cytochrome oxidase. In the naked mole-rat, mouse, and rat, the MG border was approximated using the PN staining shown. PNs are mostly absent in the MG, except in the medial region of the MG in guinea pig, mouse, and rat (large white arrows). Scale bars = 250 μm. D = dorsal, M = medial, MGd = dorsal subdivision of the MG, MGv = ventral subdivision of the MG. b. White arrows indicate high magnification examples of PNs in the medial region of the MG of mouse (top) and rat (bot-tom). MG PNs stain with both markers. Medial MG staining is also characterized by a high level of general extracellular matrix staining. Scale bars = 20 μm [Color figure can be viewed at wileyonline-library.com]

3.6 | Additional areas that stain for PNs consistently across species include the vestibular nuclei (VN), the red nucleus (RN), and the reticular formation (RF)

While this report focuses on PN staining in auditory nuclei, we noted several non-auditory areas that stained prominently for PNs in each of the species examined here: the VN, the RN, and the RF. The VN stain much more heavily for PNs than the surrounding tissue, and this staining includes both PNs and general extracellular matrix staining for unaggregated CSPGs (Figure 7a). The staining is not uniform through-out the nuclei; for example, guinea pig and rat exhibit areas of increased WFA staining (ventro-laterally in the guinea pig and dorso-medially in the rat) but staining with the two markers is similar across species. The same is true in the RN (Figure 7b), with no areas that stain more heavily with either marker. Like the VN, the RN stains much more heavily for PNs than the surrounding tissue. High magnification images of PNs from the VN and the RN show relatively homogenous staining with the two markers (Figure 7c,d). A majority of PNs in both areas stain equally with both WFA and AGG. PNs in the VN sometimes extended onto proximal processes of the surrounded cell (Figure 7c, first column), and sometimes surrounded just the soma (second column). In the RN, PNs surrounded cells in both less densely packed and more densely packed areas (Figure 7d, third column and fourth column, respectively). Areas with denser PN staining in the RN, like the area pictured in the fourth column, tended to also have more staining for unaggregated CSPGs in the neuropil.

FIGURE 7.

Additional areas that stain for PNs consistently across species include the vestibular nuclei (VN) and the red nucleus (RN). (a) Low magnification photomontages of the VN of naked mole-rat, guinea pig, rat, and mouse. WFA is shown in green and AGG is shown in magenta. The VN stain heavily for PNs in each of the species examined here. Scale bars = 250 μm. Cb = cerebellum, IV = fourth ventricle. D = dorsal, M = medial. (b) Low magnification photomontages of the RN of naked mole-rat, guinea pig, rat, and mouse. The RN stains heavily for PNs in each of the species examined here. Scale bars = 250 μm. Orientation: D = dorsal, M = medial. (c) White arrows indicate high magnification examples of PNs from the VN of naked mole-rat and mouse. PNs tend to stain heavily with both markers. Scale bars = 20 μm. NMR = naked mole-rat. (d) White arrows indicate high magnification examples of PNs from the RN of guinea pig and rat. PNs tend to stain heavily with both markers. Scale bars = 20 μm. GP = guinea pig, NMR = naked mole-rat [Color figure can be viewed at wileyonlinelibrary.com]

The RF also contained particularly prominent PNs in each of the species examined here. Figure 8 shows high magnification images of PNs from the RF in the medulla (due to the diffuse nature of the RF, spreading across much of the brainstem tegmentum, no low magnification images are shown). Reticular formation PNs almost always extended long distances onto cellular processes, and tended to stain heavily with both markers used here.

FIGURE 8.

PNs stain heavily across species in the reticular formation (RF). White arrows indicate examples of PNs from the medullary RF in naked mole-rat, guinea pig, rat and mouse. PNs in the RF stain brightly with both markers and almost always extend onto the proximal processes of surrounded cells. Scale bars = 20 μm. GP = guinea pig, NMR = naked mole-rat [Color figure can be viewed at wileyonline-library.com]

