Ganglion and “Dendrite” Populations in EAS Ears

Abstract

Background/Aims

EAS technique combines electric and acoustic stimulation in the same ear and utilizes both low frequency acoustic hearing and electric stimulation of preserved neurons. We present data of ganglion cell and dendrite populations in ears from normal individuals and those suffered from adult-onset hereditary progressive hearing loss with various residual low tone hearing. Some of these were potential candidates for EAS surgery. The data may give us information about the neuro-anatomic situation in EAS ears.

Methods

Dendrites and ganglion cells were calculated and audio-cytocochleograms constructed. The temporal bones were from the collection at the House Ear Institute in Los Angeles, USA. Normal human anatomy, based on surgical specimens, is presented.

Results

IHCs and OHCs, supporting cells, ganglion cells and dendrites were preserved in the apical region. In the mid-frequency region, around 1 kHz, the OC with inner and outer hair cells were often conserved while in the lower basal turn, representing frequencies above 3 kHz, OC was atrophic and replaced by thin cells. Despite loss of hair cells and lamina fibers ganglion cells were present even after 28 years duration of deafness.

Conclusions

Conditions with profound SNHL with preserved low tone hearing may have several causes and the pathology may vary accordingly. In our patients with progressive adult-onset SNHL (amalgamated into “presbyacusis”) neurons were conserved even after long duration of deafness. These spiral ganglion cells may be excellent targets for electric stimulation using EAS technique.

Introduction

Combined electro-acoustic stimulation paradigm in the same ear, so-called EAS strategy, uses both acoustic and electric stimulation of residual auditory nerve structures (von Ilberg et al., 1999; Kiefer et al., 2002, 2005; Gantz and Turner, 2003; Skarzynski et al 2003, Gstoettner et al., 2004, 2006; James et al., 2006; Turner et al., 2008). With less invasive surgical techniques and shorter electrodes the fragile inner ear structures may be conserved. Preservation of residual hearing is now a goal in CI surgery that should always aim to limit intracochlear damage. As the success rate of hearing preservation increase, patients with more residual hearing may become candidates for EAS surgery (Skarzynski, et al., 2003). The technique was proposed already in 1994 by William House.

It is assumed that a key issue for a successful outcome using EAS technique may be the neural potential in the basal part of the cochlea where the electrode lies close to the high- and mid-frequency coding neurons. How important neurons really are for the results of CI in general is still under debate (Linthicum and Galey 1983, Linthicum et al. 1991, Nadol et al. 2001, Khan et al. 2005, Fayad and Linthicum et al. 2006, Nadol and Eddington 2006). Fewer neurons than earlier thought seem needed and if this also holds true for the EAS principle is not known. We still know little about the way ganglion cells are stimulated within the modiolus. The neurons are located near the perilymph electrolyte and currents spread easily along the interior of the cochlea and selective stimulation within small distances using conventional monopolar stimulation seems hard to conceive. We recently identified specially arranged connexin proteins in the human spiral ganglion (Liu and Rask-Andersen, unpublished observations). Their role is still unknown. Are electrical junctions present in man’s auditory nerve and if so how can these be related to nerve synchrony and oscillations and can electric stimulation such as in EAS work in combination with acoustic hearing to replace such possible functions?

There are relatively few studies focusing on cochlear histopathology related to ears with various degree of low tone preservation in different conditions. Hinojosa and Marion (1983) and also Schuknecht (1993) analyzed ears with “profound high-tone deafness including patients with several diagnoses. Here we present data of ganglion cell and dendrite populations in ears from individuals who suffered from adult-onset hereditary progressive hearing loss with some residual low tone hearing. Some of these patients had low tone hearing making them candidates for EAS surgery. We also present information about the normal structure and innervation of the human cochlea.

