- Email: [email protected]
Micromechanical properties of sensory hairs on receptor cells of the inner ear
Hearing Research, Elsevier
11(1983) 249-260
249
Micromechanical properties of sensory hairs on receptor cells of the inner ear * Ake Flock and Steven Orman Department
**
of Physiology II, Karolinska Instituter, 104 01 Stockholm, Sweden (Received
28 February
1983; accepted
16 March
1983)
The aim of the investigation was to obtain quantitative measures of mechanical filter properties of hair cell sensory hairs which were not restricted to move by auxiliary structures. The specimen used was the crista ampullaris of the frog, dissected free and positioned in a fluid-filled chamber where it could be viewed with differential interference contrast optics. The sensory hairs were displaced by a brief jet of frog Ringer’s solution from a specially constructed microsyringe apparatus. The velocity of the jet could be stepwise controlled and was determined by measuring the speed of motion of 3 gm plastic beads propelled with the jet across the microscopic field. The stereocilia displacements were recorded on 16 mm film, and both the angle of deflection and the time for the return to the resting position were measured on the film. It was found that after displacement the sensory hairs returned to the upright position by elastic properties in the hinge region at their insertion point in the cuticular plate. Sensory hairs differed in their speed of return so that some had fast time constants, others quite slow ones. This was correlated to a difference in structural development of the stereocilia, fast sensory hairs having thick and tall stereocilia, slow ones having thin and short stereocilia. The various bundle types were identified in the scanning electron microscope and their distribution on the crista was mapped. This was found to match the distribution of nerve fibres with different functional properties. It is concluded that sensory hair cells can differ in their mechanical filter properties as a result of the structural arrangement of their stereocilia and in accordance with functional demands. Key words:
inner ear; stereocilia;
frequency
selectivity.
Introduction The sensory cells of the inner ear operate as mechanoreceptors through sensory hairs which extend like microlevers from their surface. Each cell is equipped with a large number of rod-shaped so-called stereocilia collected into a bundle. In the vestibular system they are joined by a single kinocilium occupying one side of the bundle [26], but it appears that the stereocilia are the excitable structures [ 121.
A preliminary account of this work was presented for Research in Otolaryngology [ 181. ** Present address: Department of Neurophysiology, U.S.A. l
0378-5955/83/%03.00
at the 1981 Midwinter
0 1983 Elsevier Science Publishers
University
B.V.
Meeting
of Wisconsin,
of the Association
Madison,
WI 53706,
250
In most inner-ear sense organs the sensory hairs are coupled at their tips to auxiliary structures which provide the mechanical input. In such cases the time pattern of motion of the sensory hairs will reflect the motion pattern of the auxiliary structure more or less faithfully depending on the degree of coupling. The coupling varies from one type of sense organ to another, and can also vary in different regions within one and the same organ. The most striking example is perhaps the basilar papilla in the lizard which is divided into two adjacent regions; in one of these the sensory hairs are tightly coupled to a tectorial membrane, in the other region the tectorial membrane is missing and the sensory hairs are free-standing [27]. In the case of free loosely coupled sensory hairs the mechanical response properties of the individual sensory hairs would be expected to influence the transduction of mechanical stimuli into biological signals as seen in sensory cell and nerve fibre responses. The aim of this work has been to investigate the micromechanical properties inherent in the sensory hairs of individual receptor cells unrestrained by auxiliary structures. The organ selected was the crista ampullaris of the semicircular canal in the frog. The sensory epithelium is draped over a thin connective tissue ridge spanning the floor of the ampulla. The roof of the ampulla can be opened, the cupula gently removed, and the crista can be viewed in profile with the sensory hairs accessible for micromanipulation [5]. Individual sensory hairs can be observed with differential interference contrast optics and their motion pattern in response to calibrated stimuli can be filmed and analyzed. An abundant literature describes regional variations in hair bundle structure in vestibular sensory epithelia [13,14,16]. In this study we have noted structural variations among the hair bundles investigated mechanically. Scanning electron microscopy was therefore used to characterize different types of hair bundles. This allowed the study of a possible correlation between mechanical properties and sensory hair structure, a correlation that was indeed found to exist [ 181. Furthermore, it was found by mapping the regional distribution of different types of hair bundles over the sensory epithelium that the distribution of hair bundles with fast and slow time constants matched the distribution of nerve fibres characterized by their electrophysiological response [ 111. Although dealing with a vestibular end organ the general question of sensory hairs as mechanical filters is pertinent in cochlear physiology, hence the submission of this work to Hearing Research. A second reason is that this is the first in a series of papers on sensory hair micromechanics that leads to work on the mammalian organ of Corti [22].