4 | DISCUSSION

In this study, we compared PN staining characteristics and distribution in the subcortical auditory systems of four species. We reached four main conclusions. (a) Some auditory nuclei exhibit dramatic differences in PN distribution among species whereas other nuclei show similar PN distributions across species. Table 2 summarizes the overall staining patterns for PNs through the subcortical auditory system of each species studied here. (b) Individual PNs can stain with either or with both of two general PN markers, and PNs within a nucleus can be heterogeneous or homogenous in this respect. (c) There is little evidence for a system-wide relationship between PN staining and frequency-specific (i.e., tonotopic) organization. One exception occurs in the LSO, where PN staining is heavier in low frequency regions. A similar relationship was not apparent in other nuclei with well-known tonotopic organization. Rather, non-homogeneous distribution of PNs within a nucleus appeared to reflect the distribution of different cell types. (d) There was no systematic difference in PN staining among species with different hearing ranges: PNs were present in many of the same nuclei in naked mole-rats, guinea pigs, rats, and mice. An additional, unexpected finding was that some non-auditory areas showed robust PN staining that was less variable across species than staining in auditory areas.

TABLE 2.

Summary of WFA and AGG staining across species and nuclei.

Region	Naked mole-rat		Guinea pig		Rat		Mouse
	WFA	AGG	WFA	AGG	WFA	AGG	WFA	AGG
VCN	++	++	++	++	+	+	++	++
DCN	0	0	++	++	+	+	+	+
MSO	+	++	+	++	0	0	0	0
LSO	++	++	++	++	++	++	++	++
MNTB	++	++	++	++	++	++	++	++
SPN	2	2	+	+	+	+	+	+
LNTB	2	2	+	+	0	0	++	++
VNTB	2	2	++	0	++	0	++	0
DNLL	+	++	++	+	++	+	++	+
INLL	2	2	++	++	++	++	++	++
VNLL	++	++	++	++	++	++	++	++
IC-Central	++	++	++	++	+	+	+	+
IC- Dorsal	+	+	+	+	++	++	++	++
IC-Lateral	+	+	+	+	++	++	++	++
IC-VL Border	++	++	++	++	++	++	++	++
MG-Ventral	0	0	0	0	0	0	0	0
MG-Dorsal	0	0	0	0	0	0	0	0
MG-Medial	0	0	++	++	++	++	++	++

Symbols indicate the presence of PNs stained with the corresponding marker: A “+” indicates a moderate level of staining; a “++” indicates a high level of staining, including intense staining of individual PNs; a “0” indicates lack of staining in the corresponding nucleus. For reference, see Figure 2, which shows PN staining in the DCN is absent (“0”) in naked mole rat, moderate (“+”) in rat and mouse and high (“++”) in guinea pig). A “−” symbol indicates that the nucleus was not observed or nuclear borders were too uncertain to characterize the nucleus in that species. DCN – dorsal cochlear nucleus, DNLL – dorsal nucleus of the lateral lemniscus, IC – inferior colliculus, IC-VL – ventro-lateral border region of IC, INLL – intermediate nucleus of the lateral lemniscus, LNTB – lateral nucleus of the trapezoid body, LSO – lateral superior olivary nucleus, MG – medial geniculate body, MNTB – medial nucleus of the trapezoid body, MSO – medial superior olivary nucleus, SPN – superior paraolivary nucleus, VCN – ventral cochlear nucleus, VNLL – ventral nucleus of the lateral lemniscus, VNTB – ventral nucleus of the trapezoid body

4.1 | Factors affecting PN staining

PN staining can vary between two similar markers. Although widely used, WFA and AGG probably do not stain all PNs in any species or brain region. A study in hippocampus showed that staining for brevican, a CSPG, did not completely match staining with WFA (Ajmo, Eakin, Hamel, & Gottschall, 2008). Further, different antibodies to AGG stain distinct groups of PNs that differ in the glycosylation pattern of their CSPG components (Lander, Kind, Maleski, & Hockfield, 1997; Matthews et al., 2002).