Material and Methods

Histopathology

Four female patients were analysed with adult onset hereditary progressive hearing loss. Their ages were according to the years of follow up or duration of deafness (age-duration) were case 1) 84 - 3Yr, case 2) 78 - 4Yr, case 3) 65 - 28Yr and case 4) 65 - 10Yr. Dendrites and ganglion cells were calculated in all four cases and audio-cytocochleograms constructed according to Guild (1921) and by Schuknecht (1953). All had some residual hearing at low frequencies as shown by the audiograms. The temporal bones were from the collection at the House Ear Institute in Los Angeles, USA.

Transmission electron microscopy

Ultrastructural findings described here were in part published by Tylstedt et al. (1996 a, b). Innervation of the different turns were analysed and the spiral ganglion from the lower basal, upper basal, lower middle and upper middle region were sectioned separately. Montages of the ganglion area were imaged at x 1000 and graphical reconstructions were formulated. Special attention was given to the structural relationship between the Type I ganglion cells. Thin sections were viewed in a JEOL 100 SX electron microscope. The technique for SEM processing is described elsewhere (Glueckert et al. 2005). The study was approved by the local ethics committee (no. 99398, 22/9 1999, no. C254/4, no C45/7 2007) and patient consent was obtained.

Immunohistochemistry

The study is based on human cochleae taken out at surgery during a transotic approach to remove a petroclival meningioma The cochlea was dissected out and placed in 4% buffered paraformaldehyde (PFA) in phosphate buffered saline. Sections of the cochleae were embedded and rapidly frozen and cryostat sectioned at 8–10 μm. Antibodies against Cx26, 29, 30, 31, 32, 36 and 43, Trk A, B and C receptors, parvalbumin (Pv), peripherin, class III β-tubulin and neurofilament 160 were used for immunohistochemistry of the sections in combinations. Sections were subjected to the reaction to Alexa Fluor 488 and 555 (Molecular probes) conjugated secondary antibodies.

Confocal, fluorescent and bright field imaging

Bright field and fluorescent images were obtained using an inverted fluorescent microscope (Nikon TE2000, Japan) equipped with fluorescence unit and a Digital camera with a spot digital camera with three filters (for emission spectra at 358, 461, 555 nm). For confocal microscopy we used a Nikon TE300 microscope equipped with laser imaging system using three different filters.

Plastic molding of cochlea

In order to analyse anatomic frequency maps for short electrodes the relative length of the first turn of the human cochlea was investigated using plastic castings of 95 human inner ears. Silicone and polyester resin material was used in this investigation (Erixon et al. 2008). We used the mid-point of the long diameter of the round window as reference and starting point for measuring the length of the cochlea since we believe that RW application may be the optimal technique used in future implantations. A line was drawn through the central axis of the cochlea to a distant point of the first turn and at right angles to this line through the axis of the cochlea dividing each turn of the cochlea into quadrants.

Results

In all four cases IHCs, OHCs, supporting cells, ganglion cells and spiral lamina nerve fibers were well conserved in the apical region (Figure 1). In the upper basal and middle turn (mid-frequency region) the OC with IHCs and OHCs and lamina fibers were also generally preserved while in the lower basal turn, at areas representing frequencies above 3 kHz according to Greenwood place/frequency map, OC had undergone atrophy and was replaced by a thin cell layer (Figure 2). There were no lamina fibers but ganglion cells were present with a maximal loss of 60% in cases 2 and 4, (duration of deafness 4 and 28 years) and in cases 1 and 3 there was a minor loss of ganglion cells at corresponding region (duration of deafness 3 and 10 years respectively) (Figure 1). Light microscopy showed that spiral ganglion cells at the sites of hearing loss were often arranged in cluster and physically interacted with each other. The perikarya were surrounded by a thin satellite cell (Figure 2B, 3B inset). There were no peripheral axons emerging in the direction of the organ of Corti (Figure 3B). Central axons were myelinated and had a normal appearance. Thus, remaining type I cells were unipolar in type and most of them were located together even though a few individually sited cells were also seen (Figure 3B inset). Several free cells appeared around the neural cell bodies.

Figure 1.

Audio-cytocochleograms of cases with adult-onset progressive sensorineural deafness with various preservation of low tone hearing. Duration of deafness varied from 3 to 28 years. For further information see M&M.

