THE JOURNAL OF COMPARATIVE NEUROLOGY 3303521-532 (1993)
Reorganization of the Chick Basilar
Papilla After Acoustic Trauma
Kresge Hearing Research Institute, The University of Michigan, Ann Arbor,
The auditory epithelium in birds and mammals consists of a postmitotic population of hair
cells and supporting cells. Unlike mammals, birds can regenerate their auditory epithelia dter
trauma. Recent evidence indicates that supporting cells undergo mitosis after acoustic trauma,
suggesting that supporting cells may transdifferentiate into hair cells. The goals of this study
were to 1) characterize the responses of hair cells and supporting cells to acoustic trauma, and
2) determine whether hair cell loss is a prerequisite for generation of new hair cells. Chicks were
exposed to an octave-band noise and their inner ears assayed with fluorescence or scanning
electron microscopy, In one area of the basilar papilla, defined as the center of the lesion,
extensive hair cell degeneration occurred. Expanded supporting cells obliterated degenerating
hair cells and invaded spaces normally occupied by hair cells. Aggregates of DNA were found
within the basilar papilla, suggesting that hair cell death and disintegration may occur within
the epithelium. The epithelial sheet appeared structurally confluent at all times examined.
Supporting cells exhibited altered apical contour in distal regions of the basilar papilla, where
hair cell damage was mild or inconspicuous. Four days after noise exposure, newly generated
hair cells were found in the center of the lesion and in the distal areas, where no hair cell loss
could be detected. The results suggest that supporting cells may play an important role in
maintenance and repair of the traumatized basilar papilla and raise the possibility that production of
new hair cells is not dependent on hair cell loss in the immediate vicinity.
Key words: noise, repair, regeneration, mitosis
18 iw wiiey-~ks, Inc.
In chick, hair cells that are lost after acoustic or chemical
trauma are replaced by new, normal appearing hair cells
(Cotanche, '87; Corwin and Cotanche, '88; Ryals and Rubel,
'88). The mechanisms that initiate and control this regener-
ative process are largely unknown.
Mechanisms that control regeneration vary between
different sensory epithelial tissues. In the experimentally
damaged teleost fish eye, rod precursors can give rise to
more than one retinal cell type, suggesting that cues from
the surroundings influence the regenerative process (Ray-
mond, '91). Experiments with cultured aggregates of chick
retinal pigment epithelium have shown that transdifferen-
tiation of these cells is modulated by cell shape, the
composition of the extracellular matrix, and the presence of
bFGF (reviewed in Reh et al., '911. In the olfactory neuroep-
ithelium, degenerated neurons can be replaced by new
neurons that arise via mitotic divisions of basal cells
(Graziadei and Monti Graziadei, '85). The possibility that
extracellular matrix proteins and growth factors regulate
growth and differentiation in olfactory neuroepithelium
has also been investigated (Calof et al., '911, but the
mechanism of regulation has yet to be determined. In the
postembryonic Oscar (Astronotus ocellatus), a teleost fish,
"embryonic like" neuroepithelial cells serve as stem cells,
which can give rise to new hair cells and supporting cells
(Presson and Popper, '90). Nevertheless, it is not known
how proliferation and regeneration of new hair cells and
supporting cells are regulated. Thus, data at the cellular,
molecular, and genetic levels indicate that control mecha-
nisms that regulate epithelial regeneration are species and
3H thymidine studies in traumatized chick inner ears
showed that labeled hair cells and supporting cells were
present in the basilar papilla (BPI after several days of
recovery from acoustic overstimulation (Corwin and Co-
tanche, '88; Girod et al., '89). Thus, cell divisions appear to
play a role in repopulating the BP after trauma. Recent data
showing that supporting cells divide within 2 days after
trauma (Raphael. '92) have implicated supporting cells as a
source of newly generated hair cells. The signal(s1 regulat-
Acceptcd December 23, 1992.
Preliminary result,s were presented in part at the First Meeting on
Molecular Biology of Hearing and Deafness, San Diego, May 1992, and the
29th Workshop on Inner Ear Biology, Engelberg, Septemher 1992.
Q 1993 WILEY-LISS, INC.
ing supporting cell proliferation and differentiation in the
noise exposed BP are not known.
To explain the regulatory mechanism that controls gener-
ation and differentiation of new cells in the basilar papilla,
Corwin et al. ('91) suggested that presence of intercellular
contacts with hair cells inhibits supporting cells from
proliferating. According to this hypothesis, loss of two or
more hair cells would deprive supporting cells of hetero-
philic contacts, leading to removal of lateral inhibition and
a subsequent regenerative response.
The goals of this study were 1) to characterize the process
of hair cell degeneration and the relationships between hair
cell degeneration and the response of supporting cells, and
2) to test the hypothesis that changes in cell-cell contacts
that result from the loss of hair cells and their replacement
by expanded supporting cells are necessary to induce prolif-
eration of supporting cells. To that end, the responses of
supporting cells and hair cells to acoustic trauma were
documented by fluorescence or scanning electron micros-
copy on wholemounts of the BP.
