Dynamic Patterns of Neurotrophin 3 Expression in the Postnatal Mouse Inner Ear

THE JOURNAL OF COMPARATIVE NEUROLOGY 501:30–37 (2007)

MITSURU SUGAWARA,1,2,3 JOSHUA C. MURTIE,1 KONSTANTINA M. STANKOVIC,1,2 M. CHARLES LIBERMAN,2 AND GABRIEL CORFAS1* 1Neurobiology Program, Children’s Hospital and Department of Neurology, Harvard Medical School, Boston, Massachusetts 02115 2Department of Otology and Laryngology, Harvard Medical School and Eaton-Peabody Laboratory, Massachusetts Eye and Ear Inﬁrmary, Boston, Massachusetts 02114-3096 3Department of Otolaryngology, Head and Neck Surgery, Tohoku University Graduate School of Medicine, Sendai, Japan 980-8574

ABSTRACT Recent studies indicate that neurotrophin 3 (NT3) may be important for the maintenance and function of the adult inner ear, but the pattern of postnatal NT3 expression in this organ has not been characterized. We used a reporter mouse in which cells expressing NT3 also express ␤-galactosidase, allowing for their histochemical visualization, to determine the pattern of NT3 expression in cochlear and vestibular organs. We analyzed animals from birth (P0) to adult (P135). At P0, NT3 was strongly expressed in supporting cells and hair cells of all vestibular and cochlear sense organs, Reissner’s membrane, saccular membrane, and the dark cells adjacent to canal organs. With increasing age, staining disappeared in most cell types but remained relatively high in inner hair cells (IHCs) and to a lesser extent in IHC supporting cells. In the cochlea, by P0 there is a longitudinal gradient (apex Ͼ base) that persists into adulthood. In vestibular maculae, staining gradients are: striolar Ͼ extrastriolar regions and supporting cells Ͼ hair cells. By P135, cochlear staining is restricted to IHCs and their supporting cells, with stronger expression in the apex than the base. By the same age, in the vestibular organs, NT3 expression is weak and restricted to saccular and utricular supporting cells. These results suggest that NT3 might play a long-term role in the maintenance and functioning of the adult auditory and vestibular systems and that supporting cells are the main source of this factor in the adult. J. Comp. Neurol. 501:30–37, 2007. © 2007 Wiley-Liss, Inc.

Indexing terms: NT3; postnatal; cochlea; utricle; saccule; ampulla; spiral ganglion

It is well established that the trophic factor neurotro- express a dominant-negative erbB receptor in inner ear phin 3 (NT3) and its receptor, TrkC, are expressed in the supporting cells, the cochlea develops normally until ϳ3 inner ear during embryonic development (e.g., Farinas et al., 2001) and are essential for normal development of the inner ear (Ernfors et al., 1995; Fritzsch et al., 1997a,b). This article includes Supplementary Material available via the Internet Similarly, evidence that NT3 plays important roles in the at http://www.interscience.wiley.com/jpages/0021-9967/suppmat. adult inner ear is beginning to emerge. For example, Grant sponsor: National Institute on Deafness and Other Communica- Gacek and Khetarpal (1998) showed that recovery from tion Disorders: Grant numbers: R01 DC004820 (to G.C.), Core Grant P30 unilateral surgical labyrinthectomy is impaired in mice DC05209 (to M.C.L.), and RO1 DC0188 (to M.C.L.); Grant sponsor: Mental Retardation Developmental Disabilities Research Center, National Insti- with reduced NT3 expression but not in mice with reduced tutes of Health; Grant number: P30-HD 018655 (to G.C.); Grant sponsor: brain-derived neurotrophic factor (BDNF) or NT4. More Children’s Hospital Otolaryngology Foundation Research Fund (to G.C.). recently, our analysis of transgenic mice in which *Correspondence to: Gabriel Corfas, Division of Neuroscience, Children’s neuregulin-erbB receptor signaling is blocked in cochlear Hospital, 300 Longwood Ave., Boston, MA 02115. E-mail: [email protected] supporting cells in adults suggested that neuregulin- Received 17 March 2006; Revised 31 August 2006; Accepted 29 Septem- induced NT3 production by supporting cells of the organ of ber 2006 Corti is critical for long-term survival of spiral ganglion DOI 10.1002/cne.21227 neurons (Stankovic et al., 2004). In these mice, which Published online in Wiley InterScience (www.interscience.wiley.com).