Methods The preparation
The preparation used has been briefly described elsewhere [S]. Frogs of the species Rana temporaria were decapitated and the posterior semicircular canal exposed by a ventral approach. The canal was gently tied off distal to the ampulla to
251
Fig. 1.The excised crista ampullaris is mounted between two platinum holders (h) so that it can be viewed in profile. The sensory hairs that arise from the sensory epithelium (se) are deflected by a jet from a micropipette (mp).
prevent excessive motion of the cupula during further dissection. The canal was cut distal to the tie and proximal to the ampulla. The segment was transferred to a chamber where, in reflected light, the roof of the ampulla was opened, the cupula removed and the crista freed from surrounding tissue. The dissected crista is held by a pair of platinum wires positioned in a Ringer’s solution-filled chamber, so that the edge of the crista can be viewed in profile by interference contrast optics according to Nomarski. The long working distance of the 40 X water-immersion lens (1.6 mm) allows placement of a miniature pipette beneath the objective lens close to the ridge of the crista (Fig. 1). Hair defection system The micropipette is coupled by a pressure-tight lucite holder to a micro-syringe (Hamilton, 25 ~1) and the micro-syringe plunger is coupled to a step motor microdrive (Transvertex, Sweden). The whole assembly is attached to a micromanipulator which is used to position the micropipette tip in close approximation to a stereociliary bundle and bring it into the same plane of focus. The micropipette tip was carefully cut to 50 pm under a 40 x dissecting microscope, gently scratching it with a diamond glass cutter and slightly pushing the unwanted portion. This produced a smooth, flat opening in the tip. The tip was then bent, using a heated platinum wire, to an angle of 45”. This bend placed the micropipette tip parallel to the surface of the experimental chamber and directed the stream across the plane of focus during the experiment. The microdrive was externally triggered by a digital pulse generator (Devices, U.K.) so that a predetermined number of advances (steps) of the microdrive armature could be initiated by pushing a button on a hand-held control unit. A single forward step of the microdrive depressed the syringe plunger 2 pm, this delivered 0.83 nl to the micropipette, ejected as a jet directed at stereocilia. The volume ejected could be varied by preselecting the number of steps from 1 to 10, usually delivered with 5 ms interval.
252
Visualization and calibration of flow
Since our objective was to deliver a calibrated pulse of fluid to the hairs, careful attention was paid to determining the shape of the fluid stream emerging from the micropipette and the fluid stream’s velocity profile. The velocity profile of the jet was determined, by measuring the motion of l-3 pm plastic beads (Affi-Gel; Richmond, Calif.) propelled with the jet across the microscopic field. The micropipette was filled with frog’s Ringer’s solution containing a small quantity of the plastic beads. These were sufficiently small to flow freely from the pipette with the jet of fluid. The movement of the beads across the microscopic field was recorded on 16 mm film for each jet intensity (1 to 10 steps on the microdrive). The position of individual beads was plotted frame by frame to visualize their trajectories as illustrated in Fig. 2 for three beads exiting almost simultaneously from the pipette opening. Flow velocities could thus be determined and were found to decrease with distance from the opening. The jet was essentially laminar and turbulence was negligible. During several calibration experiments the flow profile was simultaneously visualized by adding 1% sucrose to the Ringer’s/plastic bead solution. Using a dark Nomarski field the ejected Ringer’s sucrose solution appeared as a bright glow, revealing the shape of the ejected flow. The beads and the ejected fluid were found to travel in company. It was ascertained in this fashion that micropipettes selected for use in hair deflection experiments provided jets of appropriate width and velocity at the point where a selected sensory hair was to be positioned. Measurement
of hair bundle deflection
The motion pattern of sensory hairs in response to a fluid jet was recorded on 16 mm film running at 25 or 50 frames/s. A light-emitting diode synchronized to the pulse generator produced a light flash on the edge of the film at the initiation of a fluid squirt. The filmed sequence of hair deflections was analyzed by projecting single frames
1
,
.
2
. + . . - v . . 1. . ‘.
:
;
r...
y.*=...,
.
.
.
.
Fig. 2. Three plastic beads (0 v H) were ejected simultaneously from the micropipette opening indicated at the left. Their trajectories were filmed and plotted frame by frame. At arrow 1 the velocity is 240 pm/s. at arrow 2 it is 120 pm/s.