MATERIALS AND METHODS
Animals and noise exposure
One week old white Leghorn chick hatchlings were placed
for 4 hours in a mesh-wire cage beneath a horn speaker that
delivered octave-band noise with a center frequency of 1.5
kHz and intensity of 120 dB SPL. The signal was generated
by a random noise generator (General Radio Company,
model 1381), filtered, and amplified (MC 2105, McIntosh
Power Amplifier). Intensity was calibrated with a Precision
Sound Level Meter (type 2203, Bruel & Kjaer). Chicks were
killed at 0 (n = lo), 6 (n = 5), 24 (n = 41, 48.(n = 8), or 96
(n = 6) hours after noise exposure. Seven chicks were used
as untreated controls.
Histo- and immunocytochemistry
Chicks were anesthetized and perfused intracardially
with 3% paraformaldehyde in 0.15 M phosphate buffer at
pH 7.35. Temporal bones were rapidly removed and BP
harvested. Tissues were prepared as wholemounts, perme-
abilized with 0.1% Triton X-100 in phosphate buffered
saline (PBS) for 5 minutes and stained with one or more of
the following probes: cingulin-specific antibodies were used
to label tight-junctions, phalloidin to stain F-actin, and
bisbenzimide trihydrochloride (Hoechst) to label DNA.
Nonspecific immunoreactivity was blocked in normal goat
serum (5% in PBS) for 30 minutes. Samples were then
incubated in rabbit anti-cingulin antibody (1:600 dilution;
Citi et al., '88, '89) for 60 minutes, rinsed thoroughly, and
immersed in a PBS solution containing Hoechst (2 mg/ml,
Sigma): rhodamine phalloidin (1: 100; Molecular Probes,
OR), and FITC-conjugated goat anti rabbit secondary anti-
body (1:lOO; Cappel, Durham, NC) for 30 minutes. After
rinsing, papillae were mounted in 60% glycerol in sodium
carbonate buffer (pH 8.5) with p-phenylenediamine as an
anti-bleach agent. Controls for specificity of labeling in-
cluded omission of primary antibodies and labeling of
several nonauditory tissues.
Preparations were photographed with a Leitz Orthoplan
microscope equipped for epifluorescence with x 50 and
~ 1 0 0 oil objectives, or with a Bio-Rad laser confocal
microscope with an xl00 oil objective. Photography was
performed with Kodak T-max 400 film exposed at 1,600
Scanning electron microscopy (SEM)
Animals were fixed as described above for histochemistry
and then postfixed in 1% aqueous osmium tetroxide for 30
minutes. Wholemounts were processed using routine meth-
ods (Raphael, '92). Tissues were analyzed and photo-
graphed with an AMRAY lOOOB scanning electron micro-
scope operated at 10 kV.
The use of wholemounts for histochemical analysis en-
abled visualization of the BP in its entirety (Fig. 1A). Using
wholemounts with double- and triple-labeling proved to be
an excellent method for assessing the tissue organization,
distribution of junctional complexes, and location and
integrity of nuclei. In control tissues, F-actin (microfila-
ments) was present in the stereocilia and the cuticular plate
(terminal web) of every hair cell (Fig. lB), as previously
described (Baphael, '91). Hair cells were normally orga-
nized in rows with similar spaces between each other.
Co-localization of actin (Fig. 1B) and Hoechst (Fig. 1C) in
control tissues made it possible to locate the nucleus in
relation to the actin cytoskeleton of each cell. Normal hair
cells exhibited round nuclei organized in rows, visible at a
focal plane immediately beneath the luminal surface. Sup-
porting cell nuclei also labeled by Hoechst stain were
located at a deeper focal plane closer to the basement
membrane (not shown). Thus, nuclei of cells in the BP had
a laminar distribution, with hair cell nuclei immediately
beneath the reticular lamina (RL) and supporting cell
nuclei beneath them, closer to the basal lamina, as previ-
ously reported (Raphael, '92). Cingulin (Fig. lD), a protein
specific to tight-junctions (Citi et al., '88, '89), appeared as a
narrow line at the focal plane of the RL. Thus, cingulin-
specific label was found in homophilic (supporting cell-
supporting cell) and heterophilic (support cell-hair cell)
junctions at the RL. Because the apical surface area of
supporting cells was very slender, cingulin label often
appeared as a single line. Nevertheless, it should be noted
that cingulin was present all around the apical contour of
supporting cells. Therefore, the pattern of cingulin distribu-
tion was a good marker to delineate the polygonal apical
contours of hair cells and the narrow apical contour of
supporting cells (Figs. ID, 4A). In the polygon that sur-
rounds each hair cell, contact points between adjacent
supporting cells may be at the sharp corners, as depicted in
Figure 4A', or, alternatively, along the sides of the polygon.
