am: auditory meatus; bl: bony spiral lamina; bm: basilar membrane; bma: basilar membrane zona arcuata; bmp: basilar membrane zona pectinata; bv: bronx waltzer mouse mutation; cb: cell body; c: cochlea; Cx26: Connexin 26; Cx30 Connexin 30; cp: cuticular plate; dc. Deiters’ cells; fn: foramina nervosa; Hz: hertz; IHC: Inner Hair Cell; inner pillar cell; J: junction; kHz: kilohertz; kV: kilovolt; MRM: Magnetic Resonance Microscopy; m: malleus; μm: micrometre; mm: millimetre; mv: microvilli; mV: millivolt; nm: nanometre; n: nucleus; OHC: Outer Hair Cell; opc: outer pillar cell; PC: Personal Computer; pH: power of hydrogen; pnf: peripheral auditory nerve fibres; rm: Reissner’s membrane; SD: Standard Deviation; SEM: Scanning Electron Microscope; sg: Spiral ganglion; slm: spiral limbus; spo: supporting cells for the spiral organic; st: stereocilia; sv: synaptic vesicles; TEM: Transmission Electron Microscope; Tl+: thallium; tm: tectorial membrane; tym. tympanic membrane; VIII: peripheral auditory nerve; VASP: Vasodilator-stimulated phosphoprotein;

Whilst the inner ear from M. unguiculatus has been examined using a variety of histological and physiological methodologies, none describe the fine structural organization of hair cells and pillar cells of the cochlea. The inner ear is situated in the osseous labyrinth of the auditory periotic in the tympanic part of the temporal bone, lying below the squama and in front of the mastoid process. A series of interconnected membranous pouches, ducts and canals are contained within the osseous labyrinth which consists of three main parts; the vestibule, the semicircular canals and cochlea. The cochlea is divided into three longitudinal compartments; the scala tympani, scala media (containing the organ of Corti) and scala vestibuli [1,2], coiled around the modiolus (a hollow bone pillar containing the cochlea nerve). In general, the hair cells along the organ of Corti are divided into a single row of Inner Hair Cells (IHCs) and three rows of Outer Hair Cells (OHCs) [2]. The position of the cells is specified during early development and possibly follows the layout of the supporting pillar cells [3-6]. The structure of the pillar cells is formed from actin and myosin filaments, which are especially present in the apical region forming part of the cuticular plate [7,8] that increase in number with location around the basilar membrane and reticular lamina [9-11]. In addition, supporting cells have bundled microtubules containing tubulin that provides structural support to the organ of Corti [12].

Waldeyer [13] first introduced the hypothesis of radial selfmovements within the pillar cells, and Rhode & Geisler [14] came to a similar conclusion when mathematically modeling the displacement between opposing points on the tectorial membrane and reticular lamina. Schick et al. [15] found Vasodilator-stimulated phosphoprotein (VASP) and the protein Zyxin co-localised with pan-actin, suggesting actin based dynamics in pillar cells was possible. VASP is a crucial factor in the regulation of actin dynamics and the formation of actin filaments [15,16]. VASP has been found to be associated with motility in the Outer Hair Cells (OHCs), and its expression along with F-actin in pillar cells coincides with the onset of hearing [16-18]. The bronx waltzer (bv) mouse mutation is an autosomal recessive mutation that affects both the inner hair cells (IHCs) and pillar cells in the organ of Corti and vestibular system; severer mutations lead to deafness and vestibular dysfunction [19,20]. Until recently, pillar cells were assumed to form a highly rigid bridge between the IHCs and OHCs [11], though this rigidity is contrary to the findings of Gummer et al. [21], Hu et al. [22] and Hemmert et al. [23]. In light of this, it appears that pillar cells may be more dynamically involved in controlling the protracted mechanical properties of the cochlea than is generally recognised [15].