Figure 2.

Histological section of cochlea from case 4. In the lower basal turn (approx. 4 kHz area) OC is atrophic and there are no lamina fibers (a, x400). Spiral ganglion cells are preserved (b, x200). In the mid- and apical portions of the cochlea the organ of Corti (c, x400), lamina fibers and ganglion cells (d, x200) are fairly well preserved.

Figure 3.

Histological sections of the OC at the apex (top, x 400 approx. 250 Hz) and basal turn (bottom, x 400 approx. 4kHz) in a patient with progressive sensorineural hearing loss. Atrophy of the OC is associated with loss of lamina fibers but not with loss of ganglion cells (c, x 400 4 kHz). Osmium and haematoxylin.

A scanning electron micrograph (SEM) of a normal, optimally fixed, hemi-sectioned human cochlea corresponding to the level of section (figure 2) can be seen in Figure 4. The position of a short electrode and how it occupies the scala tympani space in the basal turn (360 degrees) is delineated. Its relationship to the basilar membrane, organ of Corti and spiral ganglion cells can be observed (Figure 4, 5). The normal dendrite architecture with typical arborization is seen in the basal turn in an osmium stained human specimen (Figure 5B). A graph shows the region of maximal innervation density representing upper basal and lower middle turns (Figure 5C).

Figure 4.

Scanning electron microscopy (SEM x24) of a normal human cochlea corresponding to the level of sectioning demonstrated in figure 2 (mid-modiolar section, inset right). The putative location of an EAS electrode with a tip diameter of 0.4 mm is shown (red). Circles show the anatomic location of the spiral ganglion (a; 1–2 kHz, b; 4 kHz, c; < 1 kHz). Inset (left) shows the introduction of an EAS electrode through the round window.

Figure 5.

A) Plastic corrosion cast of a left human inner ear. Solid line demarcates one turn (arrow). This represents approx. 20–24 mm distance from the mid-point of the round window. B) Surface preparation of a normal human cochlea. The distribution of peripheral processes can be seen up to one turn. (osmium tetroxide, courtesy B. Engström). C) Graph showing innervation density of the normal human cochlea. (Spoendlin and Schrott 1990). Dashed line shows EAS electrode reaching up one turn.

The distribution of perikarya at various locations is shown in figure 6 and 7. Even though the perikarya are not evenly distributed along Rosenthal’s canal it is obvious that there are normally much fewer cells appearing on radial sections in the basal region of the cochlea. From these reconstructions one can also see that perikarya often share the same satellite cells and that they often interact physically with each other, especially near the apex where the cell bodies become concentrated in a terminal bulb-like structure (Figure 6A, B). TEM shows that these cells are unmyelinated and surrounded by a thin “satellite” cell. Serial thin sections and 3-dimensional reconstructions of physically interacting perikarya show that junction-like specializations extend between neurons plasmalemma (Figure 6B). Immunofluoresence at these areas shows surprisingly an expression of connexin 30 protein (Figure 6B, unpublished observations). Such expression may be unique to humans and has not been observed in animals studied so far in our laboratory.

Figure 6.

Figure 6A. Computer-based 3-D reconstruction of the human spiral ganglion (green), dendrites (yellow) and organ of Corti (blue). Arrows shows the level of one turn (lower) and the terminal bulb. The ganglion traverses to the level of the mid-turn of the cochlea and neural perikarya end the terminal bulb-like innervate hair cells in the apical turn (from Ariyasu et al. 1989).

Figure 6B. Top left. Graphic reconstruction of a composite transmission electron micrograph (TEM) showing perikarya in the terminal bulb of a normal human spiral ganglion. Lower left. Corresponding light micrograph showing densely packed neural perikarya. Top right. Immunofluoresence shows expression of connexion 30 in these perikarya. Bottom right. 3-D reconstruction of serial thin sections shows that human perikarya form intercellular junctions (from Tylstedt and Rask-Andersen 2001). These junctions may give trophic supply to neurons that have lost their contact with their hair cells. It could explain their slow degeneration in the human cochlea.

