- Email: [email protected]
Distribution and time course of hair cell regeneration in the pigeon utricle
Hearing Research 133 (1999) 17^26
Distribution and time course of hair cell regeneration in the pigeon utricle Bradford J. Dye, Thomas C. Frank, Shawn D. Newlands, J. David Dickman * University of Mississippi Medical Center, Department of Surgery, Division of Otolaryngology, 2500 North State Street, Jackson, MS 39216-4505, USA Received 16 November 1998; received in revised form 27 February 1999; accepted 3 March 1999
Abstract Vestibular and cochlear regeneration following ototoxic insult from aminoglycoside antibiotics has been well documented, particularly in birds. In the present study, intraotic application of a 2 mg streptomycin paste was used to achieve complete vestibular hair cell destruction in pigeons (Columba livia) while preserving regenerative ability. Scanning electron microscopy was used to quantify hair cell density longitudinally during regeneration in three different utricular macula locations, including the striola, central and peripheral regions. The utricular epithelium was void of stereocilia (indicating hair cell loss) at 4 days after intraotic treatment with streptomycin. At 2 weeks the stereocilia began to appear randomly and mostly in an immature form. However, when present most kinocilia were polarized toward the developing striola. Initially, regeneration occurred more rapidly in the central and peripheral regions of the utricle as compared to the striola. As regeneration proceeded from 2 to 12 weeks, hair cell density in the striola region equaled the density noted in the central and peripheral regions. At 24 weeks, hair cell density of the central and peripheral regions was equal to normal values, however the striola region had a slightly greater hair cell density than that observed for normal animals. ß 1999 Elsevier Science B.V. All rights reserved. Key words: Regeneration; Vestibular ; Hair cell; Ototoxicity
1. Introduction The ototoxicity of aminoglycoside antibiotics to the vestibular and auditory systems has been known for decades and well documented by studies in numerous species (Berg, 1951; Wersa«ll and Hawkins, 1962). More recent research has demonstrated the regenerative ability of vestibular and auditory hair cells following administration of aminoglycoside antibiotics, particularly in birds (Cruz et al., 1987; Lippe et al., 1991; Weisleder and Rubel, 1992). For the vestibular receptors, the majority of these studies have relied on intramuscular injections of aminoglycosides in young animals with an emphasis on hair cell regeneration in the ampullary organs. However, systemic administration of streptomycin often failed to completely eliminate all of the hair cells in the vestibular end organs, particularly the receptors of the maculae, even when dosages were suf-
* Corresponding author. Tel.: +1 (601) 984 5090; Fax: +1 (601) 984 5085; E-mail: [email protected]
¢cient to produce high mortality rates (Carey et al., 1996 ; Forge et al., 1993; Frank et al., 1998; Weisleder and Rubel, 1993). Omura et al. (1989) compared intralabyrinthine, intraperitoneal, and intramuscular injections of gentamicin in bullfrogs and found that intralabyrinthine injections resulted in the greatest hair cell loss for the posterior ampulla, but no documentation of otolith organ e¡ects were provided. Carranza et al. (1997) recently described a technique of direct intraotic administration of gentamicin soaked pledgets in the bullfrog which resulted in complete ampullary hair cell destruction as demonstrated by light microscopy; however, utricular hair cell destruction was not discussed. Baird et al. (1993) examined hair cell regeneration in the utricle and saccule of bullfrogs, while Lombarte et al. (1993) studied regeneration in the utricle and lagena of a teleost ¢sh, however signi¢cant numbers of hair cells in the extrastriola regions were unaffected by aminoglycoside treatment in both investigations. With these various methods of ototoxic agent administration resulting in non-uniform hair cell loss that occurs over a period of days, quantitative investi-
0378-5955 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 5 9 5 5 ( 9 9 ) 0 0 0 4 6 - 5
HEARES 3228 17-6-99
18
B.J. Dye et al. / Hearing Research 133 (1999) 17^26
gations of hair cell regeneration become di¤cult. In addition, it is not possible to formulate with certainty the time course or distribution of receptor cell regeneration without complete hair cell loss occurring at a discrete point in time. In a previous study (Frank et al., 1998), we demonstrated that a single direct application of streptomycin sulfate into the perilymphatic space was su¤cient to produce complete vestibular hair cell loss. With one intralabyrinthine application, a minimal dose of 2 mg streptomycin was required to produce a reliable complete destruction of all vestibular hair cells in both the ampullae and maculae. The regenerative capacity of the hair cells was preserved, although the nature of the regeneration pattern for the vestibular hair cells was not extensively examined. In the present investigation, we utilized the advantages of intralabyrinthine application, including single dose administration, complete hair cell loss, and discrete point in time of ototoxic initiation to study regeneration in the otolith organs of adult birds. Of particular interest is the time course and distribution of receptor cell regeneration in the utricle, which has received little attention in regeneration studies. The results show that regeneration of utricular hair cells occurs in di¡erent regions of the adult pigeon utricle at di¡erent times following ototoxic insult. 2. Materials and methods 2.1. Animals The experiments were conducted in adult male pigeons (Columba livia), that ranged in age between 1 and 3 years. All animals were housed in the Institution Laboratory Animal Facility. The methods of study were approved by the University of Mississippi Medical Center Animal Care and Use Committee and all animal procedures were performed in accordance with the Institution Animal Care and Use Committee guidelines. 2.2. Surgical preparation Prior to surgery, the streptomycin pellets were prepared. The pellets were created by adding 0.1 cc sterile saline to 2 mg of streptomycin sulfate powder (Sigma), then mixing into a slurry paste and storing at 311³C. For intraotic administration of the streptomycin, the animals were anesthetized using intravenous pentobarbital sodium (16 mg/kg) and intramuscular ketamine (20 mg/kg). Ampicillin (37.5 mg) was given intramuscularly prior to surgery. A midline skin incision provided exposure to the mastoid cortex which allowed
visualization of the horizontal canal through the thin bone layer. The cortex was removed using a rectangular shaped bone £ap thus exposing the bony labyrinth. Using a surgical drill, the vestibule was opened (0.35 mm diameter hole) which allowed for careful placement of the 2 mg streptomycin chips into the perilymphatic space. A temporalis muscle plug was placed over the surgically created ¢stula. This was then reinforced with gelfoam and bone wax. Next, the bone £ap was placed in its original position. The procedure was repeated on the opposite side and the incisions closed. A single dose of butorphanol (0.05 mg/kg) was given postoperatively for discomfort. On the day following the surgery, the birds were begun on a hand-fed diet of chick formula (Kaytee) for 3 days. Following this, the birds were allowed to feed ad libitum. The animals were monitored for weight loss, and hand feeding was resumed as needed to prevent dehydration and starvation. A second control group of pigeons was also used, in which an identical surgical procedure was performed, except that a similar volume of normal saline (frozen chips) was placed into the perilymphatic space. 2.3. Experimental procedure Several groups of animals were used. A normal group of animals (n = 5) received no antibiotic treatment or surgical procedure, while a saline group (n = 2) received only a saline application into the labyrinth. All treated animals that received a single streptomycin application were placed into di¡erent groups based upon survival time following ototoxic treatment. The treated groups included 4 days (n = 3), 2 weeks (n = 3), 4 weeks (n = 2), 6 weeks (n = 3), 12 weeks (n = 3), and 24 weeks (n = 2) survival. Following the speci¢ed survival time, each animal was killed and the vestibular receptor organs removed from the inner ear. 2.4. Histology The pigeons were anesthetized (sodium pentobarbital, 250 mg/kg) and the right horizontal and posterior semicircular canals were opened. The labyrinth was perfused with 5 cc of an aldehyde ¢xative consisting of 2% glutaraldehyde, 3% paraformaldehyde and 1% acrolein. Immediately following the labyrinthine ¢xation, the thoracic cavity was opened and a transcardiac perfusion was performed using 1000 cc of the same 2% paraformaldehyde, 3% glutaraldehyde solution (no acrolein). The head was removed and placed into the aldehyde ¢xative overnight. On the following day, the membranous labyrinths were removed from the inner ear and the utricle isolated by careful dissection. Only utricles from the right ear were utilized for scanning electron
HEARES 3228 17-6-99
B.J. Dye et al. / Hearing Research 133 (1999) 17^26
19
Fig. 1. Scanning electron micrographs of utricular macula in normal pigeons. For all SEM preparations, the otolith membrane and otoconia have been removed. A: Whole utricle with inset boxes showing regions for enlargement and quanti¢cation of hair cell counts. Small ¢ssures in the macula appeared as the tissue was £attened for viewing. Normally, the anterior end of the utricle is curled upward, to form an L-shaped surface. Anterior: leftward; lateral: upward. Scale bar = 100 Wm. B: Striola region with dashed line representing the hair cell polarization reversal point. Scale bar = 10 Wm. C: Central region, scale bar = 10 Wm. D: Peripheral region, scale bar = 10 Wm.