NT3 EXPRESSION IN THE POSTNATAL INNER EAR 31 weeks of age, when type I spiral ganglion neurons begin to flushing the fixative solution through the oval window. degenerate. Immediately preceding this degeneration, Tissues were then washed three times for 30 minutes at there is a specific and significant reduction in the levels of room temperature and incubated in staining solution (5 ⅐ ⅐ NT3 mRNA in the cochlea. Based on these results, and the mM K3Fe(CN)6,5mMK4Fe(CN)6 3H2O, 2 mM MgCl2 observation that NT3 is expressed by inner hair cell (IHC) 6H2O, 0.01% Na-deoxycholate, 0.02% NP-40, 1 mg/ml supporting cells, we proposed that NT3 plays an impor- X-gal) at 37°C in the dark for 1 or 4 (P0), 5 (P5 and P10), tant role in promoting the long-term survival of type I or 6 hours (P15-P135). Tissues were rinsed with PBS for spiral ganglion neurons. 3–5 minutes, postfixed overnight with the same fixing Information on the pattern of NT3 expression in the solution as above, washed with PBS, and decalcified in 4% postnatal ear is limited. Some studies reported that NT3 EDTA. Tissues were then embedded in Araldite by using is expressed by IHCs of the adult cochlea (Farinas et al., a rapid dehydration protocol to minimize washout of re- 2001; Pirvola et al., 1994), whereas other studies have action product. No differences in the intensity or pattern demonstrated that this neurotrophin is expressed by IHCs of lacZ staining were observed before and after dehydra- and their supporting cells (Stankovic et al., 2004). To tion. Care was taken to treat all cochleae identically: define the pattern of expression of NT3 in the postnatal staining was done in large batches, with ears from all inner ear, we studied mice in which the Escherichia coli postnatal ages included in a single staining run. Araldite- lacZ gene is integrated into the NT3 locus (Farinas et al., embedded materials were sectioned (at 20 ␮m) on a His- 1994). In this mouse strain, cells that normally express torange (LKB Instruments). Ears of wild-type mice were NT3 also express ␤-galactosidase, allowing for their visu- processed in parallel for control. At least three wild types alization and identification by histochemical staining were processed at each age evaluated (up to and including (Fritzsch et al., 1997a). We analyzed cochlear and vestib- P20). ular organs from P0 to P135, to determine the pattern of NT3 expression during the morphological and functional Immunohistochemistry maturation in the early postnatal period and in the ma- Mice were anesthetized with 2.5% Avertin (0.2 ml/10 g ture organ. We found that NT3 is expressed in a dynamic body weight) and fixed by intracardial perfusion with 4% pattern, with the levels of expression and the areas ex- paraformaldehyde in 0.1 M PBS (pH 7.4). The temporal pressing this gene decreasing with age. In the adult inner bones were dissected, and the cochleae were perfused by ear, NT3 expression was restricted to inner hair cells and flushing the fixative solution through the oval window and their supporting cells in the cochlea and largely to sup- postfixed for 2 hours. Temporal bones were decalcified in porting cells in the vestibular epithelia. This finding 4% EDTA for 3 days at 4°C, cryoprotected in 30% sucrose strengthens the notion that supporting cells are necessary overnight at 4°C, embedded in OCT, and sectioned at 15 for the long-term maintenance and function of the adult ␮m. Sections were washed in PBS and blocked for 30 auditory and vestibular systems. minutes in 25% normal goat serum, 0.4% Triton-X, 100 mM L-lysine, 1% bovine serum albumin, and 0.05% sodium azide. Mouse anti-␤-galactosidase (1:500, Promega, MATERIALS AND METHODS Madison, WI) and rabbit anti-calretinin (1:500, Chemicon, Animals Temecula, CA) in 3% bovine serum albumin were incubated on sections overnight at 4°C. Sections were washed NT3-lacZ mice (courtesy of Dr. Reichardt, University of in PBS and blocked with 5% normal donkey serum for 20 California-San Francisco) were generated by targeted re- minutes followed by incubation with donkey anti-mouse placement of the NT3-coding exon with a construct con- Alexa-610 (1:250, Molecular Probes, Eugene, OR) and don- taining a lacZ gene cDNA and the PKCneo marker (Fari- key anti-rabbit Oregon Green (1:250, Molecular Probes) nas et al., 1994). Mice were maintained at the animal for 1 hour at room temperature. facility at Children’s Hospital, and all procedures were carried out following protocols approved by the Children’s Imaging Hospital Animal Care and Use Committee. At least two All images of LacZ-stained tissues were obtained with animals (four ears) were evaluated at every age except for digital cameras (Orca, Hamamatsu, or Spot). Images were P135, which included only one animal. Age-matched, wild- processed by gamma correction and unsharp mask algo- type littermates were used as controls. rithm (amount: 70%; radius: 9px; threshold: 1 level) by Physiological testing using Adobe Photoshop (version 6.0). Images of immuno- fluorescence were obtained with an LSM 510 Zeiss laser Because the NT3-lacZ mice have only one normal NT3 scanning confocal microscope with a 40ϫ objective. gene, we assessed cochlear function by measuring auditory brainstem responses (ABRs) and distortion product otoacoustic emissions (DPOAEs) according to standard RESULTS techniques (Kujawa and Liberman, 2006). Threshold sensitivity was no different in NT-lacZ hemizygotes and wild- Expression gradients in inner-ear whole type controls. mounts ␤ Temporal bones of NT3-LacZ mice of different postnatal LacZ ( -Gal) staining ages from P0 to P135 were processed by ␤-galactosidase Mice were anesthetized with 2.5% Avertin (0.2 ml/10 g histochemistry and embedded in plastic. The staining pat- body weight) and fixed by intracardial perfusion with tern was first analyzed in whole mounts (Figs. 1, 2), which 2% paraformaldehyde/0.2% glutaraldehyde in 0.1 M allow visualization of the general pattern of expression as phosphate-buffered saline (PBS, pH 7.4). The temporal well as of gradients over space and time. As depicted in bones were dissected, and the cochleae were perfused by Figure 1, when temporal bones of P0 mice were developed The Journal of Comparative Neurology. DOI 10.1002/cne