Additionally, PNs can change over development, in disease, and in aging (Belichenko, Miklossy, & Celio, 1997; Friauf, 2000; Mauney et al., 2013; Yamada & Jinno, 2013; Li, Li, Jin, Wang, & Zhao, 2017). We attempted to minimize variation associated with these issues by using healthy adults from each species studied. PN distribution can also be affected by experience, including reductions in sensory input or exposure to enhanced environments (Balmer, Carels, Frisch, & Nick, 2009; Nowicka, Soulsby, Skangiel-Kramska, & Glazewski, 2009; Deák, Bácskai, Gaál, Rácz, & Matesz, 2012; Faralli et al., 2016; Slaker, Barnes, Sorg, & Grimm, 2016; Sorg et al., 2016; Stamenkovic et al., 2017). All of the animals used here were housed in the same animal facility, but were housed separately in species-appropriate conditions. Differences due to experience may be especially relevant in the case of naked mole-rats, which are eusocial mammals. Subordinates are monomorphic, but are capable of undergoing processes to become sexually mature, and therefore reproductive, at any time in their lifespan (reviewed by Edrey, Park, Kang, Biney, & Buffenstein, 2011). We cannot assume that the subordinate females studied here are equivalent to sexually mature females in their auditory PN distribution, as neurobiological changes are involved in the transition to sexual maturity, and estrogenic effects on the auditory brain have been observed (Caras, 2013). It could be interesting to see if PN staining patterns in male or sexually mature female naked mole-rats differ from that described here for subordinate female naked mole-rats. In addition, naked mole-rats are the only species studied here that lack external pinnae as adults, which probably affects their sound localization abilities. This may have consequences for plasticity in the naked mole-rat central auditory system, and might be reflected in some of the differences we observed in PN staining between naked mole-rats and other species.

4.2 | Comparison with previous studies

In general, the results here agreed with previous studies of PNs in subcortical auditory nuclei. As far as we know, there have been no studies of PNs in the auditory system of naked mole-rats, and only one in guinea pigs. The study in guinea pigs found that PNs in the IC are most numerous in the central nucleus, which is consistent with the results here (Foster et al., 2014). The most extensive observations are from rats, and for most areas (CN, SOC, NLL, IC), our results agree with earlier descriptions (Seeger et al., 1994; Bertolotto et al., 1996; Friauf, 2000; Myers et al., 2012). However, in the thalamus, Bertolotto et al. (1996) described PNs scattered throughout the MG while Friauf (2000) described no PN staining in the MG. Both observations differ from our observation of PNs concentrated in the medial MG. Bertolotto et al. (1996) used an anti-body against chondroitin unsulfated proteoglycan whereas Friauf (2000) used an antibody against a specific disaccharide unit on CSPGs (D-glucuronosyl-N-acetylgalactosamine), so the differences among the studies may reflect molecular heterogeneity of the PNs. Costa et al. (2007) describe PN staining using several markers throughout the mouse brain, including in the CN, SOC, NLL, and IC, although details of distribution are not given. The heavy presence of PNs in mouse MNTB was noted by Blosa et al. (2013).

Studies of PNs have also been done in other rodent species, as well as non-rodent mammals. In gerbil VCN, octopus cells, multipolar cells, and possibly globular bushy cells are described as PN-surrounded, while the DCN is described as devoid of PNs (Cant & Benson, 2006). These observations are consistent with our descriptions of heavy PN staining in VCN; the absence of PNs in gerbil DCN matches that of naked mole-rats and differs from the other species described here. Further, gerbil MSO, LSO, and MNTB contain many PNs (Lurie, Pasic, Hockfield, & Rubel, 1997), consistent with results across species in the present study. A study in Madagascan tenrec showed PN staining in the IC, NLL, SOC, and CN (Morawski et al., 2010). Specifically, PN staining was heavier in VCN than in DCN, heavier in the LSO, and heavier along the lateral and ventromedial IC borders, all observations consistent with characteristics of PN staining that we found to be true across multiple species here. WFA stains PNs in the LSO and MSO of dog, and a stain for neurocan, a type of CSPG, showed light staining for PNs in VCN and in the central nucleus of the IC (Atoji & Suzuki, 1992; Atoji, Yamamoto, Suzuki, Matsui, & Oohira, 1997; Fech et al., 2017). Interestingly, Fech et al. (2017) describe gradients of PN staining in MSO and LSO of dog. In rhesus monkey, PNs are present across VCN and DCN, absent in MSO, heavier in medial MNTB and lighter in lateral MNTB, and heavier in the dorsal and external cortices of the IC (Hilbig et al., 2007). These observations, especially the gradient of staining across the medio-lateral axis of the MNTB are largely inconsistent with our observations in rodent. This study also suggests that rhesus monkeys may show a more rat-like and less guinea pig-like distribution of PNs in the IC. In human, PNs are associated with most cell types in VCN, but are largely absent in DCN (Wagoner & Kulesza, 2009). In human SOC, PNs are associated with the MNTB and SPN, which we observed in the rodents studied here, but are absent from LSO and MSO, which is inconsistent with our findings in rodents (Schmidt, Wolski, & Kulezsa, 2010).