Figure 7.

Computer-based graphic reconstructions based on composite transmission electron micrographs of transverse sections of the human spiral ganglion. A) Basal portion corresponding to an anatomic frequency of 10 kHz. B) Half basal turn corresponding to frequency area of 4 kHz. C) One turn corresponding to frequency area of approximately 1 kHz. D) End of terminal bulb at transition zone of “dendrites” supplying the apical turn. This corresponds to frequencies below 250 Hz. Blue cells represent Type II afferents innervating outer hair cells. Dark brown cells share satellite cells. Inset (top right) shows frequency distribution of the nerve fascicles sectioned serially and traced to the basilar membrane. A majority of the un-myelinated fibers are efferent.

The human cochleae varied in size and shape considerably resulting in different insertion depths, angles and place/frequency maps of the introduced CI electrodes. The mean length of the first turn (quadrant 1 to 4) was 22.6 mm with a range from 20.3 to 24.3 mm, representing 53% of the total length (table). According to an anatomic frequency map an electrode reaching one turn covers approximately frequencies down to around 1 kHz.

Table 1.

Outer wall length (mm)	Mean	Range	SD	n
Half diameter of the RW	1.1	0.3–1.6	0.21	65
First half of first turn	13.5	12.1–15.0	0.73	67
First turn (quadrant 1–4)	22.6	20.3–24.3	0.83	65
Second turn (quadrant 5–8)	12.4	10.7–13.3	0.63	63
Third turn (quadrant 9 to11(12)	6.1	1.5–8.2	1.40	58
Total length*	42.0	38.6–45.6	1.96	58

^*The total length of the outer wall excluding the basal half of the round window (RW). SD; standard deviation. n; number of specimens.

Discussion

We found that in progressive adult-onset profound SNHL, with various preservation of low tone hearing, the organ of Corti had undergone atrophy with degeneration of inner and outer hair cells in the lower basal turn. This was associated with a total or near total loss of lamina fibers. The spiral ganglion cells however were conserved to various degrees even after 28 years duration of deafness. These cells may be excellent targets for EAS surgery.

Conditions with profound SNHL and preserved low tone hearing may have several causes and the pathology may vary accordingly. Our results are in accordance with Hinojosa and Marion (1983) who correlated the state of the acoustic ganglion with that of the organ of Corti and of the peripheral fibers in eight patients with profound hearing loss defined as an average loss of 90 dB or greater for the hearing threshold levels at 2000, 4000, and 8000 Hz only. It included diagnoses such as presbyacusis (6), ototoxicity (6), Meniere’s disease (1) and otosclerosis (1). Cochlear reconstruction demonstrated the least amount of sensorineural injury among all bones examined (Hinojosa and Lindsey 1980), with the highest number of cells remaining in the spiral ganglion. In a 77-year-old man with residual low tone hearing audiogram revealed an advanced sensorineural hearing loss at 1000, 2000, 4000, and 8000 Hz. Ganglion cell counts were only mildy reduced in all turns. Cochlear reconstruction demonstrated total loss of the organ of Corti in the first 11 mm of the basal coil, with variable preservation of the hair cells throughout the remainder of the cochlea. Peripheral cochlear nerve fibers were also absent in the basal turn, but increased steadily into the apex.