32 SUGAWARA ET AL. for the LacZ reaction for 1 hour, the ␤-galactosidase stain- As shown in Figure 2, longer incubations times showed ing was most intense in the cochlea, where it presented in that ␤-galactosidase is expressed in the postnatal murine an apex Ͼ base gradient similar to that observed during inner ear in a dynamic pattern, with the levels of expres- embryogenesis (Farinas et al., 2001). With this incubation sion steadily decreasing from apex to base and with in- time, the vestibular sensory epithelia exhibited very light creasing postnatal age, even beyond the age at which the staining. When tissues from older animals were subjected epithelia are otherwise mature, both structurally and to similar treatment, staining was very light (not shown). functionally (about P21 for the cochlear portion; Kujawa Therefore, we extended the incubation time to increase and Liberman, 2006). Control ears (without the lacZ gene the sensitivity of the staining. insert) never showed consistent positive reaction product in any structures of the inner ear. At P0, the saccular and utricular maculae were also clearly and darkly stained, the former appearing as a dark U-shaped band and the latter as an ovoid spot. The saccular membrane was strongly stained, and the label continued into Reissner’s membrane of the cochlear duct, especially in the basal turn. The crista from each of the semicircular canals presented a more complex staining pattern, because each of these sensory organs consists of two centrally located lacZ- positive hair-cell patches (one at each end of the ampullary ridge) and two peripherally located lacZ-positive patches of dark cells, the cells responsible for ion transport in the vestibular portion of the inner ear (see below). In the cochlear portion of the inner ear, the whole mount images of Figure 2 clearly show a general decrease in staining intensity with increasing age. Also visible at P10 (and later) was the tunnel of Corti, the light strip that Fig. 1. Left: NT3-LacZ expression (blue reaction product) in whole separates the spiraling cochlear label into inner and outer mounts of the inner ear at postnatal day 0 (P0). The tissue was bands, corresponding to the inner and outer hair cell ar- developed for the LacZ reaction for 1 hour. The staining was most eas, respectively (see below). By P15, even in these low- intense in the cochlea, where it presented in an apex-to-base gradient. Tissues of wild-type (WT) animals at all postnatal ages (see P0, right power views, it is clear that the staining in the inner hair panel) did not produce any LacZ staining, even when developed for 4 cell area was more intense than that in the outer hair cell hours. Scale bar ϭ 0.5 mm in right panel (also applies to left panel). area.