During an acoustical event, the stapes in the middle ear transmits the vibrational movement of the tympanic membrane to the perilymph in the scala vestibuli via the fenestra ovalis. The motion of perilymph in turn vibrates the endolymph in the scala media, the perilymph in the scala tympani, the structure of the basilar membrane and organ of Corti, thus causing movement of the stereocilia located at the tips of the hair cells. The kinetic deflection of the stereocilia opens mechanically gated ion channels through the cilia membrane and generates receptor potentials, the strength of which depend on the orientation of the stereocilia array relative to the direction of fluid particle velocity. The endolymph in the scala media is rich in potassium salts and has a high positive charge from 80 to 120 mV [24] whilst the hair cells have a negative potential of around - 50 mV. This electrochemical gradient between the hair cells and endolymph encourages an influx of potassium ions into the negatively charged cell body. Once the cell membrane becomes sufficiently depolarised, the cell generates a brief electrical signal or non-spiking receptor potential [25]. Both inner and outer hair cells respond to sound by generating potentials [24,26]. Cochlea hair cells are also known to possess a specialised synapse and a reserve of readily releasable vesicles [27] that allows for a rapid and long lasting response of the cells when appropriately stimulated. However, OHCs are generally considered to be an effector or motor cell [28], as they receive little if any afferent innervation [29]. OHCs are also known to display electromotile properties evoked by depolarization of the cell [25,30] as part of a feedback system that contributes to both frequency selectivity and hearing sensitivity, by amplifying basilar membrane motion [28,31]. The middle ear of the Mongolian gerbil (M. unguiculatus) is voluminous, with relatively thin bone covering the cochlea and brain [32], facilitating studies of both the ear and auditory cortex [33]. Interestingly, the range of frequencies audible to M. unguiculatus is comparatively wide, and covers a bandwidth of between 0.1 kHz to 60.0 kHz, with low frequencies being comparable to man [34]

Various techniques have been used to reveal internal and external hair cell structures; Spoendlin [29] and Sato et al. [35] used a light microscope to examine the innervating nerve distributions and hair cells in the cochlea of the cat, and Wever et al. [36] used a similar methodology for examining the cochlea from the bottlenose dolphin (Tursiops truncates). Richter et al. [37] imaged soft tissue in the gerbil organ of Corti to micrometer resolution using hard X-rays. In addition, Thorne et al. [38] used Magnetic Resonance Microscopy (MRM) to generate three dimensional reconstructions of the endolymphatic and perilymphatic fluid spaces in the human, guinea pig, bat, rat, mouse, and gerbil. Lenoir et al. [39] and Ma et al. [40] used the Scanning Electron Microscope (SEM) to study maturation of the rat cochlea, whilst Kopke et al. [41] and Lovell et al. [42] used the SEM to study surface detail of the hair cells in the non mammalian vertebrate vestibule. Internal cellular structures were studied using Transmission Electron Microscopy (TEM), in accordance with Tucker et al. [6] and Ma et al. [40] who used the TEM to study the hair cells in the cochlea of the mouse. Tolomeo and Holley [7] used the TEM to study the pillar cells in the guinea pig and Souter et al. [43] used both SEM and TEM to study postnatal maturation of the organ of Corti in gerbils. Plassmann et al. [44] studied the cochlea of 5 different gerbil species using cochlear microphonic recordings, serial sections and computerized quantitative reconstructions of the cochleae and their specific morphological structures. Tange et al. [45] used the Scanning Electron Microscope (SEM) to quantify vancomycin-induced cochlear damage in M. unguiculatus.

SEM preparation methodology

All experiments were approved by the animal ethics committee of the Land Sachsen-Anhalt (No. 42502-2-825 IfN MD), in accordance with the regulations of the German Federal Law on Care and Use of Laboratory Animals, and with the European Communities Council Directive of 24 November 1986 (86/609/EEC). The heads from 6 mature gerbils were removed immediately following an overdose of ketamine, then divided medially and a section of the brainstem, auditory nerve and associated arterial system were excised from the cranium, together with the periotic bone containing the inner ear and placed in chilled fixative (3.5% glutaraldehyde in Sodium Cacodylate buffer (0.1 M pH 7.2)) within around 4 minutes. After two hours in the fixative, the ears were washed in buffer and the periotic mounted in a clear Perspex clamp placed in Petri dish, then filled with buffer until the cochlea was completely covered to a depth of 1mm. The Petri dish was fixed to a rotating vice and x,y slide and mounted on a metal table bearing a binocular microscope and Dremmel drill with a stand and micro burr tool (tip diameter 300 μm). Encapsulating bone was carefully removed from the continuously submerged cochlea using the micro burr, fine forceps and a needle probe. On completion, the cochlea was dehydrated through a graded ethanol series ranging from 35% through 50%, 70% and 90% to absolute ethanol, prior to desiccation using the critical point drying method used by Lovell et al. [46]. Fully desiccated end organs were subsequently mounted on a specimen stub using either a carbon tab or silver dag, and coated with c. 8 nm of gold in an Emitech K 550 sputter coater (working at approximately 2 x 10^-6 Torr). The processed specimens were investigated and photographed using a FEI XL30 scanning electron microscope operated at 25 kV or 2 kV, respectively and a 15mm working distance and the images digitally stored on a PC.