Actin and cingulin were co-localized at the apical intercel-
lular contacts of' cells in the BP (Fig. 1E and D, respective-
ly). The pattern of cingulin and actin labels revealed similar
cellular organization of the BP in the proximal and distal
areas. The organization of the normal sensory mosaic of the
avian BP, shown with two fluorescent markers (Figs.
lD,E), is consistent with previous reports based on scan-
ning electron microscopy (Hirokawa, '78; Tanaka and
Smith, '78; Chandler, '84).
Structural alterations that accompany hair cell degenera-
tion were consistently observed in noise exposed chicks in a
region located 1.2 mm from the proximal end of the BP.
This region of damage is referred to as the "center of the
lesion" throughout this work (see Fig. 1A for location of
this region). Supporting cells in the center of the lesion
exhibited an expanded apical area, and occupied a signifi-
cant portion of the total surface of the RL, as previously
described (Cotanche, '87; Cotanche and Dopyera, '90; Marsh
Fig. 1. Wholemounts of control (unexposed) chick basilar papilla
viewed in phase-contrast microscopy (A) or fluorescence microscopy
with probes for actin (I3 and E), DNA (C), or cingulin (Dj. A The basilar
papilla and adjacent tissues are oriented with the proximal part (p) at
the bottom of the micrograph and the distal id) regions at the top left.
The neural and abneural margins of the basilar papilla are marked by
arrows. The abneural side is on the right. The center of the lesion (cl)
induced in this work (described in the following figures) is close to the
proximal region ofthe basilar papilla. The extreme proximal tip is not in
the photographed field. B,C: Actin label reveals hair cells in organized
rows (B) and Hoechst in same field shows hair cell nuclei (C). Arrows in
B and C point to the same cell though the focal plane is slightly
different. D,E: Co-localization of cingulin and actin. Cingulin expres-
sion (D) is restricted to tight junctions, connecting the apical surfaces of
all cells in the basilar papilla. Slender double lines (arrow) on sides of
hair cells delineate supporting cells that surround every hair cell,
forming a polygonal shape with 5 or 6 sides (h, apical surface of a hair
cell surrounded by 6 supportingcellsj. Actin label IE) is in the normally
packed bundles of stereocilia (sj and in the cuticular plate ic) of hair
cells ih marks the apical surface of the same hair cell in E and D). Actin
label is also in a belt of adherens junctions (arrow). Arrows in D and E
point to identical field, showing tight junctions around a supporting cell
in D and actin in adherens junctions of the same junctional complex in
E. Abneural side ofbasilar papilla faces top in B-E. Bars: 250 fim in A; 5
Fm in C and E.
Fig. 2. SEM (A, B) and fluorescence microscopy (C-Fj photomicro-
graphs oflesioned area in hasilar papilla of chicks exposed to noise for 4
hours. Inset in A shows the normal apical organization of thc basilar
papilla, in a control specimen. A: Immediately after exposure the apical
surfaces of hair cells arc smaller than normal, whereas supporting cells
are expanded and occupy more space than they normally do. Most hair
cells have a circular-shaped apical surface. The stereocilia appear
normal, and the ordcrly rows of hair cclls are maintained with no
apparent hair cell loss. A-inset: In control tissue, the apical mem-
branes of hair cells occupy most of the surface area of the basilar
papilla, whereas the apical extensions of supporting cells form straight
and ncrrow lines that appear as a polygon around hair cells. Note that
the distance between centers of stereocilia bundles on neighboring hair
cells is similar in exposed and control tissues. B: Six hours after noise
exposure several hair cells (h) still remain in the center of the lesion.
The apical surfaces of some hair cells [areas devoid of microvilli) are
severely constricted. One hair cell is completely obliterated by invading
supporting cells (arrowheads), so that only the stereocilia protrude
from the reticular lamina. C,D Co-localization of adin (focused at
stereocilia, in C) and Hoechst (focal plane of hair cell nuclei, in D),
showing hair cell loss and damaged hair cells (Cj and missingnuclei and
degenerating nuclei (arrow) (D) 6 hours after noise exposure. DNA
aggregates (arrowhead) arc scattered within the epithelium. E: Low
magnification of the lesion 6 hours after noise exposure reveals
extensive hair cell loss in the center of the lesion (curved arrow points to
the region with complete hair cell loss). In some areas in the lesion
several hair cells were present (open arrows) and in more remote
regions no hair cell loss is seen [arrow). White bars mark the abneural
border of the basilar papilla. F: Six hours after exposure, many hair cell
nuclei are missing in the lesioned area (curved arrow), whereas a
normal distribution of nuclei is observed in the more distal reb~on
(arrow). The abncural side of basilar papilla faces the bottom of the
micrographs. White bars mark the abneural border of the basilar
papilla. Bars: 5 pm in A (for A and inset), B, and D; 50 Fm in E and F.