Fig. 2. NT3-LacZ expression (blue reaction product) in whole the cristae of the three semicircular canals (posterior [P], lateral [L], mounts of the inner ear at different postnatal ages. Key structures are and superior [S]) and the maculae of the utricle and sacccule. The identiﬁed in the image of the ear at postnatal day 0 (P0). Cochlear postnatal age, which ranged from P0 to P135, is shown. Scale bar ϭ apex and base are indicated. Vestibular organs are labeled, including 0.5 mm in lower right panel (applies to all). The Journal of Comparative Neurology. DOI 10.1002/cne

NT3 EXPRESSION IN THE POSTNATAL INNER EAR 33

pression was not detected at any age in the spiral liga- ment, the spiral limbus, or the stria vascularis, or in any neural structures in the spiral ganglion or modiolus. (These neural structures do not appear in the images of Fig. 4.) Examination of the organ of Corti (right panels of Fig. 4) showed NT3-lacZ signal in all structures of the nascent sensory epithelium at P0; however, even at this early stage, the darkest signal was in the region of the hair cells and their immediately adjacent supporting cells (see dashed circles in P0, fourth column). By P10–P15, the label was more clearly restricted to hair cells and supporting cells (see P15, third column), i.e., Deiters’ cells (supporting the outer hair cells), pillar cells (surrounding the tunnel of Corti), and the inner border and inner phalan- geal cells (surrounding inner hair cells). By P15, a radial Fig. 3. Immunostaining for ␤-galactosidase (␤-Gal). Sections through the apical region of the P15 cochlea were stained with anti- gradient (inner hair cell area darker than outer hair cell bodies against ␤-galactosidase (red) and the inner hair cell marker area) was clearly established. In this regard, the pillar-cell calretinin (green). The ␤-galactosidase staining is present in all hair staining (both inner and outer pillars) appeared to track cells and supporting cells (SCs) but absent from the tectorial mem- with that of the outer hair cell area, not the inner hair cell brane. As is the case for the LacZ staining (Fig. 4), the intensity of the area. At the oldest age examined, label appeared to be immunostaining in the inner hair cell (IHC) is stronger than that in restricted to the inner hair cell area. In the basal turn, OHCs and supporting cells. Scale bar ϭ 25 ␮m in upper right panel (applies to all). DIC: differential interference contrast. label appeared only in the inner hair cells themselves and not in their support cells. Saccule and utricle. The low-power cross-sections of the saccule and utricle (Fig. 5, left panels) showed that In the vestibular organs of the inner ear, the whole NT3-lacZ staining is restricted at all ages beyond P0 to the mount images of Figure 2 also show a decrease in the sensory epithelia, i.e., the macular regions containing hair staining intensity of all the hair cell organs. At P20 and cells, and the saccular membrane. At P0, there was some beyond, the expression levels appeared highest in the staining of cells in the marginal zone, visible in the image saccule and lowest in the semicircular canal cristae; how- from the utricle. No staining at any age was visible in the ever, even at the oldest age investigated (P135), there was utricular membrane, nor in the mesenchymal cells under- still detectable staining in all hair cell organs of the ves- lying the epithelia or in neural structures including Scar- tibular system. pa’s ganglion. The sensory epithelia of the saccule and utricle display Tissue localization in sectioned material regional variation in their cellular, synaptic, and neural To investigate the tissue localization of NT3 expression architecture (reviewed by Desai et al., 2005). One impor- with higher resolution, i.e., to identify the cells expressing tant organizational landmark in these hair cell organs is the ␤-galactosidase, the plastic-embedded whole mounts the striola, a swathe of hair cells running along the middle were sectioned and analyzed at higher magnification. Rep- of the epithelium, which marks the dividing line between resentative images of the cochlear duct, the utricular and groups of hair cells with opposing hair-bundle polarities. saccular epithelia, and the epithelia of the posterior am- At all ages, NT3 expression was higher in striolar than pulla are shown in Figures 4, 5, and 6, respectively. Im- extrastriolar regions in both the saccule and the utricle portantly, to ensure that the blue reaction product truly (Fig. 5, left panels). In the mouse inner ear, the saccular identified ␤-galactosidase expressing, cells we also stained striola is more curvilinear than the utricular striolar (as sections of the apical region of the P15 cochlea with anti- seen clearly in the whole mounts of Fig. 2); thus the ␤-galactosidase antibodies. Figure 3 shows that saccular striola typically appears as two dark lacZ- ␤-galactosidase antibodies labeled all hair cells and sup- positive bands in the sections, whereas the utricular st- porting cells, the same cells that were labeled by the riolar appears as only one band. ␤-galactosidase reaction (Fig. 4). Furthermore, the rela- Examination at higher magnification (Fig. 5, right pan- tive intensity of the immunostaining of the different cell els) showed that, at the earliest ages (P0–P5), the staining type was similar to that of the histochemistry (IHC Ͼ intensity within the sensory epithelium was comparable OHCs ϭ supporting cells). Thus, these results indicate in hair cells and their supporting cells. With increasing that the blue-labeled cells represent NT3-expressing cells age, the signal became weaker and, in contrast to the and that the intensity of the labeling reaction provides cochlea, became increasingly limited to the supporting semiquantitative information about the levels of NT3 ex- cells. By P30, hair cell staining was very weak. pression by the different cells. Crista ampullaris. Ampullae are expanded regions in Cochlea. In the cochlear portion of the inner ear, de- each semicircular canal that contain the sense organs for velopment of hair cells proceeds from basal to apical turns. rotational motion, the crista ampullaris. Each ridge-like Thus, at any given age, apical regions appear more imma- crista is covered with patches of sensory cells and flanking ture than basal regions. The images of the entire cochlear regions of dark cells surrounding the ridge, which play duct (left panels of Fig. 4) across all ages and cochlear important roles in the regulation of ionic composition of regions indicate that strong NT3-lacZ expression is re- endolymph. As seen in the low-power views of Figure 6 stricted to the cells of the organ of Corti, with weaker (left panels), NT3 expression was seen in hair cells and expression seen in the cells of Reissner’s membrane. Ex- supporting cells at the center of the sensory cell patch and The Journal of Comparative Neurology. DOI 10.1002/cne