microscopy (SEM) analysis in all animals, while the left ear was used for a separate neural tracing investigation. The utricle was next prepared for SEM. First, the otoconia and otolith membranes were removed by manual dissection. Next, the tissue was placed into a 0.5% osmium and 0.1 M phosphate bu¡er solution for 2 h. After six serial washes of 10 min each in distilled water, the tissue was dehydrated. This was done with serial washes of 10 min each in 70%, 90% and 95% acetone followed by three washes of 15 min in 100% acetone. The tissue was then incubated in a 1:1 mixture of tetramethylsilane (TMS) and acetone for 45 min followed by an incubation of 45 min in a 3:1 mixture of TMS and acetone. The tissue was then incubated twice in 100% TMS for 1 h. Finally, the tissue was further dried in the oven for 30 min, allowing the TMS to sublimate (Dey et al., 1989). The tissue was then mounted onto aluminum studs and gold coated. The tissue was scanned and photographed using a JEOL TSM-3000 scanning electron microscope (15 kV). 2.5. Data analysis The entire right utricle (U75) and three separate re-
gions of the neuroepithelium (U1000) for each animal were photographed, in order to quantify the distribution and time course of regeneration. The three regions included the striola, the central (juxtastriola), and the peripheral regions, as shown in Fig. 1. The striola region was identi¢ed in the normal and regeneration animals by the reversal of the stereocilia bundle morphological polarization. The striola line was placed into the center of the view ¢eld for quanti¢cation. The central region was de¢ned at approximately the midpoint in the utricular epithelium, while the peripheral region was de¢ned by the medial border of the epithelium. Hair cells within a speci¢ed 9111 Wm2 rectangular area window for each region were counted, using the presence of a stereocilia bundle to indicate the existence of a hair cell. For newly regenerating hair cells, the identi¢cation of a kinocilium was required in order for the cell to be counted. The hair cell counts for each region within each treatment group were averaged. Statistical comparisons between regions and treatment groups were made using repeated measures ANOVA (Statistica). A single exponential curve was ¢t to the hair cell density data as a function of post-treatment time, using a leastsquares algorithm (SigmaPlot).
HEARES 3228 17-6-99
20
B.J. Dye et al. / Hearing Research 133 (1999) 17^26
Fig. 2. Scanning electron micrographs of utricular macula 4 days following treatment with streptomycin. A: Striola region with no stereocilia bundles present. B: Peripheral region with no stereocilia visible. Polygonal shaped supporting cells appear to comprise the entire neuroepithelium with a single microvillus protruding from the apical surface of each cell. Scale bars = 10 Wm.
3. Results 3.1. Normal utricle The normal pigeon utricle measures approximately 1.2 mm in length and 1.0 mm in width. In all ¢ve of the normal utricles examined, the macula surface was densely populated with hair cells, as shown in Fig. 1. As previously noted by many investigators (Flock, 1964 ; Wersa«ll, 1956; Lindeman, 1969) the hair cells are morphologically polarized according to the position of the kinocilium, which nearly always lay at the edge of the hair cell closest to the striola. In addition, the kinocilium of each cell was longer and thicker than the neighboring stereocilia. The striola was typically approxi-
mately C shaped as it curved along the outer lateral and anterior edges of the neuroepithelium. The striola line, as de¢ned by the reversal of hair cell morphological polarization, was not strictly uniform but generally lay within 100^175 Wm of the lateral/anterior border (Fig. 1A,B). The hair cells had a staircase array of stereocilia, with a lone larger kinocilium. Each hair cell was typically surrounded in a polygon pattern by 4^6 supporting cells, characterized by a single apical cilium. The density of hair cells was quanti¢ed for each of the three regions in ¢ve utricles from di¡erent animals. The central and peripheral regions had similar mean hair cell densities of 41 ( þ 6 S.D.) and 40 ( þ 8 S.D.) cells/1000 Wm2 , respectively. The striola region had fewer hair cells
HEARES 3228 17-6-99
B.J. Dye et al. / Hearing Research 133 (1999) 17^26
21
than the other two regions in all animals examined with a mean value of 34 ( þ 6 S.D.) cells/1000 Wm2 . However, this di¡erence was not signi¢cant (F(1,8) = 3.63, P 6 0.09). In two animals, normal saline instead of the ototoxic antibiotic was placed into the perilymphatic space using the same surgical technique. No observable hair cell destruction was noted at 4 days survival following the treatment. The mean hair cell density values in the control animals for the striola (27 þ 3 S.D.), central (37 þ 3 S.D.), and peripheral (38 þ 2 S.D.) regions were not signi¢cantly di¡erent from those observed for the normal animals. 3.2. E¡ects of streptomycin At 4 days after treatment with streptomycin, no visible stereocilia were observed in any of the six utricles examined, as illustrated in Fig. 2. For quantitative measures, three animals were photographed and counted, according to regions. The striola region could not be identi¢ed precisely, due to the lack of any visible stereocilia. Thus, the striola region was approximated by placing the view ¢eld at 100 Wm inside the lateral border, close to the average striola location in untreated animals. Still, no stereocilia were visible in any of the cell surfaces throughout the maculae. As shown in Fig. 2A for the striola region and in Fig. 2B for the peripheral region, support cells appear to comprise the entire neuroepithelium. Support cells had a characteristic polygonal shape and a single small cilium protruding from the apical surface. These lone cilia were also observable in the support cells of normal non-treated animals. Thus, at 4 days post-treatment following intraotic administration of streptomycin, a complete destruction of hair cells was observed. 3.3. Regeneration of utricular hair cells The time course of hair cell regeneration was studied over a 24 week recovery time continuum. As shown in Fig. 3, relatively few stereocilia bundles were present at early post-treatment periods, however over time the utricle neuroepithelium became repopulated to normal values. For example, at 2 weeks following streptomycin application, the number of visible hair cells was only a fraction of the normal pattern (compare Figs. 3A and 1A). However, the stereocilia from a number of cells had begun to appear distributed in a random fashion across the epithelial surface. When examined at higher magni¢cation, as shown in Fig. 4, some hair cells in the 2 week macula contained mature stereocilia bundles. However, many of the stereocilia bundles examined at 2 weeks were immature in appearance with only a few cilia and no distinguishable kinocilium. When present,
Fig. 3. Scanning electron micrographs of whole utricular maculae at di¡erent time intervals following treatment with streptomycin. A: Utricle at 2 weeks post-treatment time with few stereocilia bundles present. B: Utricle at 6 weeks post-treatment. Hair cell density has greatly increased, as evidenced by the presence of stereocilia bundles. Striola region is visibly less dense than central or peripheral regions. C: Utricle at 12 weeks following streptomycin treatment. Hair cell density continues to increase for all regions. Anterior: leftward; lateral: upward. Scale bars = 100 Wm for all micrographs.
most of the kinocilia were directionally oriented, even in the maculae of the 2 week post-treatment survival animals. For example, the morphological polarization of hair cells in the lateral margins of the epithelium were developing a de¢nable striola, as evidenced by oppositely directed stereocilia bundles, similar to that
HEARES 3228 17-6-99
22
B.J. Dye et al. / Hearing Research 133 (1999) 17^26
Fig. 4. Scanning electron micrographs of striola region at di¡erent time intervals following treatment with streptomycin. A, B, C, D: Striola region at 2, 4, 6, and 12 weeks post-treatment time, respectively. Dashed lines represent the morphological polarization reversal of hair cells, indicating the striola line. Scale bars = 10 Wm for all micrographs.