253
on the screen of a viewer (Recordak Magnaprint Reader, Kodak). A transparent graded angle protractor was fixed to the front of the screen so that the protractor appeared superimposed over the image of the hair. For the sensory hair being examined, the protractor was aligned prior to the squirt (the so-called resting position) so that the bundle axis lay at 0”. This axis was easily determined since the Nomarski optics produced a distinct straight-edge shadow along the center of each bundle. Measurement of the angle of deflection then proceeded frame by frame, plotting hair motion at 40 or 20 ms intervals. Scanning electron microscopy The ampulla of the posterior semicircular canal was exposed and fixed in situ in a fixative composed as follows: 2% glutaraldehyde + 4% paraformaldehyde + 2 mM MgCl, in 0.1 M sodium phosphate buffer at pH 6.2. After 30 min the ampulla was separated and immersed in the fixative for another 30 min and was then post-fixed for 1 h in 2% 0~0, in the same buffer and rinsed in distilled water. The crista was dissected free as described above, treated by the OTO-method [ 141 and critical point dried. The specimen was covered with gold by sputter coating (Baltzer) and examined in a Cambridge Stereoscan scanning electron microscope. Results General features The stimulus is defined as the fluid pulse directed at a selected sensory hair from the micropipette, and the response-is the resulting deflection as measured from the film recording. In calibration tests with plastic beads it was found that the velocity of the stimulus jet for a given volume ejected was highly reproducible between repeated pulses. This was confirmed in the preparation with the sensory hairs as indicators of fluid displacement. All hairs within the jet deflect in unison and then return to the resting position. Because the jet expands with distance from the pipette, hairs at the periphery of the jet may exhibit a vertical component of deflection taking their tip out of focus. Such behaviour was avoided by vertical alignment of the center of the pipette to the same focal plane as a selected hair bundle. This procedure ensured that the sensory hair was located in the most laminar part of the
400
Fig. 3. The angle of deflection
800 msec
of a sensory
hair bundle
subjected
twice to identical
stimuli.
254
-
1 step
-
3 step
-
5 step
msec
Fig. 4. Plot of the deflection
of a sensory
hair in response
to three different
velocities
of the jet
flow. For a 50 pm wide pipette the ejected fluid jet was wide enough to easily aim at and fully envelop a hair bundle selected for observation. Fig. 3 shows the time course of deflection of a hair displaced twice by repeated fluid pulses. The accuracy of the measurement of deflection is about 0.5’ and the reproducibility is of approximately the same order. In response to the fluid displacement the hair bundle deflects as a stiff rod hinged at the base. All the stereocilia join the motion as if connected to one another. In Fig. 4 a family of deflection curves for three different pulse velocities are plotted. The angle of deflection reaches a maximum and then returns to the original position. For increasing velocities the angle of deflection increases as seen in Fig. 5. For large displacements there is a
2
3
4
5
step
Fig. 5. The angle of maximum deflection was measured The mean and the standard deviation are plotted.
for three velocities
of the Jet in 7 sensory
hairs.
255
tendency for the hair to return with a higher velocity than at small displacements as seen from the slope of the return phase. Hence, the elastic restoring force increases with deflection of the hair. Time course of sensory hair motion The time course of the return phase varied for different sensory hair bundles as illustrated with the examples given in Fig. 6: (A) Some bundles returned rapidly to the resting position. (B) Other bundles returned more slowly to the resting position. They also tended to reach maximum excursion later. (C) A third type of response was a large initial deflection followed by a slow return. When the time constants for different bundles (displaced by equal velocity pulses) are compared one finds a clear separation into two groups, one below and one above 760 ms. Among the faster time constants there is no clearcut division but a smooth transition from a fast group centered about 80 ms and a population of intermediate time constants. Correlation of sensory hair stiffness to architecture of stereociiium bundle Bundles with slow and fast return can be distinguished in the microscope during the experiment and may be found side by side in the same preparation. Also, slow return can be seen in a bundle placed between two fast ones. Therefore, the observed variability does not depend on differences between preparations or differences in the position of a particular bundle in the jet. Rather, we have observed a correlation between the time constant of a given bundle and its structure in terms of length,
25
400
600
msec Fig. 6. The time pattern of motion of three different hair bundles.
800
1000
256
number and diameter of the stereocilia within the bundle and the relative length of the kinocilium as compared to the stereocilia. Response types A, B and C of Fig. 6 correlate to sensory hairs with the general structural characteristics seen in Fig. 7: (A) The sensory hair bundle is about 5 pm wide at the base and tapers gradually
8A
B
C
Fig. 7. Microphotographs of sensory hair bundles development of stereocilia are seen. Fig. 8. Scanning electron matching Fig. 7.
micrographs
observed
of hair bundles
in the experimental
showing differences
microscope.