REORGANIZATION IN TRAUMATIZED BASILAR PAPILLA
Fig. 3. Distal basilar papilla immediately after noise exposure (A-C)
or 6 hours after exposure (D), laheled for actin (A): cingulin (B, D). or
DKA (C). The abneural side of the basilar papilla faces left side in all
micrographs. A Actin is distrihuted normally in stereocilia and the
cuticular plate of hair cells. Actin in adherens junctions delineates the
borders of cells in the reticular lamina and reveals that the surface area
of hair cells is slightly reduced, whereas supporting cells in the lower
half of micrograph (arrows) are expanded (compare to control in Fig.
1E). There is a sharp transition between the region with expanded
supporting cells (in the lower part of the micrograph) and the undam-
aged region (in the upper part of the micrograph). Note that all hair
cells are present and bundles of stereocilia appear intact. B: Cingulin
distribution shows tight junctions at the reticular lamina. Hair cells (h)
are smaller than normal (compare to control in Fig. 1D). Many
supporting cells that border the sides of hair cells arc expanded and
appear as a "figure eight" (arrows! or as 2 rings sharing a common
middle line. Note that no hair cell loss is seen. C: The distribution of
Hoechst stain reveals no loss of hair cell nuclei. The orderly rows are
generally preserved, although 2 nuclei on the right side of micrograph
are slightly out of place. D: Cingulin expression 6 hours after exposure
shows more pronounced supporting cell expansion. Hair cell contours
are irregular due to supporting cell bulging. Areas of constriction
between the 2 halves of each figure eight appear. as lines of cingulin
(small arrow) or as crossover points (large arrow). Bars: 10 km in B (for
A and B); 10 km in D (for C and D).
et al., '90; Raphael, '92). The apical surface area of hair cells
was reduced, relative to normal, immediately after the
noise exposure (Fig. 2A, compare to control in inset), as
previously described by Marsh et al. ('90). In addition to
being constricted, the apical contour of hair cells in the
center of the lesion appeared oval or circular, instead of the
normally occurring polygonal contour. Stereocilia on hair
cells with constricted surface often appeared normal (Fig.
Six hours after noise exposure, hair cells in the center of
the lesion had very constricted apical surfaces and support-
ing cells were extremely expanded (Fig. 2B). Hair cells that
apical surface of supporting cell
B / B’/
Fig. 4. A schematic representation of the cellular elements that
construct the reticular lamina in control basilar papilla (A, A’) and in
noise exposed basilar papilla at the distal region (B, B’). A: The basal
pole of supportingcells is in contact with the basal lamina and the apical
pole extends to the RL. The apical membrane of each normal support-
ing cell appears like a slender linc. A’: Top view, showing the apical
surfaces of supporting cells that border a hair cell, together forming a
polygon. Six (or, occasionally, five) supporting cells construct the
polygon, but it is not clear whether contact points between these
were nearly completely engulfed by expanded supporting
cells could often be identified by the remaining stereocilia
that protruded from the RL (Fig. 2B). In many cases, all
5-6 supporting cells surrounding a hair cell were expanded
and appeared to invade the spaces normally occupied by
A substantial reduction in the number of nuclei in the
lesion was revealed by the distribution of Hoechst stain in
the lesion 6 hours after the noise exposure (Fig. 2D;
compared to Fig. 1C). This reduction was noted at the focal
plane of hair cell nuclei. In this area, actin distribution
revealed damage to hair cells, whereas Hoechst stain showed
degenerating nuclei as well as scattered DNA aggregates
(Fig, ZC,D), suggesting that cells were disintegrating in
situ. At this stage, phallaidin-labeled tissues revealed a
lesion with extensive hair cell loss (Fig. 2E). Inspection of
Hoechst-labeled BP at this stage revealed that in the area of
lesion, very few nuclei remained at the focal plane just
beneath the RL (Fig. 2F), confirming that hair cells were
The region referred to as “distal BP” throughout this
work was located distally to the center of the lesion,
approximately 2 inm from the proximal end of the BP. The
distal BP exhibited moderate structural alterations immedi-
ately after noise exposure. The distribution of actin in the
stereocilia and cuticular plate of hair cells was unchanged
in the distal BP after noise exposure (compare Fig. 3A to
Fig. 1E). However, the relative size of hair cells and
supporting cells was altered, as was the shape of supporting
cells (Fig. 3A,B). Cingulin-specific label at the apical con-
tour of supporting cells showed that these supporting cells
were expanded, whereas apical surfaces of the hair cells
were smaller than normal and irregularly shaped (compare
supporting cells are truly at the sharp corners, as described here B:
After noise exposure, the apical surface of supporting cells in the distal
basilar papilla expands and appears like a figure eight B’: Top view of
six supporting cells surrounding the apical surface of one hair cell after
noise exposure. The apical surface of hair cells is reduced compared to
normal The polygon shape is replaced by an irregular one. Noise-
induced changes in the basolateral domain of supportmg cells are not
depicted in this schematic
Fig. 3B to Fig. 1D). The most distal part of the BP did not
display morphological changes, and the border between
affected and unaffected regions in the distal BP was clearly
observed (Fig. 3A). The pattern of cingulin distribution in
the distal BP after noise exposure showed hair cells orga-
nized in rows as in the normal BP, and did not provide
evidence for hair cell loss. Based on the distribution of
cingulin, supporting cells in the distal BP appeared wider
than normal at their apical surfaces and in many cases the
contour of their apical membranes resembled “figure eights”
(Figs. 3B, 4B). The distribution of Hoechst in the distal BP
revealed that all hair cell nuclei were present, although
some nuclei were slightly dislocated relative to normal (Fig.