of normal animals. In the periphery (Fig. 5A), all of the kinocilium were directed toward the lateral or anterior margin (site of the developing striola). At the early time point of 2 weeks, the cell densities for the central and peripheral regions were similar, both with mean values of 10 ( þ 0.5 and þ 1.1 S.D.) cells/1000 Wm2 , respectively. Cell densities in the striola region were lower for all animals examined, with a mean 8 ( þ 1.8 S.D.) cells/1000 Wm2 . The cell densities for all regions at 2 weeks post-treatment were signi¢cantly less than those observed in normal utricles (striola: F(1,6) = 46.9, P 6 0.05; central: F(1,6) = 63.3, P 6 0.05 ; peripheral : F(1,6) = 41.9, P 6 0.05). At 4 weeks, the stereocilia appeared in increasing numbers and level of maturity, as shown in Figs. 4B and 5B. In fact, the density of cells had doubled over the 2 week mean values. The peripheral and central regions continued to be more densely populated with mature stereocilia than the striola region. In addition, the striola line became better de¢ned, as greater numbers of hair cells exhibited morphological polarizations (Fig. 4B). At 6 weeks, an increasing number of hair cells with
stereocilia were visible, as shown in the whole utricle example of Fig. 3B. The overall organization and distribution of the hair cells had an appearance that was more similar to that of the normal utricle. Still, a relative paucity of cells in the striola region was evident. Most of the stereocilia bundles examined were mature, with well-de¢ned kinocilium and staircase bundles. However, there were also a number of immature stereocilia bundles present. The striola line was easily identi¢ed by the polarization reversal of hair cells (Fig. 4C). The hair cell density for the 6 week animals continued to be greatest at the peripheral and central regions with total hair cell density approximately 50% of normal values (Fig. 5C). The hair cell densities were as follows : striola 16 ( þ 5.8 S.D.), central 22 ( þ 6.6 S.D.), and periphery 21 ( þ 5.3 S.D.) cells/1000 Wm2 , with all means remaining signi¢cantly below normal values (striola: F(1,6) = 15.7, P 6 0.05 ; central: F(1,6) = 16.6, P 6 0.05; peripheral: F(1,6) = 14.1, P 6 0.05). At 12 weeks post-streptomycin treatment, a continued increase in hair cell density was noted, as shown in Fig. 3C. The density of hair cells was essentially equivalent for all regions of the maculae. The most dramatic
HEARES 3228 17-6-99
B.J. Dye et al. / Hearing Research 133 (1999) 17^26
23
Fig. 5. Scanning electron micrographs of peripheral region at di¡erent time intervals following treatment with streptomycin. A: Two weeks following treatment. B: Four weeks following treatment. C: Six weeks following treatment. D: Twelve weeks following treatment. Scale bars = 10 Wm for all micrographs.
di¡erence in hair cell number was observed in the striola region (Fig. 4D). The striola had a mean density value of 30 ( þ 5.9 S.D.) cells/1000 Wm2 that was not signi¢cantly di¡erent from normal utricles (F(1,6) = 0.45, P s 0.5). The mean hair cell density values for the peripheral and central regions were 30 ( þ 2.5 S.D.) and 29 ( þ 5.5 S.D.) cells/1000 Wm2 , respectively. Although the striola region had returned to normal density levels by 12 weeks survival, the central (F(1,6) = 7.7, P 6 0.05) and peripheral (F(1,6) = 4.7, P 6 0.05) regions remained signi¢cantly less dense than normal utricles. All of the observed kinocilia were eccentrically placed to polarize the cells toward the well-de¢ned striola. The kinocilia were thicker, longer and more mature in appearance. In some samples, pockets of epithelium were less dense than others (Fig. 3C). At 24 weeks, the utricular surface was nearly normal in appearance except that a greater density of hair cells in the striola region was often visible. With rare exceptions, the kinocilia were oriented towards the striola throughout the macula surface. Still, there were a few rare hair cells that appeared to have random polariza-
Fig. 6. Hair cell density plotted as a function of post-treatment time. The mean hair cell densities for the striola (circles), central (squares) and peripheral (diamonds) regions are shown. The mean density of hair cells for the striola region is lower than either the central or peripheral regions initially, but by 12 weeks all three were equivalent. Normal mean hair cell densities are shown for comparison. Single exponential function (thick dashed line) ¢t to the mean central hair cell density values is plotted (HCD = hair cell density, PTT = post-treatment time in days).
HEARES 3228 17-6-99
24
B.J. Dye et al. / Hearing Research 133 (1999) 17^26
tions. The striola region, in all animals examined, actually had a higher than normal cell density with a mean value of 38 ( þ 0.8 S.D.) cells/1000 Wm2 , but this di¡erence was not statistically signi¢cant. The central and peripheral regions had cell densities that had returned to normal values, with means of 43 ( þ 2.2 S.D.) and 42 ( þ 3.3 S.D.) cells/1000 Wm2 , respectively. In general, at 24 weeks survival, the regenerated neuroepithelium was indistinguishable from that of the normal utricle except for a relatively more dense striola region. As shown in Fig. 6, the rate of regeneration in the utricular macula, regardless of region measured, proceeded in a non-linear fashion. The time course of hair cell repopulation could best be described using a single exponential curve (hair cell density = a(13e3bpostÿtreatment time )). Curves were ¢t to the longitudinal data for each of the macula regions, however only the central region ¢t is shown for clarity. Using the best-¢t parameter values, a `time constant' for regeneration based upon hair cell density was calculated to be approximately 71 days for the central region data. The time constant represents the post-treatment time (from day 0) that 63% of hair cell density observed at 24 weeks was obtained. 4. Discussion In the present study, SEM was used to quantify the regeneration of hair cells in the utricular macula following ototoxic insult. An important factor in the analysis was the complete elimination of all hair cells from the neuroepithelium at a discrete point in time. With SEM, only the presence of stereocilia bundles was used as evidence of the existence of hair cells. In a previous investigation in our lab, Frank et al. (1998) established the parameters and administration route for the antibiotic treatment method currently employed. Both SEM and light microscopy of utricular maculae showed a similar complete loss of stereocilia and the absence of underlying hair cells. Supporting cells and some other cells that appeared to be undergoing di¡erentiation and division in the neuroepithelium were still present at 4 days post-treatment time. Thus, based upon the present results and those of the previous methods study (Frank et al., 1998), it seems clear that at 4 days following a single intralabyrinthine application of streptomycin, no functional hair cells were present in the utricular neuroepithelium. This time point of 4 days can then serve as a baseline for the present and future studies regarding anatomical and functional regenerative recovery. At various time intervals from 2 to 24 weeks, the utricle epithelium revealed progressive hair cell regener-