34 SUGAWARA ET AL.

Fig. 4. NT3-LacZ expression in plastic sections through the co- from the middle of the ﬁrst turn. These two regions correspond to best chlear duct (left columns) and the organ of Corti (right columns) show frequencies of roughly 4 and 25 kHz, respectively. IHC, inner hair cell; the postnatal changes in expression level and locus. Each row repre- OHC, outer hair cell; SC, supporting cell; IP, inner pillar cell; OP, sents a different postnatal age, as indicated. The “apex” views are outer pillar cell; RM, Reissner’s membrane. Scale bar ϭ 100 ␮m for taken from the middle of the second turn; the “base” views are taken left two columns; 25 ␮m for right two columns.

in ampullary dark cells. Expression in both cell types was only in the central zone. As seen in the saccule and utricle, strong at P0 and decreased with age thereafter (Fig. 6). NT3 expression was robust in both supporting cells and Based on hair cell density, the sensory cell patches in hair cells at P0 (Fig. 6, right panels). After P0, NT3 ex- the ampulla are divided into three regions of approxi- pression levels decreased and became restricted to sup- mately equal area (Lindeman, 1969): the innermost cen- porting cells. tral zone, the outermost peripheral zone, and the transi- tional intermediate zone. Similar to the striolar region of the utricle and saccule, type I hair cells outnumber type II DISCUSSION hair cells in the central zone of the crista with roughly By using NT3-LacZ mice, we show that NT3 is ex- equal numbers of type I and type II hair cells in the pressed in the postnatal inner ear in dynamic patterns peripheral zone (Desai et al., 2005). As seen in the low- and that these patterns continue to change with age well power images in Figure 6, NT3 expression was observed beyond the periods in which the inner ear structures be- The Journal of Comparative Neurology. DOI 10.1002/cne