TEM preparation methodology

The TEM preparation methodology employed in this study was based on techniques used by Lovell & Harper [46]. As with the SEM methodology, the heads of four mature male gerbils were divided medially and the periotic bone containing the inner ears from the left and right sides were removed and placed into chilled fixative for two hours. Following this, the periotic bone was trimmed and placed in a transparent micro-vice in a small dish filled with chilled buffer solution until it covered the cochlea to a depth of approximately 1mm. Removal of the bone encapsulating the cochlea was achieved manually under a binocular microscope using the aforementioned SEM methodology. Bone was removed a turn at a time and the canal cut into sections at each turn, then freed from the modiolus by running a micro-scalpel around the inner edge of the spiral lamina (the bony shelf projecting partially into the cochlea canal that supports the basilar membrane). After cutting, the canal was lifted from the periotic and placed into chilled Sodium Cacodylate buffer. The canal was then secondary fixed in a 1% Osmium Tetroxide in Sodium Cacodylate buffer for one hour for the first set of samples, and omitted for the second. The tissue was then rinsed twice in buffer, then dehydrated in an ethanol series. Once in 100% ethanol, the tissue was placed in increasing concentrations of Durcupan epoxy resin until it was fully infiltrated to 100%, in accordance with Glauert [47]. The resin and samples were then placed in a small plastic dish filled to a depth of approximately 2mm, with polymerisation achieved by placing the dish in a 70^°C oven for 48 hours. The resulting resin block was divided into smaller blocks each containing a section of cochlea, then sectioned using a Reichert- Jung Ultracut and a Diatome diamond knife. Ultra thin sections were picked up using 200 mesh thin bar copper grids or coated slot grids, and stained with a saturated ethanol solution of Uranyl Acetate. The fully processed images were taken using a LEO EM 912 transmission electron microscope and analysed using the ImageJ software.

General dissection

Partial removal of the bone covering the inner ear from M. unguiculatus reveals the cochlea encased in the dense auditory periotic bone, along with the eardrum and malleus. Further removal of the remaining bone covering the posterior part of the inner ear reveals the membranous labyrinth of the vestibule, which contains the saccule, utricle and semi-circular canals; each filled with endolymph. Six complete gerbil cochlea were imaged using the Scanning Electron Microscope and measured along the organ of Corti from the lower basal end to the upper tip (approximately 2.75 turns). The innervated organ of Corti had a mean length of 11.67 mm (min 11.44mm, max 11.91 mm) with a standard deviation of 0.208.

Hair cell morphology

The rodent cochlea (in common with other mammals) has two types of hair cell, Inner Hair Cells (IHCs) and Outer Hair Cells (OHCs), orientated in a highly ordered arrangement toward the outer wall of the cochlea. Inner hair cells occur in a single row around the inner edge of the lamina, whilst the outer hair cells are arranged in a band of three rows. Each of the OHCs bears up to 60 stereocilia arranged in a W formed from three to four consecutively shorter rows. The approximate hair cell count along the entire length of the organ of Corti was calculated to be 1260 inner hair cells to 5000 outer hair cells, configured in four highly ordered rows with a ratio of approximately 1 inner hair cell to between 3 and 4 outer hair cells. Figure 1a to Figure 1d present micrographs of the cochlea at the second and basal turns.

Figure 1: Low magnification SEM micrograph of the cochlea at the second turn. cb. cell body, n. nucleus, slm. spiral limbus. 1b: SEM micrograph of the stereocilia from the single row of IHCs. 1c: SEM micrograph of the stereocilia from an OHC from the second turn of the cochlea. 1d: SEM micrograph of an IHC showing rows of stereocilia.