ation and maturation. This progressive regeneration was consistent between the birds composing each group. Site speci¢c di¡erences were noted as the hair cell regeneration took place. Initially, regeneration occurred more rapidly in the central and peripheral regions of the utricle as compared to the striola (Figs. 4^6). In birds, these regions have a much higher concentration of type II hair cells as compared to type I (JÖrgensen and Andersen, 1973). Type II hair cells have been reported to regenerate faster than type I hair cells in several species (Carey et al., 1996 ; Lopez et al., 1997; Masetto and Correia, 1997; Weisleder and Rubel, 1993). Initially, hair cell density in the striola region during regeneration lagged behind that observed for the central and peripheral regions. The striola region in birds has the highest concentration of type I hair cells, except for the unique 4^6 type II cell band that de¢nes the striola line (JÖrgensen and Andersen, 1973). Thus, as type I cells regenerate more slowly, it is consistent to expect that the striola region would have lower initial densities. By the 12 week post-treatment time, however, densities in all three measured regions were equivalent. Hair cell regeneration in the striola region at 24 weeks appeared to be slightly greater than that observed for normal control animals, however the difference was not signi¢cant. At 24 weeks post-treatment time, the density of hair cells in the central and peripheral regions was similar to normal controls. The rate of regeneration appeared to follow an exponential function with a hair cell density recovery time constant of approximately 70 days. Still, at the longest survival time examined, 24 weeks, an asymptotic level for hair cell density may not have been reached, so that the derived time constant value may be underrepresented. It is possible that hair cell proliferation continues after this time, although the rate of density increase appears to be slowing, if not completed. To date, only a few studies addressing recovery of functional loss in the vestibular system during regeneration have been reported. For example, evoked potentials elicited during naso-occipital linear accelerations in chicks were examined following ototoxic treatment, with response thresholds increasing an order of magnitude and response amplitude decreased to near zero on the ¢rst day following streptomycin injections (Jones and Nelson, 1992). No examination of end organs was performed, so it is unknown how many residual hair cells remained after the systemic application was completed. However, initial low level responses progressed steadily so that at 2 weeks post-treatment, response threshold was near normal while latency was still increased and amplitude dramatically lower. By 8 weeks post-treatment, normal response values were obtained. In a second study, recovery of the horizontal vestibulo-ocular re£ex (HVOR) in chicks was examined
HEARES 3228 17-6-99
B.J. Dye et al. / Hearing Research 133 (1999) 17^26
in correlation with hair cell regeneration (Carey et al., 1996). Systemic treatment with streptomycin over an 8 day period produced signi¢cant hair cell loss in the horizontal semicircular canals, however when examined histologically at 1^4 days post-treatment up to 40% of the normal value of hair cells (mostly type II) remained. As noted by the authors, it is di¤cult to assess whether these remaining hair cells were residual or newly regenerated. Still, when the HVOR was tested immediately following ototoxic treatment, essentially no or very little response was present in the chicks. HVOR response returned within the ¢rst 1^3 weeks, as hair cell density in the horizontal ampulla increased. By 8^9 weeks posttreatment time, both the HVOR gain and phase values and the hair cell densities had returned to normal. Both of these functional recovery investigations provide important data, in that vestibular responses were greatly reduced or eliminated as a result of ototoxic insult and gradually returned as the neuroepithelia regenerated. Still, the uncertainty of complete hair cell loss following treatment in these studies leaves quantitative evaluation of anatomical and functional recovery di¤cult. As a comparison, in the present study it was observed that at 6 weeks post-treatment time, the utricular hair cell density was approximately half that of the normal macula regardless of epithelial region examined. Even with 12 weeks survival, the density of hair cells in the central and peripheral regions was signi¢cantly below normal values. Although no empirical assessment of vestibular related responses was performed, it seems unlikely that complete functional recovery could be expected by 12 weeks following complete utricular hair cell destruction. Thus, the quicker hair cell proliferation and functional recovery at 8 weeks regeneration as described for the chick studies (Carey et al., 1996; Jones and Nelson, 1992) may have been accelerated due to the presence of residual receptor cells and neural connections that were una¡ected by the ototoxic treatment. With the present intralabyrinthine administration method, both the structural and functional aspects of regeneration can be more thoroughly evaluated, since a discrete time point identi¢ed with no functional hair cells in the vestibular end organs are present. In conclusion, we have developed a technique of intraotic application of streptomycin in which complete utricular hair cell destruction is achieved while maintaining regenerative capacity. Following streptomycin application, the regeneration of hair cells in the utricle macula was progressive, consistent and reproducible. We noted site speci¢c di¡erences in the regeneration rate, with the striola developing more slowly than the central and peripheral regions of the macula. The quantitative veri¢cation of this technique solidi¢es it as a reliable method of studying utricular hair cell regeneration in the adult pigeon.
25
Acknowledgements The authors would like to gratefully acknowledge the ¢ne assistance provided to this project by Vijaya Onguri, Wanda Dorsett-Martin, DVM, and Glenn Hoskins. This work was supported in part by funds provided from the NIH (R01 DC03286) and NASA (NAG2-1204).
References Baird, R.A., Torres, M.A., Schu¡, N.R., 1993. Hair cell regeneration in the bullfrog vestibular otolith organs following aminoglycoside toxicity. Hear. Res. 65, 164^174. Berg, K., 1951. The toxic e¡ect of streptomycin on the vestibular and cochlear apparatus experimental study on cats. Acta Otolaryngol. Suppl. 97, 1. Carey, J.P., Fuchs, A.F., Rubel, E.W., 1996. Hair cell regeneration and recovery of the vestibulo-ocular re£ex in the avian vestibular system. J. Neurophysiol. 76, 3301^3312. Carranza, A., Lopez, I., Castellano, P., Ho¡man, L., Honrubia, V., 1997. Intraotic administration of gentamicin: A new method to study ototoxicity in the crista ampullaris of the bullfrog. Laryngoscope 107, 137^142. Cruz, R., Lambert, P., Rubel, E., 1987. Light microscopic evidence of hair cell regeneration after gentamicin toxicity in chick cochlea. Arch. Otolaryngol. Head Neck Surg. 113, 1058^1062. Dey, S., Baul, T.S.B., Roy, B., Dey, D., 1989. A new rapid method of air-drying for scanning electron microscopy using tetramethylsilane. J. Microsc. 156, 259^261. Flock, A., 1964. Structure of the macula utriculi with special reference to directional interplay of sensory responses as revealed by morphological polarization. J. Cell Biol. 22, 413^431. Forge, A., Lie, L., Corwin, J.T., Nevill, G., 1993. Ultrastructural evidence for hair cell regeneration in the mammalian inner ear. Science 259, 1616^1619. Frank, T.C., Dye, B.J., Newlands, S.D., Dickman, J.D., 1998. Streptomycin ototoxicity and hair cell regeneration in the pigeon utricle. Laryngoscope (in press). Jones, T.A., Nelson, R.C., 1992. Recovery of vestibular function following hair cell destruction by streptomycin. Hear. Res. 62, 181^ 186. JÖrgensen, J.M., Andersen, T., 1973. On the structure of the avian maculae. Acta Zool. 54, 121^130. Lindeman, H.H., 1969. Regional di¡erences in sensitivity of the vestibular sensory epithelia to ototoxic antibiotics. Acta Otol. 67, 177^189. Lippe, W.R., Westbrook, E.W., Ryals, B.M., 1991. Hair cell regeneration in the chicken cochlea following aminoglycoside toxicity. Hear. Res. 56, 203^210. Lombarte, A., Yan, H.Y., Popper, A.N., Chang, J.S., Platt, C., 1993. Damage and regeneration of hair cell ciliary bundles in a ¢sh ear following treatment with gentamicin. Hear. Res. 64, 166^174. Lopez, I., Honrubia, V., Lee, S.C., Schoeman, G., Beykirch, K., 1997. Quanti¢cation of the process of hair cell loss and recovery in the chinchilla crista ampullaris after gentamicin treatment. Int. J. Dev. Neurosci. 15, 447^461. Masetto, S., Correia, M.J., 1997. Ionic currents in regenerating avian vestibular hair cells. Int. J. Dev. Neurosci. 15, 387^399. Omura, R., Harada, Y., Suzuki, M., Hirakawa, H., 1989. Structural and physiologic change of bullfrog semicircular canal due to gen-
HEARES 3228 17-6-99
26
B.J. Dye et al. / Hearing Research 133 (1999) 17^26
tamycin intoxication. Acta Otolaryngol. (Stockh.) 468 (Suppl.), 41^47. Weisleder, P., Rubel, E., 1992. Hair cell regeneration in the avian vestibular epithelium. Exp. Neurol. 115, 2^6. Weisleder, P., Rubel, E., 1993. Hair cell regeneration after streptomycin toxicity in the avian vestibular epithelium. J. Comp. Neurol. 331, 97^110.
Wersa«ll, J., 1956. Studies on the structure and innervation of the sensory epithelium of the cristae ampullares in the guinea pig. Acta Otol. Suppl. 126, 1^85. Wersa«ll, J., Hawkins, J.E., Jr., 1962. The vestibular sensory epithelia in the cat labyrinth and their reactions in chronic streptomycin intoxication. Acta Otol. 54, 1^23.