3C). Degenerating nuclei were not observed in the distal
BP. The distribution of cingulin in the distal BP 6 hours
after noise exposure (Fig. 3D) was similar to that observed
immediately after noise exposure (see Fig. 4 for schematic).
In some areas, the apical contour of each supporting cell
was composed of two rounded portions, forming a figure
eight. Hair cells were slightly constricted and their side
borders were irregularly shaped. Nevertheless, hair cell loss
was not observed in the distal BP.
In the center of the lesion, expanded supporting cells with
a thick cover of microvilli occupied a large region with
extensive hair cell loss 24 hours after noise exposure (Fig.
5A). A few hair cells with stereocilia bundles and a smooth
apical membrane survived the trauma and remained in the
region, among expanded supporting cells. In some cases,
the remains of hair cells were visible at the surface of the
epithelium. In contrast, no hair cell loss occurred in the
distal BP (Fig. 5B). Nevertheless, supporting cells in the
distal BP were slightly expanded and their apical surface
resembled a figure eight (compare Fig. 5B to Fig. 3D).
REORGANIZATION IN TRAUMATIZED BASILAR PAPILLA
Fig. 5. SEM of the center of lesion (A) and the &ski1 portion of
basilar papilla (B) 24 hours after noise exposure. A: Several surviving
hair cells (h) remain in the lesion, whereas numernus expanded
supporting cells cover an extensive region with hair cell loss. Stereocilia
of a hair cell in final stage of degeneration protrude from the surface
During all stages of hair cell degeneration, the confluence of
the RL was uninterrupted, and protrusion or ejection of
hair cell bodies from the RL was not observed.
Forty-eight hours after noise exposure, the center of thc
lesion contained a few hair cells with features of noise-
damaged cells. These hair cells had a large apical surface
and a well developed, although irregularly arranged, bundle
of stereocilia, suggesting that they survived the trauma
(Fig. 6A). These surviving (preexisting) hair cells werc
surrounded by supporting cells with expanded apical sur-
faces (Fig. 6A). At this time, mitotic chromosomes were
found in the lesion at a focal plane immediately beneath the
RL (Fig. 6B), as previously described (Raphael, '92). Cells
that contained mitotic chromosomes (Fig. 6C) exhibited
spherical cell bodies (Fig. 6D). However, the distribution of
actin in the junctional complexes at the level of the RL
directly above the dividing chromosomes did not disclose a
round or spherical cell contour (Fig. 6E), suggesting that
the apical surface of dividing cells is not necessarily round.
(arrow). B: No hair cell loss is seen in the distal basilar papilla, and
stereocilia on most cells are normal, though some bundles are slightly
damaged. Expanded supporting cells shaped as a figure eight (arrows)
are positioned between hair cclls. Bars: 2 pin.
Ninety-six hours after noise exposure, hair cells with a
small, embryonic-like apical surface were observed in the
center of the lesion (Figs. 7A, 8A) and in the distal BP (Figs.
7B, 8B). Small hair cells were considered to be new based on
their size and shape. Thus, cells with (apical) surface area
not exceeding half that of normal (mature) hair cells, and
immature organization of stereocilia, were considered to be
newly generated hair cells.
In the center of the lesion, expanded supporting cells
formed the immediate borders of new hair cells (Figs. 7A,
8A). In the distal BP (Figs. 7B, 8B), hair cell loss was not
evident, and new hair cells werc situated among preexisting
Most, but not all of the newly added hair cells were found
in pairs in the lesioned as well as the distal BP. As won as
they could be recognized as hair cells, regenerated hair cells
were appropriately oriented, with the kinocilia and the
tallest stereocilia at the abneural pole of the cell.
Fig. 6. Fluorescence micrographs of center of lesion in bwilar
papilla 48 hours after noise exposure. A: Actin label shows extensive
hair cell loss, although a few surviving hair cells (h) remain. B: DNA
label shows mitotic chromosomes (arrow) in lesion, at focal plane
immediately beneath reticular lamina. C,D: Mitotic chromosomes
(arrow in C) are enclosed by a belt of actin labcl (in D, showing t,he Same
region as C), showing rounded shape of dividing cell. E : The same
region shown in D and F, but at focal plane of the reticular lamina.