NT3 EXPRESSION IN THE POSTNATAL INNER EAR 35

Fig. 5. NT3-LacZ expression in plastic sections through the utric- number of hair cells was maximal, roughly corresponding to the ular and saccular hair cell epithelia at low power (left columns) and middle section through the maculae. HC, hair cell; SC, supporting high power (right columns) show the postnatal changes in expression cell. Scale bar ϭ 100 ␮m for left two columns; 25 ␮m for right two level and locus. Postnatal ages are as indicated. For each organ, the columns. section chosen for photographic documentation was that for which the come functionally and morphologically mature. Therefore, observations in postnatal ears, and none provides data it is possible that NT3 plays a long-term role in the main- from an age-graded series of postnatal animals. Thus the tenance and functioning of the adult auditory and vestib- present report complements the existing literature. The in ular systems. situ study in rats suggests that an apical-basal gradient in Previous studies have used either in situ hybridization NT3 expression levels may appear in the P7 ear, although or an NT3-lacZ reporter to study the expression patterns no supporting images are presented. Our data clearly of NT3 in embryonic and fetal cochleae from either rat or show an apical-basal gradient in NT3 expression in all mouse (Pirvola et al., 1992, 1994; Fritzsch et al., 1999; ears at ages beyond P0. The postnatal gradient in expres- Qun et al., 1999; Farinas et al., 2001). All studies agree sion of this neurotrophic factor could be a contributor to that NT3 expression can be seen in hair cells and/or their neural presbycusis, i.e., the observation that age-related supporting cells throughout late embryonic development. degeneration of cochlear neurons in the absence of signif- However, these earlier studies include only fragmentary icant hair cell loss occurs in a base-apex gradient, with a The Journal of Comparative Neurology. DOI 10.1002/cne

36 SUGAWARA ET AL.

2004). When DN-erbB4 was expressed in the inner ear by supporting cells, mice with one copy of the transgene had a reduction in high-frequency hearing, whereas mice with two copies had hearing loss extending to all frequency regions. These results suggest that a relatively low level of expression of DN-erbB4 could induce a significant reduction in NT3 expression in the base, where expression of this neurotrophin is normally the lowest, thus resulting in the frequency-specific hearing loss. The presence of NT3 in adult IHC supporting cells, in addition to IHCs, suggests that supporting cells provide trophic support to sensory neurons in the spiral ganglion. These supporting cells closely ensheath the unmyelinated peripheral terminals of neurons underneath IHCs and OHCs, and they express multiple glial markers (Anniko et al., 1986; Furness and Lehre, 1997; Vega et al., 1999; Rio et al., 2002). Thus, they may well play roles similar to those of glial cells, which have been implicated in the promotion of neuronal survival. The importance of supporting cells in regulating long-term neuronal survival and neuronal sprouting in the organ of Corti was suggested by the above-mentioned study in which blockade of the neuregulin signaling pathway in cochlear supporting cells led to a hearing loss due to postnatal degeneration of cochlear neurons (Stankovic et al., 2004). The importance of supporting cells was further supported by our study of long-term effects of ototoxic drugs, which showed that, in areas of hair cell loss, neuronal survival was enhanced when supporting cells remained intact, even years after the insult (Sugawara et al., 2005). Interestingly, the long- term (Ͼ 5-year) neuronal survival was particularly strong in the apical turn, where, according to the present study, the levels of NT3 expression remain highest well into adulthood. The functional significance of NT3 in cells of the Reiss- ner’s membrane, which separates two markedly different cochlear fluids, potassium (Kϩ)-rich endolymph and sodium (Naϩ)-rich perilymph, is unclear. However, previous workers have also noted the fetal expression of BDNF by cells of the Reissner’s membrane (Farinas et al., 2001). Cells of Reissner’s membrane express high levels of many ion transporters (Yoshihara and Igarashi, 1987; Stankovic et al., 1997; King et al., 1998; Yeh et al., 1998), and they ϩ Fig. 6. NT3-LacZ expression in plastic sections through the am- participate in absorption of Na from endolymph (Lee and pulla of the posterior canal at low power (left column) and high power Marcus, 2003). Similarly, NT3 is expressed by cells in the (right column). Postnatal ages are as indicated. Sections are through limiting membrane of the saccule and in dark cells of the the middle of one of the pair of hair-cell-rich “central zones, which form mirroring patches on either end of the ampullary crest. HCs, ampulla, which have roles in regulating vestibular ionic hair cells; SCs, supporting cells. Scale bar ϭ 100 ␮m for left two environment (Wangemann, 1995; Stankovic et al., 1997). columns; 25 ␮m for right two columns. It is not known whether cells of Reissner’s and limiting membranes express TrkC, but it is interesting to note that Trk signaling is involved in regulation of the expression and activities of ion channels (reviewed by Huang and prominent steady decline in the numbers of surviving cells Reichardt, 2003). For example, Trk receptor signaling in the basal turn (Otte et al., 1978). leads to modulation of Naϩ and Caϩ currents, and activa- This gradient of NT3 expression may also explain re- tion of several members of the TRP family of cation chan- sults from our previous studies using transgenic mice nels (reviewed by Huang and Reichardt, 2003). Therefore, expressing a dominant-negative form of the receptor the possibility that NT3 acts in an autocrine manner to erbB4 in supporting cells of the inner ear (Stankovic et al., facilitate ionic regulation through activation of TrkC in 2004). This dominant-negative receptor was designed to both the auditory and vestibular systems needs to be abolish glial response to neuregulin, a factor that can considered. induce the expression of glial signals that have retrograde Unlike the adult cochlea, in which IHCs and supporting effects on their associated neurons (Chen et al., 2003; cells express NT3, the adult vestibular end organs express Prevot et al., 2003). In the cochlea, this retrograde signal NT3 only in supporting cells. Our data suggest that NT3 appears to be NT3, which contributes to the long-term in vestibular supporting cells provides trophic support for survival of adult spiral ganglion neurons (Stankovic et al., peripheral terminals of neurons that synapse on type I The Journal of Comparative Neurology. DOI 10.1002/cne