HEARES 3228 17-6-99
Distribution and time course of hair cell regeneration in the pigeon utricle Bradford J. Dye, Thomas C. Frank, Shawn D. Newlands, J. David Dickman * University of Mississippi Medical Center, Department of Surgery, Division of Otolaryngology, 2500 North State Street, Jackson, MS 39216-4505, USA Received 16 November 1998; received in revised form 27 February 1999; accepted 3 March 1999
Abstract Vestibular and cochlear regeneration following ototoxic insult from aminoglycoside antibiotics has been well documented, particularly in birds. In the present study, intraotic application of a 2 mg streptomycin paste was used to achieve complete vestibular hair cell destruction in pigeons (Columba livia) while preserving regenerative ability. Scanning electron microscopy was used to quantify hair cell density longitudinally during regeneration in three different utricular macula locations, including the striola, central and peripheral regions. The utricular epithelium was void of stereocilia (indicating hair cell loss) at 4 days after intraotic treatment with streptomycin. At 2 weeks the stereocilia began to appear randomly and mostly in an immature form. However, when present most kinocilia were polarized toward the developing striola. Initially, regeneration occurred more rapidly in the central and peripheral regions of the utricle as compared to the striola. As regeneration proceeded from 2 to 12 weeks, hair cell density in the striola region equaled the density noted in the central and peripheral regions. At 24 weeks, hair cell density of the central and peripheral regions was equal to normal values, however the striola region had a slightly greater hair cell density than that observed for normal animals. ß 1999 Elsevier Science B.V. All rights reserved. Key words: Regeneration; Vestibular ; Hair cell; Ototoxicity
1. Introduction The ototoxicity of aminoglycoside antibiotics to the vestibular and auditory systems has been known for decades and well documented by studies in numerous species (Berg, 1951; Wersa«ll and Hawkins, 1962). More recent research has demonstrated the regenerative ability of vestibular and auditory hair cells following administration of aminoglycoside antibiotics, particularly in birds (Cruz et al., 1987; Lippe et al., 1991; Weisleder and Rubel, 1992). For the vestibular receptors, the majority of these studies have relied on intramuscular injections of aminoglycosides in young animals with an emphasis on hair cell regeneration in the ampullary organs. However, systemic administration of streptomycin often failed to completely eliminate all of the hair cells in the vestibular end organs, particularly the receptors of the maculae, even when dosages were suf-
* Corresponding author. Tel.: +1 (601) 984 5090; Fax: +1 (601) 984 5085; E-mail: [email protected]
¢cient to produce high mortality rates (Carey et al., 1996 ; Forge et al., 1993; Frank et al., 1998; Weisleder and Rubel, 1993). Omura et al. (1989) compared intralabyrinthine, intraperitoneal, and intramuscular injections of gentamicin in bullfrogs and found that intralabyrinthine injections resulted in the greatest hair cell loss for the posterior ampulla, but no documentation of otolith organ e¡ects were provided. Carranza et al. (1997) recently described a technique of direct intraotic administration of gentamicin soaked pledgets in the bullfrog which resulted in complete ampullary hair cell destruction as demonstrated by light microscopy; however, utricular hair cell destruction was not discussed. Baird et al. (1993) examined hair cell regeneration in the utricle and saccule of bullfrogs, while Lombarte et al. (1993) studied regeneration in the utricle and lagena of a teleost ¢sh, however signi¢cant numbers of hair cells in the extrastriola regions were unaffected by aminoglycoside treatment in both investigations. With these various methods of ototoxic agent administration resulting in non-uniform hair cell loss that occurs over a period of days, quantitative investi-
0378-5955 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 5 9 5 5 ( 9 9 ) 0 0 0 4 6 - 5
HEARES 3228 17-6-99
18
B.J. Dye et al. / Hearing Research 133 (1999) 17^26
gations of hair cell regeneration become di¤cult. In addition, it is not possible to formulate with certainty the time course or distribution of receptor cell regeneration without complete hair cell loss occurring at a discrete point in time. In a previous study (Frank et al., 1998), we demonstrated that a single direct application of streptomycin sulfate into the perilymphatic space was su¤cient to produce complete vestibular hair cell loss. With one intralabyrinthine application, a minimal dose of 2 mg streptomycin was required to produce a reliable complete destruction of all vestibular hair cells in both the ampullae and maculae. The regenerative capacity of the hair cells was preserved, although the nature of the regeneration pattern for the vestibular hair cells was not extensively examined. In the present investigation, we utilized the advantages of intralabyrinthine application, including single dose administration, complete hair cell loss, and discrete point in time of ototoxic initiation to study regeneration in the otolith organs of adult birds. Of particular interest is the time course and distribution of receptor cell regeneration in the utricle, which has received little attention in regeneration studies. The results show that regeneration of utricular hair cells occurs in di¡erent regions of the adult pigeon utricle at di¡erent times following ototoxic insult. 2. Materials and methods 2.1. Animals The experiments were conducted in adult male pigeons (Columba livia), that ranged in age between 1 and 3 years. All animals were housed in the Institution Laboratory Animal Facility. The methods of study were approved by the University of Mississippi Medical Center Animal Care and Use Committee and all animal procedures were performed in accordance with the Institution Animal Care and Use Committee guidelines. 2.2. Surgical preparation Prior to surgery, the streptomycin pellets were prepared. The pellets were created by adding 0.1 cc sterile saline to 2 mg of streptomycin sulfate powder (Sigma), then mixing into a slurry paste and storing at 311³C. For intraotic administration of the streptomycin, the animals were anesthetized using intravenous pentobarbital sodium (16 mg/kg) and intramuscular ketamine (20 mg/kg). Ampicillin (37.5 mg) was given intramuscularly prior to surgery. A midline skin incision provided exposure to the mastoid cortex which allowed
visualization of the horizontal canal through the thin bone layer. The cortex was removed using a rectangular shaped bone £ap thus exposing the bony labyrinth. Using a surgical drill, the vestibule was opened (0.35 mm diameter hole) which allowed for careful placement of the 2 mg streptomycin chips into the perilymphatic space. A temporalis muscle plug was placed over the surgically created ¢stula. This was then reinforced with gelfoam and bone wax. Next, the bone £ap was placed in its original position. The procedure was repeated on the opposite side and the incisions closed. A single dose of butorphanol (0.05 mg/kg) was given postoperatively for discomfort. On the day following the surgery, the birds were begun on a hand-fed diet of chick formula (Kaytee) for 3 days. Following this, the birds were allowed to feed ad libitum. The animals were monitored for weight loss, and hand feeding was resumed as needed to prevent dehydration and starvation. A second control group of pigeons was also used, in which an identical surgical procedure was performed, except that a similar volume of normal saline (frozen chips) was placed into the perilymphatic space. 2.3. Experimental procedure Several groups of animals were used. A normal group of animals (n = 5) received no antibiotic treatment or surgical procedure, while a saline group (n = 2) received only a saline application into the labyrinth. All treated animals that received a single streptomycin application were placed into di¡erent groups based upon survival time following ototoxic treatment. The treated groups included 4 days (n = 3), 2 weeks (n = 3), 4 weeks (n = 2), 6 weeks (n = 3), 12 weeks (n = 3), and 24 weeks (n = 2) survival. Following the speci¢ed survival time, each animal was killed and the vestibular receptor organs removed from the inner ear. 2.4. Histology The pigeons were anesthetized (sodium pentobarbital, 250 mg/kg) and the right horizontal and posterior semicircular canals were opened. The labyrinth was perfused with 5 cc of an aldehyde ¢xative consisting of 2% glutaraldehyde, 3% paraformaldehyde and 1% acrolein. Immediately following the labyrinthine ¢xation, the thoracic cavity was opened and a transcardiac perfusion was performed using 1000 cc of the same 2% paraformaldehyde, 3% glutaraldehyde solution (no acrolein). The head was removed and placed into the aldehyde ¢xative overnight. On the following day, the membranous labyrinths were removed from the inner ear and the utricle isolated by careful dissection. Only utricles from the right ear were utilized for scanning electron
HEARES 3228 17-6-99
B.J. Dye et al. / Hearing Research 133 (1999) 17^26
19
Fig. 1. Scanning electron micrographs of utricular macula in normal pigeons. For all SEM preparations, the otolith membrane and otoconia have been removed. A: Whole utricle with inset boxes showing regions for enlargement and quanti¢cation of hair cell counts. Small ¢ssures in the macula appeared as the tissue was £attened for viewing. Normally, the anterior end of the utricle is curled upward, to form an L-shaped surface. Anterior: leftward; lateral: upward. Scale bar = 100 Wm. B: Striola region with dashed line representing the hair cell polarization reversal point. Scale bar = 10 Wm. C: Central region, scale bar = 10 Wm. D: Peripheral region, scale bar = 10 Wm.