Several lines of actin traverse the region directly above the dividing
chromosomes. In the area where division occurs, no remaining hair
cells are seen. Arrows point to the same area in D and E. Bars: 10 km in
B; 10 bm in E (for A,C,D,E).
REORGANIZATION IN TRAUMATIZED BASILAR PAPILLA
Fig. 7. Actin distribution in the ccnter of lesion (A) and in the distal
basilar papilla (B) 96 hours after noisc exposure. A: Fluorescence
micrograph ofthe center of lesion where small hair cells with immature
appearance are found among expanded supporting cells and surviving
hair cells (h). Some immature hair cells appear in pairs (long arrows)
and others appear individually (short arrow). B: Laser confocal micro-
graph of the distal basilar papilla. A pair of small hair cells (arrows) is
present among normal looking hair cells in an area where no hair cell
loss is detected. Bars: 5 pm.
ate damaged hair cells and that the basolateral portion of
degenerating hair cells remains within the epithelium
(Raphael and Altschuler, '91). In contrast, dying hair cells
are ejected (or extruded) from the epithelium after severe
acoustic overstimulation in the chick BP (Cotanche et al.,
'87) and after laser microbeam irradiation in the lateral line
organ (Balak et al., '90). Irradiation may affect supporting
cell nuclei and abolish their ability to quickly expand and
replace hair cells. Nevertheless, in view of the present
results it appears that elimination of hair cells from the
epithelium can occur via two different routes: I) extrusion
and ejection to the luminal fluid or 2) obliteration and
degeneration of hair cells within the epithelial layer.
Future experiments will determine the causal relation-
ship between the extent of trauma and the route of hair cell
degeneration. I speculate that after severe trauma caused
by lengthy noise exposure, hair cells may be extruded,
whereas with moderate noise exposure hair cells degenerate
within the epithelium and are not ejected. In either case,
supporting cells appear to expand in a synchronized way,
preventing discontinuities in the RL.
Supporting cell expansion and hair
Expansion of supporting cells in the BP after noise
exposure was first described by Cotanche et al. ('87) and
later reported in additional works (Cotanche, '87; Marsh et
al., '90; Raphael and Altschuler, '92). The mechanism of
supporting cell expansion is still unknown. In this work,
supporting cells expanded not only in areas of hair cell loss
but also in remote areas without hair cell loss, suggesting
that hair cell loss is not a prerequisite for supporting cell
response to acoustic trauma.
The data presented here provide strong evidence that
hair cells may degenerate without extrusion into the lumi-
nal space. This evidence is based on two observations. First,
partly obliterated hair cells were observed protruding from
the center of a cluster of supporting cells, suggesting that
supporting cell expansion may trap hair cell bodies within
the epithelium. Second, small aggregates of DNA were
found in the region where hair cells degenerate after noise
exposure, strongly suggesting that cell death occurs within
the epithelium. Since supporting cells do not appear to
degenerate after noise exposure, aggregates of DNA are most
probably the remains of hair cell nuclei that disintegrate in situ.
Studies on hair cell degeneration in the organ of Corti
have also indicated that expanding supporting cells obliter-
Reorganization of apical junctions
A change in the apical contour of supporting cells in the
distal BP was invariably observed 4 hours after the onset of
noise exposure. In control tissues, the apical contour of
Fig. 8. SEM micrograph of the center of lesion (A) and the distal
basilar papilla (B) 96 hours after noise exposure. A A pair of small,
immature hair cells is located in an area with partial hair cell loss. Large
hair cells that survived the noise exposure are also seen (hl. The
immature hair cells are oriented with the tallest stereocilia and the
kinocilium facing the ahneural side of the basilar papilla. B: A pair of
supporting cells appeared slender, forming a polygon around
the hair cells. In contrast, supporting cells after trauma
appeared as a figure eight. The figure eight pattern was
evident with actin-specific or cingulin-specific labeling, as
well as with SEM. It is unlikely that new cells played a role
in altering the organization of the RL and forming the
figure eight contours, since figure eights could be detected
as early as 4 hours after the onset of noise exposure, a time
period too short for mitosis to occur.
Double-label experiments co-localized actin with cingulin
in the figure eight, indicating that elements of adherens
junctions (actin) and tight junctions (cingulin) participated
in creating this figure eight. It is likely, therefore, that
change in the apical contour of supporting cells is accompa-
nied by reorganization of the apical junctional complexes. I
speculate that the reorganization of actin bundles in sup-
porting cells represents a stress response of these cells,
caused directly or indirectly by overstimulation. The pur-
pose of this stress response could be to maintain the surface
tension of the epithelium at the RL, or to regulate support-
ing cell expansion and hair cell constriction. Based on this
tentative role, I propose to name the shape transition of
supporting cells “SAFE” (Stress-Associated Figure Eight).
immature hair cells (arrowheads) is located among mature hair cells.