NT3 EXPRESSION IN THE POSTNATAL INNER EAR 37 and type II hair cells. A possible role for NT3 in vestibular Gacek RR, Khetarpal U. 1998. Neurotrophin 3, not brain-derived neuro- function in adults was hinted at in a study showing that trophic factor or neurotrophin 4, knockout mice have delay in vestibular compensation after unilateral labyrinthectomy. Laryngoscope 108: recovery from unilateral surgical labyrinthectomy takes 671–678. much longer in NT3 heterozygous mice compared with Huang EJ, Reichardt LF. 2003. Trk receptors: roles in neuronal signal wild-type mice and mice with reduced expression of BDNF transduction. Annu Rev Biochem 72:609–642. or NT4 (Gacek and Khetarpal, 1998). This suggested that King M, Housley GD, Raybould NP, Greenwood D, Salih SG. 1998. Expres- NT3 is the early trophic regulator of vestibular compen- sion of ATP-gated ion channels by Reissner’s membrane epithelial sation. In light of our current and earlier findings (Stank- cells. Neuroreport 9:2467–2474. ovic and Corfas, 2003), we suggest that NT-3 involved in Kujawa SG, Liberman MC. 2006. Acceleration of age-related hearing loss vestibular compensation may also be derived from the by early noise exposure: evidence of a misspent youth. J Neurosci 26:2115–2123. vestibular periphery. Lee JH, Marcus DC. 2003. Endolymphatic sodium homeostasis by Reiss- It is clear that NT3 plays a significant role in the devel- ner’s membrane. Neuroscience 119:3–8. oping auditory and vestibular systems. The current study Lindeman HH. 1969. Studies on the morphology of the sensory regions of demonstrating nonmonotonic changes in NT3 expression the vestibular apparatus with 45 figures. Ergeb Anat Entwicklungsge- in the adult auditory and vestibular systems suggests that sch 42:1–113. NT3 may have long-lasting roles in maintenance and func- Otte J, Schuknecht HF, Kerr AG. 1978. Ganglion cell populations in tioning of these systems. Furthermore, the specific expres- normal and pathological human cochleae. Implications for cochlear sion patterns suggest that this long-term role for NT3 may implantation. Laryngoscope 88:1231–1246. be necessary not only for the most accepted roles for neu- Pirvola U, Ylikoski J, Palgi J, Lehtonen E, Arumae U, Saarma M. 1992. Brain-derived neurotrophic factor and neurotrophin 3 mRNAs in the rotrophins, e.g., the preservation of synaptic contacts and peripheral target fields of developing inner ear ganglia. Proc Natl Acad neuronal survival, but possibly also for the maintenance of SciUSA89:9915–9919. ionic balance in the inner ear through regulation of the Pirvola U, Arumae U, Moshnyakov M, Palgi J, Saarma M, Ylikoski J. 1994. expression and activity of ion channels. Coordinated expression and function of neurotrophins and their receptors in the rat inner ear during target innervation. Hear Res 75:131– 144. LITERATURE CITED Prevot V, Rio C, Cho GJ, Lomniczi A, Heger S, Neville CM, Rosenthal NA, Ojeda SR, Corfas G. 2003. Normal female sexual development requires Anniko M, Thornell LE, Gustavsson H, Virtanen I. 1986. Intermediate neuregulin-erbB receptor signaling in hypothalamic astrocytes. J Neu- filaments in the newborn inner ear of the mouse. ORL J Otorhinolar- rosci 23:230–239. yngol Relat Spec 48:98–106. Qun LX, Pirvola U, Saarma M, Ylikoski J. 1999. Neurotrophic factors in Chen S, Rio