microscopy (SEM) analysis in all animals, while the left ear was used for a separate neural tracing investigation. The utricle was next prepared for SEM. First, the otoconia and otolith membranes were removed by manual dissection. Next, the tissue was placed into a 0.5% osmium and 0.1 M phosphate bu¡er solution for 2 h. After six serial washes of 10 min each in distilled water, the tissue was dehydrated. This was done with serial washes of 10 min each in 70%, 90% and 95% acetone followed by three washes of 15 min in 100% acetone. The tissue was then incubated in a 1:1 mixture of tetramethylsilane (TMS) and acetone for 45 min followed by an incubation of 45 min in a 3:1 mixture of TMS and acetone. The tissue was then incubated twice in 100% TMS for 1 h. Finally, the tissue was further dried in the oven for 30 min, allowing the TMS to sublimate (Dey et al., 1989). The tissue was then mounted onto aluminum studs and gold coated. The tissue was scanned and photographed using a JEOL TSM-3000 scanning electron microscope (15 kV). 2.5. Data analysis The entire right utricle (U75) and three separate re-
gions of the neuroepithelium (U1000) for each animal were photographed, in order to quantify the distribution and time course of regeneration. The three regions included the striola, the central (juxtastriola), and the peripheral regions, as shown in Fig. 1. The striola region was identi¢ed in the normal and regeneration animals by the reversal of the stereocilia bundle morphological polarization. The striola line was placed into the center of the view ¢eld for quanti¢cation. The central region was de¢ned at approximately the midpoint in the utricular epithelium, while the peripheral region was de¢ned by the medial border of the epithelium. Hair cells within a speci¢ed 9111 Wm2 rectangular area window for each region were counted, using the presence of a stereocilia bundle to indicate the existence of a hair cell. For newly regenerating hair cells, the identi¢cation of a kinocilium was required in order for the cell to be counted. The hair cell counts for each region within each treatment group were averaged. Statistical comparisons between regions and treatment groups were made using repeated measures ANOVA (Statistica). A single exponential curve was ¢t to the hair cell density data as a function of post-treatment time, using a leastsquares algorithm (SigmaPlot).
HEARES 3228 17-6-99
20
B.J. Dye et al. / Hearing Research 133 (1999) 17^26
Fig. 2. Scanning electron micrographs of utricular macula 4 days following treatment with streptomycin. A: Striola region with no stereocilia bundles present. B: Peripheral region with no stereocilia visible. Polygonal shaped supporting cells appear to comprise the entire neuroepithelium with a single microvillus protruding from the apical surface of each cell. Scale bars = 10 Wm.
3. Results 3.1. Normal utricle The normal pigeon utricle measures approximately 1.2 mm in length and 1.0 mm in width. In all ¢ve of the normal utricles examined, the macula surface was densely populated with hair cells, as shown in Fig. 1. As previously noted by many investigators (Flock, 1964 ; Wersa«ll, 1956; Lindeman, 1969) the hair cells are morphologically polarized according to the position of the kinocilium, which nearly always lay at the edge of the hair cell closest to the striola. In addition, the kinocilium of each cell was longer and thicker than the neighboring stereocilia. The striola was typically approxi-
mately C shaped as it curved along the outer lateral and anterior edges of the neuroepithelium. The striola line, as de¢ned by the reversal of hair cell morphological polarization, was not strictly uniform but generally lay within 100^175 Wm of the lateral/anterior border (Fig. 1A,B). The hair cells had a staircase array of stereocilia, with a lone larger kinocilium. Each hair cell was typically surrounded in a polygon pattern by 4^6 supporting cells, characterized by a single apical cilium. The density of hair cells was quanti¢ed for each of the three regions in ¢ve utricles from di¡erent animals. The central and peripheral regions had similar mean hair cell densities of 41 ( þ 6 S.D.) and 40 ( þ 8 S.D.) cells/1000 Wm2 , respectively. The striola region had fewer hair cells
HEARES 3228 17-6-99
B.J. Dye et al. / Hearing Research 133 (1999) 17^26
21
than the other two regions in all animals examined with a mean value of 34 ( þ 6 S.D.) cells/1000 Wm2 . However, this di¡erence was not signi¢cant (F(1,8) = 3.63, P 6 0.09). In two animals, normal saline instead of the ototoxic antibiotic was placed into the perilymphatic space using the same surgical technique. No observable hair cell destruction was noted at 4 days survival following the treatment. The mean hair cell density values in the control animals for the striola (27 þ 3 S.D.), central (37 þ 3 S.D.), and peripheral (38 þ 2 S.D.) regions were not signi¢cantly di¡erent from those observed for the normal animals. 3.2. E¡ects of streptomycin At 4 days after treatment with streptomycin, no visible stereocilia were observed in any of the six utricles examined, as illustrated in Fig. 2. For quantitative measures, three animals were photographed and counted, according to regions. The striola region could not be identi¢ed precisely, due to the lack of any visible stereocilia. Thus, the striola region was approximated by placing the view ¢eld at 100 Wm inside the lateral border, close to the average striola location in untreated animals. Still, no stereocilia were visible in any of the cell surfaces throughout the maculae. As shown in Fig. 2A for the striola region and in Fig. 2B for the peripheral region, support cells appear to comprise the entire neuroepithelium. Support cells had a characteristic polygonal shape and a single small cilium protruding from the apical surface. These lone cilia were also observable in the support cells of normal non-treated animals. Thus, at 4 days post-treatment following intraotic administration of streptomycin, a complete destruction of hair cells was observed. 3.3. Regeneration of utricular hair cells The time course of hair cell regeneration was studied over a 24 week recovery time continuum. As shown in Fig. 3, relatively few stereocilia bundles were present at early post-treatment periods, however over time the utricle neuroepithelium became repopulated to normal values. For example, at 2 weeks following streptomycin application, the number of visible hair cells was only a fraction of the normal pattern (compare Figs. 3A and 1A). However, the stereocilia from a number of cells had begun to appear distributed in a random fashion across the epithelial surface. When examined at higher magni¢cation, as shown in Fig. 4, some hair cells in the 2 week macula contained mature stereocilia bundles. However, many of the stereocilia bundles examined at 2 weeks were immature in appearance with only a few cilia and no distinguishable kinocilium. When present,
Fig. 3. Scanning electron micrographs of whole utricular maculae at di¡erent time intervals following treatment with streptomycin. A: Utricle at 2 weeks post-treatment time with few stereocilia bundles present. B: Utricle at 6 weeks post-treatment. Hair cell density has greatly increased, as evidenced by the presence of stereocilia bundles. Striola region is visibly less dense than central or peripheral regions. C: Utricle at 12 weeks following streptomycin treatment. Hair cell density continues to increase for all regions. Anterior: leftward; lateral: upward. Scale bars = 100 Wm for all micrographs.
most of the kinocilia were directionally oriented, even in the maculae of the 2 week post-treatment survival animals. For example, the morphological polarization of hair cells in the lateral margins of the epithelium were developing a de¢nable striola, as evidenced by oppositely directed stereocilia bundles, similar to that
HEARES 3228 17-6-99
22
B.J. Dye et al. / Hearing Research 133 (1999) 17^26
Fig. 4. Scanning electron micrographs of striola region at di¡erent time intervals following treatment with streptomycin. A, B, C, D: Striola region at 2, 4, 6, and 12 weeks post-treatment time, respectively. Dashed lines represent the morphological polarization reversal of hair cells, indicating the striola line. Scale bars = 10 Wm for all micrographs.