Preexisting hair cells are regularly arranged and no missing cells can be
detected. New hair cells are oriented in register with other cells in area,
with taller rows of stereocilia facing ahneural side (lower left side of
micrograph). Bars: 5 pm.
Constriction of the apical surface of hair cells is likely to
result from active constriction of the circumferential actin
belt associated with the adherens junction. Constricted
actin belts have also been observed in outer hair cells in the
organ of Corti (Raphael and Altschuler. ’92). I speculate
that the force and speed of hair cell constriction after noise
may be directly related to the amount of mechanical stress,
which in turn is a function of the severity of the trauma.
Supporting cells attempt to restrain hair cell constriction,
but if the contractile force of the hair cell exceeds a critical
level, supporting cells expand and invade the space of the
constricting hair cell. In the distal BP, overstimulation is
less severe, contractile forces are weaker, and therefore
supporting cells are able to restrain hair cells and prevent
their contraction The role of the SAFE in this proposed
model would be to reinforce the apical domain of supporting
cells, to help restrain hair cell constriction and thereby
prevent hair cell loss.
Is hair cell death active?
Active cell death, also known as apoptosis, occurs as a
normal type of cell death in developing and mature tissues,
and may also result from trauma (Fawthrop et al., ’91;
REORGANIZATION IN TRAUMATIZED BASILAR PAPILLA
Gerschenson and Rotello, '92; Raff, '92). It is likely that
constriction of the apical actin belt in hair cells is an active
process, similar to constriction of the contractile ring that
separates two daughter cells from each other during cytoki-
nesis (Schroeder, '73). If true, the apical constriction in
damaged hair cells may constitute the first step in a cascade
of cellular activities leading to cell death. Thus, I raise the
possibility that hair cells are actively involved in their own
demise. A future goal will be to determine if degenerating
hair cells express any number of genes that encode apoptosis-
associated gene products.
Signals for supporting cell division and hair
A major finding in the present work is that new hair cells
may appear in distal areas of the BP, where hair cell
damage is minimal and hair cell loss could not be detected.
This finding raises the possibility that hair cell loss is not
necessary for production of new hair cells. It should be
noted that in this work, the lack of hair cell loss has been
determined by a subjective assessment based on the pres-
ence of orderly, uninterrupted rows of normal looking hair
cells. This subjective assessment cannot conclusively rule
out that single cells are eliminated unnoticed. Nevertheless,
it is rather unlikely that two or more neighboring hair cells
would be missing without being detected. Due to the
cellular organization at the RL, to completely deprive a
supporting cell of heterophilic contacts with hair cells, at
least two hair cells must be lost (see Corwin et al., '91, for
review). Thus, the data indicate that removal of lateral
inhibition is not necessary for producing new hair cells in
the traumatized BP. These findings corroborate previous
evidence that supporting cells that divide often maintain
contacts with preexisting hair cells (Raphael, '92).
Ionic leaks due to holes in the RL are also thought to
constitute a signal for regeneration after hair cell loss in the
BP (Corwin et al., '91). However, in this work the RL and
the network of tight junctions were structurally uninter-
rupted during hair cell degeneration and regeneration,
suggesting that changes in ionic composition do not play a
role in the signalling mechanism of regeneration. Neverthe-
less, a direct measurement of the functional integrity of the
RL as an ionic barrier should be performed before endolym-
phatic leak is unequivocally ruled out as a signal for
In conclusion, the present work demonstrated that after
intense noise exposure, damaged hair cells often display
constricted apical surface area and altered apical contour.
Dying hair cells may disintegrate within the epithelium
while the RL remains structurally undisturbed. Structural
changes in response to noise occur in supporting cells in the
distal BP, where hair cell damage is minimal and hair cell
loss is undetected. New hair cells are observed in the BP 4
days after acoustic trauma, in regions with or without
noticeable hair cell loss. Regenerated hair cells are appropri-
ately oriented in the BP as soon as they can be detected with
the present methods.
These findings emphasize the important role of support-
ing cells in the course of inner ear trauma. Based on the
present results, it is proposed that degenerating hair cells
may undergo active cell death. It is further speculated that
hair cell degeneration, supporting cell expansion, and hair
cell regeneration are mediated by tensile forces at the RL
and changes in cell shape. Experiments are now underway
to examine these hypotheses, and to determine whether
production of new hair cells in the center of the lesion is
regulated by the same mechanisms as in distal regions of
I offer many thanks to Peter Finger, Michael Lee, and Dr.
Yu Wang, who provided excellent technical support. I am
extremely grateful to Drs. Stephen Easter, Donna Martin,
Pamela Raymond, and Kathryn Tosney for their very
helpful comments on the manuscript. I thank Dr. Sandra
Citi (Cornell University Medical School, New York) for
kindly donating anti-cingulin antibodies. This work was
supported by a grant from the Deafness Research Founda-
Balak, K.J., J.T. Corwin, and J.E. Jones (1990) Regeneratcd hair cells can
originate from supporting cell progeny: Evidence from phototoxicity and
laser ablation experiments in the lateral line system. J. Neurosci.