C, Ji RR, Dikkes P, Coggeshall RE, Woolf CJ, Corfas G. 2003. the auditory periphery. Ann N Y Acad Sci 884:292–304. Disruption of ErbB receptor signaling in adult non-myelinating Rio C, Dikkes P, Liberman MC, Corfas G. 2002. Glial fibrillary acidic Schwann cells causes progressive sensory loss. Nat Neurosci 6:1186– protein expression and promoter activity in the inner ear of developing 1193. and adult mice. J Comp Neurol 442:156–162. Desai SS, Ali H, Lysakowski A. 2005. Comparative morphology of rodent vestibular periphery. II. Cristae ampullares. J Neurophysiol 93:267– Stankovic KM, Corfas G. 2003. Real-time quantitative RT-PCR for low- 280. abundance transcripts in the inner ear: analysis of neurotrophic factor expression. Hear Res 185:97–108. Ernfors P, Van De Water T, Loring J, Jaenisch R. 1995. Complementary Stankovic KM, Brown D, Alper SL, Adams JC. 1997. Localization of pH roles of BDNF and NT-3 in vestibular and auditory development. ϩ Ϫ Neuron 14:1153–1164. regulating proteins H ATPase and Cl /HCO3-exchanger in the guinea pig inner ear. Hear Res 114:21–34. Farinas I, Jones KR, Backus C, Wang XY, Reichardt LF. 1994. Severe sensory and sympathetic deficits in mice lacking neurotrophin-3. Na- Stankovic K, Rio C, Xia A, Sugawara M, Adams JC, Liberman MC, Corfas ture 369:658–661. G. 2004. Survival of adult spiral ganglion neurons requires erbB receptor signaling in the inner ear. J Neurosci 24:8651–8661. Farinas I, Jones KR, Tessarollo L, Vigers AJ, Huang E, Kirstein M, de Caprona DC, Coppola V, Backus C, Reichardt LF, Fritzsch B. 2001. Sugawara M, Corfas G, Liberman MC. 2005. Influence of supporting cells Spatial shaping of cochlear innervation by temporally regulated neu- on neuronal degeneration after hair cell loss. J Assoc Res Otolaryngol rotrophin expression. J Neurosci 21:6170–6180. 6:136–147. Fritzsch B, Farinas I, Reichardt LF. 1997a. Lack of neurotrophin 3 causes Vega JA, San Jose I, Cabo R, Rodriguez S, Represa J. 1999. Trks and p75 losses of both classes of spiral ganglion neurons in the cochlea in a genes are differentially expressed in the inner ear of human embryos. region-specific fashion. J Neurosci 17:6213–6225. What may Trks and p75 null mutant mice suggest on human develop- Fritzsch B, Silos-Santiago I, Bianchi LM, Farinas I. 1997b. The role of ment? Neurosci Lett 272:103–106. neurotrophic factors in regulating the development of inner ear inner- Wangemann P. 1995. Comparison of ion transport mechanisms between vation. Trends Neurosci 20:159–164. vestibular dark cells and strial marginal cells. Hear Res 90:149–157. Fritzsch B, Pirvola U, Ylikoski J. 1999. Making and breaking the inner- Yeh TH, Herman P, Tsai MC, Tran Ba Huy P, Van den Abbeele T. 1998. A vation of the ear: neurotrophic support during ear development and its cationic nonselective stretch-activated channel in the Reissner’s mem- clinical implications. Cell Tissue Res 295:369–382. brane of the guinea pig cochlea. Am J Physiol 274:C566–576. Furness DN, Lehre KP. 1997. Immunocytochemical localization of a high- Yoshihara T, Igarashi M. 1987. Cytochemical localization of Caϩϩ-ATPase affinity glutamate-aspartate transporter, GLAST, in the rat and activity in the lateral cochlear wall of the guinea pig. Arch Otorhino- guinea-pig cochlea. Eur J Neurosci 9:1961–1969. laryngol 243:395–400.