of normal animals. In the periphery (Fig. 5A), all of the kinocilium were directed toward the lateral or anterior margin (site of the developing striola). At the early time point of 2 weeks, the cell densities for the central and peripheral regions were similar, both with mean values of 10 ( þ 0.5 and þ 1.1 S.D.) cells/1000 Wm2 , respectively. Cell densities in the striola region were lower for all animals examined, with a mean 8 ( þ 1.8 S.D.) cells/1000 Wm2 . The cell densities for all regions at 2 weeks post-treatment were signi¢cantly less than those observed in normal utricles (striola: F(1,6) = 46.9, P 6 0.05; central: F(1,6) = 63.3, P 6 0.05 ; peripheral : F(1,6) = 41.9, P 6 0.05). At 4 weeks, the stereocilia appeared in increasing numbers and level of maturity, as shown in Figs. 4B and 5B. In fact, the density of cells had doubled over the 2 week mean values. The peripheral and central regions continued to be more densely populated with mature stereocilia than the striola region. In addition, the striola line became better de¢ned, as greater numbers of hair cells exhibited morphological polarizations (Fig. 4B). At 6 weeks, an increasing number of hair cells with
stereocilia were visible, as shown in the whole utricle example of Fig. 3B. The overall organization and distribution of the hair cells had an appearance that was more similar to that of the normal utricle. Still, a relative paucity of cells in the striola region was evident. Most of the stereocilia bundles examined were mature, with well-de¢ned kinocilium and staircase bundles. However, there were also a number of immature stereocilia bundles present. The striola line was easily identi¢ed by the polarization reversal of hair cells (Fig. 4C). The hair cell density for the 6 week animals continued to be greatest at the peripheral and central regions with total hair cell density approximately 50% of normal values (Fig. 5C). The hair cell densities were as follows : striola 16 ( þ 5.8 S.D.), central 22 ( þ 6.6 S.D.), and periphery 21 ( þ 5.3 S.D.) cells/1000 Wm2 , with all means remaining signi¢cantly below normal values (striola: F(1,6) = 15.7, P 6 0.05 ; central: F(1,6) = 16.6, P 6 0.05; peripheral: F(1,6) = 14.1, P 6 0.05). At 12 weeks post-streptomycin treatment, a continued increase in hair cell density was noted, as shown in Fig. 3C. The density of hair cells was essentially equivalent for all regions of the maculae. The most dramatic
HEARES 3228 17-6-99
B.J. Dye et al. / Hearing Research 133 (1999) 17^26
23
Fig. 5. Scanning electron micrographs of peripheral region at di¡erent time intervals following treatment with streptomycin. A: Two weeks following treatment. B: Four weeks following treatment. C: Six weeks following treatment. D: Twelve weeks following treatment. Scale bars = 10 Wm for all micrographs.
di¡erence in hair cell number was observed in the striola region (Fig. 4D). The striola had a mean density value of 30 ( þ 5.9 S.D.) cells/1000 Wm2 that was not signi¢cantly di¡erent from normal utricles (F(1,6) = 0.45, P s 0.5). The mean hair cell density values for the peripheral and central regions were 30 ( þ 2.5 S.D.) and 29 ( þ 5.5 S.D.) cells/1000 Wm2 , respectively. Although the striola region had returned to normal density levels by 12 weeks survival, the central (F(1,6) = 7.7, P 6 0.05) and peripheral (F(1,6) = 4.7, P 6 0.05) regions remained signi¢cantly less dense than normal utricles. All of the observed kinocilia were eccentrically placed to polarize the cells toward the well-de¢ned striola. The kinocilia were thicker, longer and more mature in appearance. In some samples, pockets of epithelium were less dense than others (Fig. 3C). At 24 weeks, the utricular surface was nearly normal in appearance except that a greater density of hair cells in the striola region was often visible. With rare exceptions, the kinocilia were oriented towards the striola throughout the macula surface. Still, there were a few rare hair cells that appeared to have random polariza-
Fig. 6. Hair cell density plotted as a function of post-treatment time. The mean hair cell densities for the striola (circles), central (squares) and peripheral (diamonds) regions are shown. The mean density of hair cells for the striola region is lower than either the central or peripheral regions initially, but by 12 weeks all three were equivalent. Normal mean hair cell densities are shown for comparison. Single exponential function (thick dashed line) ¢t to the mean central hair cell density values is plotted (HCD = hair cell density, PTT = post-treatment time in days).
HEARES 3228 17-6-99
24
B.J. Dye et al. / Hearing Research 133 (1999) 17^26
tions. The striola region, in all animals examined, actually had a higher than normal cell density with a mean value of 38 ( þ 0.8 S.D.) cells/1000 Wm2 , but this di¡erence was not statistically signi¢cant. The central and peripheral regions had cell densities that had returned to normal values, with means of 43 ( þ 2.2 S.D.) and 42 ( þ 3.3 S.D.) cells/1000 Wm2 , respectively. In general, at 24 weeks survival, the regenerated neuroepithelium was indistinguishable from that of the normal utricle except for a relatively more dense striola region. As shown in Fig. 6, the rate of regeneration in the utricular macula, regardless of region measured, proceeded in a non-linear fashion. The time course of hair cell repopulation could best be described using a single exponential curve (hair cell density = a(13e3bpostÿtreatment time )). Curves were ¢t to the longitudinal data for each of the macula regions, however only the central region ¢t is shown for clarity. Using the best-¢t parameter values, a `time constant' for regeneration based upon hair cell density was calculated to be approximately 71 days for the central region data. The time constant represents the post-treatment time (from day 0) that 63% of hair cell density observed at 24 weeks was obtained. 4. Discussion In the present study, SEM was used to quantify the regeneration of hair cells in the utricular macula following ototoxic insult. An important factor in the analysis was the complete elimination of all hair cells from the neuroepithelium at a discrete point in time. With SEM, only the presence of stereocilia bundles was used as evidence of the existence of hair cells. In a previous investigation in our lab, Frank et al. (1998) established the parameters and administration route for the antibiotic treatment method currently employed. Both SEM and light microscopy of utricular maculae showed a similar complete loss of stereocilia and the absence of underlying hair cells. Supporting cells and some other cells that appeared to be undergoing di¡erentiation and division in the neuroepithelium were still present at 4 days post-treatment time. Thus, based upon the present results and those of the previous methods study (Frank et al., 1998), it seems clear that at 4 days following a single intralabyrinthine application of streptomycin, no functional hair cells were present in the utricular neuroepithelium. This time point of 4 days can then serve as a baseline for the present and future studies regarding anatomical and functional regenerative recovery. At various time intervals from 2 to 24 weeks, the utricle epithelium revealed progressive hair cell regener-