4.3 | Molecular heterogeneity of PNs is likely to reflect a variety of functions

Individual PNs can stain with WFA, with AGG, or with both. These two markers bind to the same molecule, but to different parts of that mole-cule (WFA is generally considered to bind to the GAG side chains of CSPGs, especially AGG [Giamanco, Morawski, & Matthews, 2010], while the anti-AGG antibody binds to the protein portion of AGG). Therefore, differences in staining between the two markers, like those we observed in the LSO or the IC, indicate molecular heterogeneity in the CSPGs making up those PNs. Molecular heterogeneity of PNs has been previously reported within and between brain regions, including the IC, the red nucleus, and the hippocampus (Ajmo et al., 2008; Foster et al., 2014; Dauth, et al., 2016; Rácz, Gaál, & Matesz, 2016). In the hippocampus, the molecular makeup of PNs varies in a cell subtype-specific manner (Yamada & Jinno, 2017). Our observations in guinea pig DCN seemed to echo this: PNs in the fusiform cell layer stained more heavily with WFA, while PNs in the deep layer stained more heavily with the anti-AGG antibody. Molecular heterogeneity of PNs probably indicates functional differences between the cells they surround, and differences in the staining of PNs could be used to lead researchers to previously undifferentiated subtypes of neurons.

One implication of molecular heterogeneity is that there is no single “best” stain for PNs. The markers used here are widely considered to be broad markers for PNs, which is consistent with the widespread distribution of PNs seen in all four species in the current study. However, neither marker is ubiquitous. The dorsal nucleus of the lateral lemniscus provides one of numerous examples. In this nucleus, WFA stains PNs prominently in guinea pig, rat and mouse, but shows almost no staining in naked mole-rats, where AGG stains PNs intensely in naked mole-rats. Experiments to study PNs, particularly in an uncharacterized species or nucleus, may require testing of numerous PN markers.

PNs have been implicated in numerous functions, including formation of a structural barrier to axon growth and synapse formation, buffering of cations and protection from glutamate excitotoxicity, promotion of synaptic plasticity, maintenance of an ionic environment to support fast spiking, and regulation of firing rate and spike timing (Corvetti & Rossi, 2005; Beurdeley et al., 2012; Suttkus et al., 2012; Cabungcal et al., 2013; de Vivo et al., 2013; Blosa et al., 2015; Balmer, 2016; Sorg et al., 2016). Molecular heterogeneity of PNs probably has functional consequences (Yamaguchi, 2000). For example, incorporation of smaller CSPGs, such as brevican, results in a tighter PN matrix than incorporation of larger CSPGs, such as AGG. This tighter matrix concentrates the negative charge of the PN molecules (buffering more cations and repelling negatively charged molecules, such as glutamate, to a greater extent), and provides more of a barrier to new synapse formation. These may be examples of functional differences between different types of PNs (for example, PNs that exhibit more WFA-binding vs. less). The heterogeneity of PNs that was observed in the present study suggests that PNs could be serving a variety of functions even within individual nuclei.

4.4 | Homogeneity of staining characterizes PNs in some areas

While heterogeneity of PN staining was common, it was not universal among the auditory nuclei. In some nuclei, or cell groups within a nucleus, the PNs showed relatively uniform staining across species. Such similarities may reflect a specific and conserved function of those PNs. Similar staining characterized the VCN, most SOC nuclei, the ventro-lateral border of the IC and the medial region of the MG. Here and across the literature, the MNTB stands out as staining heavily for PNs. MNTB principal cells are glycinergic cells that provide inhibitory signals to brainstem circuits critical for binaural processing of inter-aural cues for sound localization. These principal cells are postsynaptic to the calyces of Held, specialized giant synapses that have become a model for many studies of synaptic physiology. The association of PNs with the MNTB principal cells and their calyces provides a unique opportunity for studying PN functions. The capability of MNTB principal cells to exhibit high firing rates and precise spike timing is present across species (Kopp-Scheinpflug, Tolnai, Malmierca, & Rübsamen, 2008), and subsequent studies showed that PNs in the MNTB are important to both firing rate and precise spike timing (Blosa et al., 2013; Balmer, 2016).