Injury to the organ of Corti seems to result in retrograde degeneration of peripheral cochlear nerve fibers. Otte et al. (1978) stated that loss of the spiral ganglion cells seems to parallel the extent of injury of supporting cells rather than to hair cells. According to studies by Spoendlin (1975) based on cats, loss of inner hair cells resulted in retrograde degeneration of the cochlear neurons. Hinojosa and Marion (1983) however found that the degree of degeneration of the organ of Corti in humans including the supporting cells does not correlate with the number of remaining lamina fibers or spiral ganglion cells. They assumed that factors other than the state of the organ of Corti, supporting cells, and peripheral fibers have an effect on the extent of ganglion cell loss. Thus, contrary to what happens in animals, loss of hair cells and even loss of the entire organ of Corti, does not lead to complete neuron loss. This disparity is of fundamental importance for CI as well as EAS. The reason for the slow degeneration remains puzzling. We speculated that neural perikarya may obtain local trophic supply from each other through their physical interaction. It is made possible through their lack of myelin (Rask-Andersen and Tylstedt 1997, Tylsedt et al. 1997). This interaction is noticeable in the apical region where even specific junctions occur between bordering cells (Figure 6B) (Rask-Andersen et al. 2000, Tylstedt and Rask-Andersen 2001). However, histology from cases 1–4 shows that some type I cells have no physical contact with each other so this explanation cannot fully explain the degree of preservation. Interestingly, human perikarya are enclosed by a different type of cell here named “satellite” cell that contains no myelin. Degeneration of the peripheral axons together with their myelin containing Schwann cells starts at the periphery and may seize at the axon hillock where the “satellite” cells enclose the perikaryon. Satellite cells may not be affected by the degeneration. In the situation where the peripheral axon and the cell body are myelinated the degeneration process may also involve the cells surrounding the perikarya. This cell is of paramount importance for the protection, function and survival of the neuron. It produces survival factors such as glia cell-line-derived neurotrophic factor (GDNF). We found tyrosine kinas receptor B (Trk B) selectively expressed on perikarya; a target for the neurotrophin BDNF (Liu and Rask-Andersen 2008 unpublished observations). One way to further prevent degradation of neurons could therefore be to add small amounts of these growth factors via the implant using drug delivery systems.

If preservation of peripheral processes improves electrical stimulation or maintains the integrity of spiral ganglion cells is not known with certainty. Reports of animal studies have indicated that it is the cell bodies or possibly central axons that are being stimulated by the implant (Clopton et al. 1980) and human implanted temporal bones suggest that performance is unrelated to the percent of remaining peripheral processes and spiral ganglion cells (Fayad and Linthicum 2006). The degeneration of lamina fibers seems to occur more promptly following inner hair cell loss which is thought to head fiber degeneration. This deterioration then seems to decelerate once reaching the cell bodies. However as stated earlier the conditions may be highly varying and maintained lamina fibers can also be noticed despite total loss of inner hair cells (Hinojosa and Marion 1980, Teufert et al., 2006).

The apical part of the cochlea is known to be more resistant to degeneration with conservation of sensory structures. This biological principle seems to remain through animal series and may have evolutionary background. Rosenthal’s canal is well defined only in the first turn of the human cochlea where neuron’s location matches innervated hair cells. More apically neurons coalesce into a less well defined bony canal where neurons are tonotopically compressed. The spiral ganglion in humans goes only as high as the middle segment where dendrites spread out to innervate the third turn (Ariyasu et al., 1989, figure 6A, B, 8). Electrodes reaching the second turn may therefore stimulate neurons coding for lower frequencies than represented by their location. This may imply that selective stimulation of neurons coding for particular frequencies is more intricate in the apical region than in the basal region. This condition may have bearing on EAS strategy since the use of acoustic stimulation in the low frequency region may provide more temporal fine structure hard to obtain through electric stimulation.

The optimal length and design of the implanted electrodes as well as the most suitable surgery for the EAS patients are still a matter of debate. Different short electrodes ranging from 6 to 10 to 21–24 mm have been constructed (von Ilberg et al., Kiefer et al., Skarzynski et al. 2004, Gantz et al. 2005, Adunka et al 2005, Gstoettner et al. 2006). The original 6 mm device was lengthened to 10 mm with the most apical electrode at approximately 2,500 to 3,000 Hz according to the Greenwood place frequency map. The tip of the electrode curves into the ascending segment but does not extend to the upper basal turn of the cochlea. The most innervated region of the human cochlea is the upper basal and lower middle turns (Figure 5C, Spoendlin and Schrott-Fischer 1990). An electrode reaching one turn is close to the most densely innervated areas which may be advantageous in case of inadvertent deafness caused by EAS surgery. We found that at one turn (approximately 1 kHz) both hair cells and lamina fibers were well preserved but apparently non-functional.