The IHCs in the cochlea have a cell body at the apex of the helicotrema of approximately 37.5 μm (SD 1.284, min 35.6, max 38.8) in length and approximately 9.1 μm (SD 1.005, min 7.6, max 10.0) in width, and containing a nucleus with a mean diameter of 3.5 μm (SD 0.438, min 2.96, max 4.1). The cell body lies at almost 90° from the OHCs, which are orientated almost vertically; though this angle can vary by as much as 45° from vertical. The mean length of the cuticular plate was 5.4 μm (SD 0.485, min 4.7, max 6.0) and positioned on the side of the inner hair cell, so the stereocilia protrude upward to a mean height of 10.0 μm (SD 0.642, min 8.8, max 10.5) and at a similar angle to the cilia from the outer hair cells. IHCs from the first turn of the cochlea are approximately 22.0 μm (SD 0.790, min 21.1, max 23.3 μm) in length and approximately 8.3 μm (SD 0.742, min 7.311, max 9.229 μm) in width, and containing a nucleus with a mean diameter of 3.8 μm (SD 0.316, min 3.327, max 4.207 μm). The mean length of the cuticular plate was 4.9 μm (SD 0.689, min 4.2, max 5.8 μm), with stereocilia protruding upward to a mean height of 5.6 μm (SD 0.407, min 5.0, max 6.2 μm). At the basal end of the second turn of the cochlea the IHCs are approximately 20.5 μm (SD 1.374, min 18.4, max 21.8) in length and approximately 8.2 μm (SD 1.201, min 6.1, max 9.3 μm) in width, and containing a nucleus with a mean diameter of 2.5 μm (SD 0.383, min 2.0, max 3.0 μm). However, it should be noted that these data may not include the maximum nucleus width at this location, as other sections from a similar sample had a mean width of 4.9 μm (SD 0.669, min 3.4, max 5.5 μm). The mean length of the cuticular plate was 4.2 μm (SD 0.161, min 4.0, max 4.4 μm), and stereocilia had a mean height of 2.7 μm (SD 0.283, min 2.5, max 3.2 μm). The body of the outer hair cells (OHCs) are slightly smaller than the IHCs, being approximately 31.8 μm (SD 3.082, min 26.5, max 35.5 μm) in length and approximately 6.9 μm (SD 0.818, min 5.9, max 8.3 μm) in width, and containing a nucleus with a mean diameter of 4.7 μm (SD 0.746, min 3.4, max 5.6 μm) at the apex. The uppermost portion of the cell contains the cuticular plate; in the outer hair cells it is approximately 4.8 μm across, supporting around 60 stereocilia and is in close proximity to the outer pillar cell. The base of the OHCs are connected to Deiters’ cells below which lies the terminal end of the outer pillar cells, which extend upward to the IHCs and roof of the tunnel of Corti. Each OHC is separated from its neighbour by a region void of any obvious structure approximately 10 μm wide, containing perilymph from the scala tympani.

Pillar Cells

Figure 2a through Figure 2e present ultra thin sections of the pillar cells, viewed using the Transmission Electron Microscope (TEM). The thinnest part of the pillar cell body is approximately 2 μm in thickness and projects from the cuticular plate at the top of the OHCs (Figure 2a and Figure 2b), running just below the epithelial surface toward the single row of IHCs positioned approximately 25 μm from the outer hair cells (Figure 2d).

Figure 2: 2a TEM micrographs of the outer hair cell tip and connecting pillar cell from the second turn of the cochlea. Note; the striated myofibril like structures below the OHC tip are the pillar cells projecting from the two proximal hair cells. Bar = 2 μm. 2b: Micrograph of the junction between the tip of an outer hair cell and the connecting pillar cell. Bar = 0.5 μm. 2c: Micrograph of a transverse section through a pillar cell. Bar = 1 μm. 2d: Micrograph of an Outer Hair Cell from the first half turn of the cochlea. Bar = 10 μm. 2e: Micrograph of the basal inner pillar cell (pc.) and bony spiral lamina (bl.) containing peripheral fibres of the auditory nerve (pnf.), entering the foramina nervosa (fn.) located 10 to 15 μm below the distal end of the inner hair cell. cp. cuticular plate, J. junction, n. nucleus, st. stereocilia, sv. synaptic vesicles.

Transverse sectioning of the pillar cell (Figure 2c) revealing thin diameter actin and thicker diameter myosin filaments producing a cross striated architecture (lower structure in Figure 2a). As Figure 2d shows, the pillar cell thickens as it passes below the tip of the IHC, then becomes thinner again as it descends to the bony spiral lamina (Figure 2e). The lamina contains peripheral fibres from the auditory nerve, which are located between 10 to 15 μm below the distal end of the inner hair cell. Synaptic vesicles can be clearly seen above the peripheral fibres, which project toward the vesicles through the foramina nervosa. The terminal end or foot of the inner pillar cell can be seen in close proximity to the VIII nerve fibres and synaptic vesicles.