ation and maturation. This progressive regeneration was consistent between the birds composing each group. Site speci¢c di¡erences were noted as the hair cell regeneration took place. Initially, regeneration occurred more rapidly in the central and peripheral regions of the utricle as compared to the striola (Figs. 4^6). In birds, these regions have a much higher concentration of type II hair cells as compared to type I (JÖrgensen and Andersen, 1973). Type II hair cells have been reported to regenerate faster than type I hair cells in several species (Carey et al., 1996 ; Lopez et al., 1997; Masetto and Correia, 1997; Weisleder and Rubel, 1993). Initially, hair cell density in the striola region during regeneration lagged behind that observed for the central and peripheral regions. The striola region in birds has the highest concentration of type I hair cells, except for the unique 4^6 type II cell band that de¢nes the striola line (JÖrgensen and Andersen, 1973). Thus, as type I cells regenerate more slowly, it is consistent to expect that the striola region would have lower initial densities. By the 12 week post-treatment time, however, densities in all three measured regions were equivalent. Hair cell regeneration in the striola region at 24 weeks appeared to be slightly greater than that observed for normal control animals, however the difference was not signi¢cant. At 24 weeks post-treatment time, the density of hair cells in the central and peripheral regions was similar to normal controls. The rate of regeneration appeared to follow an exponential function with a hair cell density recovery time constant of approximately 70 days. Still, at the longest survival time examined, 24 weeks, an asymptotic level for hair cell density may not have been reached, so that the derived time constant value may be underrepresented. It is possible that hair cell proliferation continues after this time, although the rate of density increase appears to be slowing, if not completed. To date, only a few studies addressing recovery of functional loss in the vestibular system during regeneration have been reported. For example, evoked potentials elicited during naso-occipital linear accelerations in chicks were examined following ototoxic treatment, with response thresholds increasing an order of magnitude and response amplitude decreased to near zero on the ¢rst day following streptomycin injections (Jones and Nelson, 1992). No examination of end organs was performed, so it is unknown how many residual hair cells remained after the systemic application was completed. However, initial low level responses progressed steadily so that at 2 weeks post-treatment, response threshold was near normal while latency was still increased and amplitude dramatically lower. By 8 weeks post-treatment, normal response values were obtained. In a second study, recovery of the horizontal vestibulo-ocular re£ex (HVOR) in chicks was examined
HEARES 3228 17-6-99
B.J. Dye et al. / Hearing Research 133 (1999) 17^26
in correlation with hair cell regeneration (Carey et al., 1996). Systemic treatment with streptomycin over an 8 day period produced signi¢cant hair cell loss in the horizontal semicircular canals, however when examined histologically at 1^4 days post-treatment up to 40% of the normal value of hair cells (mostly type II) remained. As noted by the authors, it is di¤cult to assess whether these remaining hair cells were residual or newly regenerated. Still, when the HVOR was tested immediately following ototoxic treatment, essentially no or very little response was present in the chicks. HVOR response returned within the ¢rst 1^3 weeks, as hair cell density in the horizontal ampulla increased. By 8^9 weeks posttreatment time, both the HVOR gain and phase values and the hair cell densities had returned to normal. Both of these functional recovery investigations provide important data, in that vestibular responses were greatly reduced or eliminated as a result of ototoxic insult and gradually returned as the neuroepithelia regenerated. Still, the uncertainty of complete hair cell loss following treatment in these studies leaves quantitative evaluation of anatomical and functional recovery di¤cult. As a comparison, in the present study it was observed that at 6 weeks post-treatment time, the utricular hair cell density was approximately half that of the normal macula regardless of epithelial region examined. Even with 12 weeks survival, the density of hair cells in the central and peripheral regions was signi¢cantly below normal values. Although no empirical assessment of vestibular related responses was performed, it seems unlikely that complete functional recovery could be expected by 12 weeks following complete utricular hair cell destruction. Thus, the quicker hair cell proliferation and functional recovery at 8 weeks regeneration as described for the chick studies (Carey et al., 1996; Jones and Nelson, 1992) may have been accelerated due to the presence of residual receptor cells and neural connections that were una¡ected by the ototoxic treatment. With the present intralabyrinthine administration method, both the structural and functional aspects of regeneration can be more thoroughly evaluated, since a discrete time point identi¢ed with no functional hair cells in the vestibular end organs are present. In conclusion, we have developed a technique of intraotic application of streptomycin in which complete utricular hair cell destruction is achieved while maintaining regenerative capacity. Following streptomycin application, the regeneration of hair cells in the utricle macula was progressive, consistent and reproducible. We noted site speci¢c di¡erences in the regeneration rate, with the striola developing more slowly than the central and peripheral regions of the macula. The quantitative veri¢cation of this technique solidi¢es it as a reliable method of studying utricular hair cell regeneration in the adult pigeon.
25
Acknowledgements The authors would like to gratefully acknowledge the ¢ne assistance provided to this project by Vijaya Onguri, Wanda Dorsett-Martin, DVM, and Glenn Hoskins. This work was supported in part by funds provided from the NIH (R01 DC03286) and NASA (NAG2-1204).
References Baird, R.A., Torres, M.A., Schu¡, N.R., 1993. Hair cell regeneration in the bullfrog vestibular otolith organs following aminoglycoside toxicity. Hear. Res. 65, 164^174. Berg, K., 1951. The toxic e¡ect of streptomycin on the vestibular and cochlear apparatus experimental study on cats. Acta Otolaryngol. Suppl. 97, 1. Carey, J.P., Fuchs, A.F., Rubel, E.W., 1996. Hair cell regeneration and recovery of the vestibulo-ocular re£ex in the avian vestibular system. J. Neurophysiol. 76, 3301^3312. Carranza, A., Lopez, I., Castellano, P., Ho¡man, L., Honrubia, V., 1997. Intraotic administration of gentamicin: A new method to study ototoxicity in the crista ampullaris of the bullfrog. Laryngoscope 107, 137^142. Cruz, R., Lambert, P., Rubel, E., 1987. Light microscopic evidence of hair cell regeneration after gentamicin toxicity in chick cochlea. Arch. Otolaryngol. Head Neck Surg. 113, 1058^1062. Dey, S., Baul, T.S.B., Roy, B., Dey, D., 1989. A new rapid method of air-drying for scanning electron microscopy using tetramethylsilane. J. Microsc. 156, 259^261. Flock, A., 1964. Structure of the macula utriculi with special reference to directional interplay of sensory responses as revealed by morphological polarization. J. Cell Biol. 22, 413^431. Forge, A., Lie, L., Corwin, J.T., Nevill, G., 1993. Ultrastructural evidence for hair cell regeneration in the mammalian inner ear. Science 259, 1616^1619. Frank, T.C., Dye, B.J., Newlands, S.D., Dickman, J.D., 1998. Streptomycin ototoxicity and hair cell regeneration in the pigeon utricle. Laryngoscope (in press). Jones, T.A., Nelson, R.C., 1992. Recovery of vestibular function following hair cell destruction by streptomycin. Hear. Res. 62, 181^ 186. JÖrgensen, J.M., Andersen, T., 1973. On the structure of the avian maculae. Acta Zool. 54, 121^130. Lindeman, H.H., 1969. Regional di¡erences in sensitivity of the vestibular sensory epithelia to ototoxic antibiotics. Acta Otol. 67, 177^189. Lippe, W.R., Westbrook, E.W., Ryals, B.M., 1991. Hair cell regeneration in the chicken cochlea following aminoglycoside toxicity. Hear. Res. 56, 203^210. Lombarte, A., Yan, H.Y., Popper, A.N., Chang, J.S., Platt, C., 1993. Damage and regeneration of hair cell ciliary bundles in a ¢sh ear following treatment with gentamicin. Hear. Res. 64, 166^174. Lopez, I., Honrubia, V., Lee, S.C., Schoeman, G., Beykirch, K., 1997. Quanti¢cation of the process of hair cell loss and recovery in the chinchilla crista ampullaris after gentamicin treatment. Int. J. Dev. Neurosci. 15, 447^461. Masetto, S., Correia, M.J., 1997. Ionic currents in regenerating avian vestibular hair cells. Int. J. Dev. Neurosci. 15, 387^399. Omura, R., Harada, Y., Suzuki, M., Hirakawa, H., 1989. Structural and physiologic change of bullfrog semicircular canal due to gen-
HEARES 3228 17-6-99
26
B.J. Dye et al. / Hearing Research 133 (1999) 17^26
tamycin intoxication. Acta Otolaryngol. (Stockh.) 468 (Suppl.), 41^47. Weisleder, P., Rubel, E., 1992. Hair cell regeneration in the avian vestibular epithelium. Exp. Neurol. 115, 2^6. Weisleder, P., Rubel, E., 1993. Hair cell regeneration after streptomycin toxicity in the avian vestibular epithelium. J. Comp. Neurol. 331, 97^110.
Wersa«ll, J., 1956. Studies on the structure and innervation of the sensory epithelium of the cristae ampullares in the guinea pig. Acta Otol. Suppl. 126, 1^85. Wersa«ll, J., Hawkins, J.E., Jr., 1962. The vestibular sensory epithelia in the cat labyrinth and their reactions in chronic streptomycin intoxication. Acta Otol. 54, 1^23.
HEARES 3228 17-6-99