Cochlear inner and outer wall length differ greatly and an electrode runs deeper if placed against the modiolus than the outer wall. This owes the larger diameter of the first turn and the modiolus. The large variations in cochlear anatomy may favor the idea of using more individually shaped electrodes. Since the medial wall of the cochlea is fragile it would seem imperative not to exert any physical pressure on this structure (Rask-Andersen et al., 2006). Our studies of the human cochlea show that each person’s cochlea is individually shaped (Erixon et al. 2009). Variations in size and shape result in different insertion depths, angles and place/frequency maps of the introduced CI electrodes (table). When performing CI surgery in patients intended for combined acoustic and electric hearing it is necessary to consider these anatomic variations. Based on measurements of 65 plastic molds of the human cochlea, the first turn extended 20.3–24.3 mm with an average of 22.6 mm. This represents the length of an electrode reaching one turn to approximately 1 kHz. These results are in accordance with those by Adunka et al. (2005) and Gstoettner et al. (2006) who found that a 360-degree insertion of the array entering this region, which defines the end of electric stimulation and beginning of acoustic stimulation, corresponds to 18–24 mm. This depth was found to provide good cochlear implant performance for both combined EAS and electrical stimulation alone in case of loss of residual hearing without the need for re-implantation. A shallow insertion reduces the risk of damage to apical cochlear structures while a deep insertion of the array may improve cochlear implant performance in case residual hearing is lost. Another concern is the potential deterioration of residual low frequency hearing over time.

EAS technologies aiming at preserving acoustic hearing and provide electric stimulation of residual ganglion cell bodies may improve esthetics of sound including speech comprehension in noise and multi-talker backgrounds as well as music perception (Turner et al., 2004, 2008, Kong et al., 2005). In a recent study of hybrid patients they were found to score nearly as accurate as normals for melody recognition, whereas long-electrode patients performed very poorly (Gfeller et al. 2006). The question arise how many residual ganglion cells in the basal turn that are necessary to provide benefits from such a strategy. Are they necessary or can neurons higher up in the cochlea compensate for their loss? How can we assess preoperatively the amount of residual ganglion cells still present in the lower turn? Humans have approximately 30 000 neurons (Otte et al., 1978, Rasmussen 1940) and almost 50% of these are located in the first turn. At present it is not possible to predict the neural potential of the cochlea and its different parts. Despite a profound hearing loss we know the numbers of surviving cells may be considerable. The ganglion cell populations have been found to be largest in ears deafened by sudden deafness, Meniere’s disease, and ototoxic drugs; somewhat less in vascular occlusion, temporal bone fracture, otosclerosis, and cochlear dysplasia and least in measles, bacterial labyrinthitis and congenital syphilis (Hinojosa and Lindsey 1980). In addition, fewer ganglion cells than previously thought seem necessary to achieve useful auditory sensation from electrical stimulation even as few as 10% of the normal number (Linthicum and Galey 1983, Linthicum et al. 1991, Nadol et al. 2001, Fayad and Linthicum et al. 2006). Recent data from the FDA Iowa/Nucleus Hybrid clinical trial suggests that those with more than 35 years of severe-profound hearing loss above 2000 Hz often did not do well with the added electric stimulation, telling that there were not enough viable ganglion cells in the base of the cochlea to take advantage of a 10-mm electrode (Turner et al., 2008). Linthicum and Fayad (2007) demonstrated though a case with a woman who had a life time, 89 years of deafness, with no hair or supporting cells with 23,000 ganglion cells.

William House already in 1994 anticipated that a rising number of patients with different kinds of sensorineural HL will benefit from “use of complementary assistive systems one acoustic and one electric”. We see a new era of implantation emerging. Innovative strategies is necessary to design even more sophisticated electrodes not only for saving residual hearing but also for more selective stimulation of the neural components of the human cochlea.