All mammalian cochleae appear to function according to the same basic principles; however, the effective frequency range differs between species [48]. For example, the range of audible frequencies is about 20 Hz to 16 kHz in the human cochlea and about 100 Hz to 60 kHz in M. unguiculatus [34]. In humans, the cochlea is coiled approximately two and one-half turns around the modiolus, and in M. unguiculatus it is coiled two and three-quarter turns. The diversity in the number of spiral turns in the cochlea and variation in the basilar membrane length between rodents can be seen in Table 1 [46]. The number of cochlea turns and the basilar membrane length from M. unguiculatus, (identified in this study) have also been included, along with the hearing frequency range as defined by Ryan [34].

Table 1: The spiral turns of the cochlea and basilar membrane length with the range of audible frequencies from a selection of rodents (adapted from Lovell and Harper [46]) and includes data for M. unguiculatus.

In this work, the structure of pillar cells is investigated, and striated banding within the cell body can be clearly seen in Figure 2a. Transverse sectioning shows this architecture has a repeating pattern of alternating thin diameter actin and thicker diameter myosin filaments [7,8]. The number of actin and myosin filaments within the pillar cell increase progressively with location in the cochlea [9-11]. Outer pillar cells express Cx26 while the inner pillar cells express Cx30; it is considered that these proteins form pathways between cells [37], and can be imaged using bright field, fluorescent and confocal fluorescent techniques. Genetic mutations in Connexin 26 (Cx26) and Connexin 30 (Cx30) have been linked to prelingual deafness in humans.

It is therefore proposed here, that it is not only the outer hair cells that are motile [25,30] but the pillar cells themselves have a structure that may responds sympathetically with the movement of the outer hair cells; thus pillar cell motion may resemble the movement of cilia or flagella. Rapid movement of pillar cells would further increase hair cell motility; in line with the findings of Waldeyer, Rhode & Geisler and Schick et al. [13-15]. The schematic diagram in Figure 3 has been annotated with arrows that suggest possible directions of motion during an acoustical event.

Figure 3: Schematic showing the relationship between the inner and outer hair cells and the pillar cells from M. unguiculatus. The arrows labelled A, B and C indicates the possible direction of movement caused by the motion of the pillar cells. bm. basilar membrane, dc. outer phalangeal (Deiters) cells, Hc. Hensen cells, IHC. Inner Hair Cell, ipc. inner pillar cell, iphal. inner phalangeal cells, OHC. Outer Hair Cell, opc. outer pillar cell, pha. phalanx of the outer phalangeal cells, sv. synaptic vesicles, VIII. auditory nerve fibres. Note: the apical region of the outer hair cells is in contact with the and the outer hair cells from the first row do not directly contact the outer pillar cell.

The cochlea from the mature Mongolian gerbil (M. unguiculatus)has a mean length of 11.67mm and spirals for approximately 2.75 turns, with a single row of approximately 1260 Inner hair cells and 5000 outer hair cells arranged in three ordered rows. The IHC dimensions range from approximately 38µm in length and 9µm in width at the tip of the cochlea to approximately 20µm in length and 8µm in width at the basal end. Stereocilia protrude upward to a mean height of 10.0µm at the tip down to around 2.7µm at the base and corresponds to the tonotopic organisation of the cochlea. OHCs are more standardised than the IHCs being approximately 32µm in length and 7µm in width at several locations around the spiral turns; however, the length gradient of stereocilia conforms with measurements from adjacent IHCs. Pillar cells project from the cuticular plate at the tip and from the base of the OHCs, and support both IHCs and OHCs as they traverse the organ of Corti to the tympanic lip of the spiral lamina. The histological examination of these cells reveals thin diameter actin and thicker diameter myosin filaments which may contribute to pillar cell motion.

Thanks to Anton Ilango from the Leibniz Institute for Neurobiology for kindly allowing the removal of the cochlea from gerbils during pathological preparation for an unrelated experiment, and the authors would also like to thank Michael Reppin from the Institute for Materials Technology for his help preparing the samples for SEM microscopy.