Calof, A.L., A.D. Landner, and D.M. Chikaraishi (1991) Regulation of
neurogenesis and neuronal differentiation in primary and immortalized
cells from mouse olfactory epithelium. In Regeneration of Vertebrate
Sensory Receptor Cells. Ciba Foundation Symposium 160. New York:
John Wiley & Sons, p. 249-276.
Chandler, J.P. (1984) Light and electron microscopic studies of the basilar
papilla in the duck, Anus platyrhynchos. I. The hatchling. J. Comp.
Citi, S., H. Sabanay, R. Jakes, B. Geiger, and J. Kendrick-Jones (1988)
Cingulin, a new peripheral component of tight junctions. Nature 333272-
Citi, S., H. Sabanay, J. Kendrick-Jones, and B. Geiger (1989) Cingulin:
Characterization and localization. J. Cell Sci. 93:107-122.
Convin, J.T., and D.A. Cotanche (1988) Regeneration of sensory hair cells
after acoustic trauma. Science 240:1772-1774.
Corwin, J.T., J.E. Jones, A. Katayama, M.W. Kelley, and M.E. Warchol
(1991) Hair cell regeneration: The identities of progenitor cells, potential
triggers and instructive cues. In Regeneration of Vertebrate Sensory
Receptor Cells. Ciba Foundation Symposium 160. New York: John Wiley
&Sons, p. 103-130.
Cotanche, D.A. (1987) Regeneration of hair cell stereociliary bundles in the
chick cochlea following severe acoustic trauma. Hear. Res. 3Ot181-196.
Cotanche, D.A., and C.E.J. Dopyera (1990) Hair cell and supporting cell
response to acoustic trauma in the chick cochlea. Hear. Res. 46:29-40.
Cotanche, D.A.? J.C. Saunders, and L.G. Tihey (1987) Hair cell damage
produced by acoustic trauma in the chick cochlea. Hear. Res. 25267-
Fawthrop, D.J., A.R. Boobis, and D.S. Davies (1991) Mechanisms of cell
death. Arch. Toxicol. 65t437-444.
Gerschenson, L.E., and R.J. Rotello (1992) Apoptosis: A different type of cell
death. FASEB J. 6.2450-2455.
Girod, D.A., L.G. Duckert, and E.W Rubel (1989) Possible precursors of
regenerated hair cells in the avian cochlea following acoustic trauma.
Hear. Res. 42175-194.
Graziadei. P.P.. and G.A. Monti Graziadei 11985) Neurogenesls and plastic-
ity of the olfactory sensory neurons. Ann. N.Y. Acad. Sci. 457:127-142.
Hirokawa, N. (1978) The ultrastructure of the basilar papilla of the chick. J.
Comp. Neurol. 181t361-374.
Marsh, R.R., L. Xu, J.P. Moy, and J.C. Saunders (1990) Recovery of the
basilar papilla following intense sound exposure in the chick. Hear. Res.
Presson, J.C., and A.K. Popper (1990) Possible precursors to new hair cells,
support cells, and Schwann cells in the ear of a post-embryonic fish.
Hear. Res. 46t9-22.
Raff, M.C. (19921 Social controls on cell sunival and cell death. Nature
Raphael, Y. 11991) Damage to the tectorial membrane may protect chick hair
cells from noise overstimulation. Hear. Res. 53t173-184.
Raphael, Y. i1992: Evidence for supporting cell rnitusis in response i.o
acoustic trauma in the avian inner ear. J. Neurocytol. 21:663-671.
Raphael, y., and R,A. Altschuler ( 1991) Modulation of cytoskeletal and
Reh, T.H., M. Jones, and C Pittack (1991) Common mechanisms of retinal
regeneration in the larval frog and embryonic chick. In Regeneration of
Vertebrate Sensory Receptor Cells. Ciba Foundation Symposium ~ 0 .
New York: John Wiley & Sons, p. 192-207.
Ryals, B.M., and E.W Rube1 (1988) Hair cell regeneration after acoustic
Schroeder, T.E. (1973) Actin in dividingcells: Contractile ring filaments bind
heavy meromyosin. Proc. Natl. Acad. Sci. USA 7Or1688-1692.
T ~ ~ & ~ ,
K., and C.A. Smith (1978) Structure ofthe chicken's inner Far: SEM
and'I'EM study. Am. d. Anat. 153:251-271.
junctionalproteins during hair cell degeneration. Ccll Motil. 18:215-227.
Raphael, Y., and R.A. Altschuler (1992) Early microfilament reorganization
in injured auditory epithelia. Exp. Neurol. 11532-36.
Raymond, P.A. (1991) Retinal regeneration in teleost fish. In Regeneration
of Vertebrate Sensory Receptor Cells. Ciba Foundation Symposium 160.
NewYork: John Wiley& Sons,p. 171-191.