We also observed prominent groups of PNs in the ventro-lateral IC, medial MG and ventral VCN. A group in the ventral VCN may correspond to cochlear root neurons, which have been identified in some rodents but not others (reviewed in Cant & Benson, 2003). In some species, these neurons are collected in a distinct cluster called the acoustic nerve nucleus, but such a nucleus is not apparent in other species (for example, guinea pig). Cochlear root neurons have been implicated as “sentinels”, detecting sounds and sending direct signals to specific brainstem reflex centers (López, Saldaña, Nodal, Merchán, & Warr, 1999). Is such a circuit absent in species without an identifiable acoustic nerve nucleus? If one assumes that guinea pigs have neurons analogous to the cochlear root neurons, the PNs may serve as a marker for these neurons. López et al. (1993) point out that cochlear root neurons stain heavily for calcium binding proteins, which may allow them to fire rapidly in the presence of loud, higher frequency stimuli. In cortex, PNs surround parvalbumin positive, fast-spiking interneurons, establishing a link between PNs and calcium binding protein expression (Härtig et al., 1992). Cochlear root neurons may represent more evidence of this link.

The placement of prominent, large PNs in the ventro-lateral IC is reminiscent of a region described in mustached bats (where it was considered a ventral part of the external nucleus of the IC; Gordon & O’Neill, 2000). The cells in this region were selective for directionality and rate of frequency-modulated sweeps, which bats use in echolocation but which are also important properties of many complex sounds, including speech. In mustached bats, this region receives input from the nucleus of the central acoustic tract and sends projections to extra-lemniscal auditory thalamus and deep layers of the superior colliculus, leading to its classification as an extralemniscal region of the IC.

The group of PNs we observed in the auditory thalamus appears to be in the medial MG (although MG subdivision borders can be difficult to ascertain without special stains; Anderson et al., 2007). Anderson and Linden (2011) described neurons in the MGm of mice that have shorter response latencies and a higher probability of response to acoustic stimuli than neurons in other parts of the MG. PNs could very well contribute to these properties.

Sonntag et al. (2015) suggested that the auditory system is ideal for studying the functions of PNs due to their high density in auditory nuclei relative to other nuclei. We observed similarly high densities of PNs in auditory nuclei. In addition, the greater heterogeneity of PNs in auditory regions may facilitate correlation of PN molecular constitution with the wide variety of functions attributed to PNs.

4.5 | PNs are prominent regardless of the animal’s hearing range

Härtig et al. (1999) showed a relationship between PNs and cells that express potassium channels with the Kv3.1b subunit, implicating PNs in support of fast-spiking cells. Subsequent work provides support for a relationship between PNs and KV3.1b in some but not all auditory brainstem regions that have fast-spiking cells (see Sonntag et al., 2015, for review). Hilbig et al. (2007) concluded that WFA staining in rhesus monkeys demonstrated an association of PNs with high frequency processing. This was most evident in the MNTB, where WFA staining varied from intense in the medial (high-frequency) part of the nucleus to light in the lateral (low-frequency) part of the nucleus. We observed a subtle gradient of staining along the tonotopic axis in the LSO, with PNs more intensely stained in the lateral, low-frequency area. The MNTB, as well as most other tonotopically organized nuclei, did not display a clear gradient of PN staining in the species studied here.

A relationship between PNs and sound frequency could also potentially be related to the range of frequencies that an animal can hear. We examined four rodents with different hearing ranges (Figure 1b). Comparisons of PNs across these species failed to reveal a simple relationship between PNs and hearing range. PNs are prominent in high frequency specialists like mice and rats, but they are also prominent in naked mole-rats, which hear much lower frequencies with much higher overall thresholds. Moreover, PNs are prominent across tonotopic regions (i.e., are found in both low frequency and high frequency regions) of nuclei in animals such as guinea pig that have a wide hearing range. While these data do not exclude a role for PNs related to processing high frequency sounds, they suggest that the functions of PNs are broad, with presumably significant functions for the processing of low frequency sounds as well.