in architecture
Differences
in
of the stereocilia
251
towards the tip. The stereocilia within the bundle can be discerned and reach to the tip of the bundle. (B) This type of bundle is somewhat more slender with a width at the base of about 4 pm, and there is a gradual taper to the tip. The stereocilia are more difficult to discern and seem to end below the tip of the hair. (C) The sensory hair bundles have a small basal bundle of stereocilia and a tall solitary kinocilium. On the basis of these results scanning electron microscopy was performed on the crista ampullaris in order to search for and examine in closer detail the corresponding types of sensory hairs. Scanning electron micrographs of such sensory hairs are shown in Fig. 8 and are described below: (A) The stereocilia are massive with a width of about 0.4 pm and insert in a broad cuticular plate. The tallest stereocilia accompany the kinocilium up to just beneath the tip. (B) The total width of the stereociliary bundle is smaller and the diameter of the individual stereocilia is less, measuring about 0.3 pm. The tallest stereocilia end about halfway to the tip of the kinocilium. (C) This type of bundle has a solitary kinocilium joined at its base by a group of short stereocilia having a diameter of about 0.15 pm. The cuticular plate has a correspondingly small diameter. Regional distribution of different types of bundles The distribution of hair bundles type A, B and C is shown in Fig. 9. Fast response type A bundles are present in the central part of the crista only. Medium fast type B bundles dominate the lateral planum region and are also scattered in the center. Slow type C bundles are present mainly along the edges of the sensory epithelium with a few found higher up on the ridge centrally in the crista.
Fig. 9. Distribution of sensory hairs type A, B and C.
258
Discussion The stiffness of vestibular sensory hairs was first noted by Schultze [21] and Retzius [20] but was forgotten, probably because microdissection of fresh tissue was abandoned when fixation, embedding and sectioning techniques were introduced around the turn of the century. It is now known that fixation of the tissue results in loss of sensory hair stiffness unless special fixation techniques are used [5]. The stereocilium stiffness relies on an inside core of actin filaments [3] probably cross-linked by the protein fimbrin [6]. A similar structure of the core has been reported in the lizard basilar papilla where a paracrystalline packing of the actin filaments was noted [23]. Such a core implies nonelastic properties, and the elastic component responsible for the return of the sensory hair has to be looked for elsewhere. Restoring force could be supplied by the network of fibrillar material that intertwines the ciliary bundles. A more important site may be at the tapered neck of the stereocilia where the actin filaments are tightly joined and penetrate as rootlets into the cuticular plate, a tight mesh of actin filaments beneath the apical membrane. The present work shows that the micromechanical behaviour of a sensory hair is closely related to the architecture of the stereocilia of which it is composed. Ranging from fast to slow the sensory hairs would in physiological terms be described as dynamic rapidly adapting velocity sensitive units, or static slowly adapting displacement sensitive units. This differentiation of mechanical response patterns may derive ‘passively’ from the structural differentiation of the sensory hairs or it could rely on act:ve mechanisms related to those which cause ciliary beat or muscle contraction. Invoked in the latter case is a dependence on membrane related ionic mechanisms. The mechanical properties or the sensory hairs described here do not, however, appear to depend critically on an intact membrane [4] and are therefore more likely to be built into hair structure than being derived from active mechanisms. The different sensory hair response types would provide individual filtering characteristics for individual hair cells, if allowed to become expressed in the physiological situation. This will depend on the physical nature of the coupling between sensory hairs and the material of the cupula, a question which is under discussion [2,9,14]. In the vestibular system differences in firing patterns, particularly in adaptive behaviour, among nerve fibres innervating the crista ampullaris have been described by several authors [8,17,19]. There are two types of hair cells in the mammalian crista ampullaris which differ significantly in structure and innervation patterns [26], a fact that has been suggested as a basis for differences in nerve response properties. Lim [14] has noted that Type I and Type II hair cells in the mammalian vestibular system also have different bundle structures. Hair cells in the frog crista ampullaris are all Type II according to innervation patterns. but they can be subdivided into at least two groups according to sensory hair structure ([9], and the present study). We find that the different bundle types are differentially distributed over the crista, fast thick bundles in the centre, slow slender bundles on the two sides. Highly significant in this context is the work of Honrubia et al. [ 1 l] correlating anatomical and physiological characteristics of primary afferents in the frog crista ampullaris. They found that neurons with an evoked impulse discharge of
259
short time constant innervate the central part of the crista whereas neurons with slow time constant characteristics as a whole tended to innervate peripheral regions. The positive correlation between hair bundle mechanical response and nerve discharge properties does not prove a causative relationship but shows that receptors and neurons are tuned in harmony. The concept of sensory hairs acting as mechanical filters does have implications for other inner ear sense organs. For example, in the basilar papilla of the lizard only hair cells in a particular region are covered by a tectorial membrane, the remaining being free-standing in the endolymph [27]. The free-standing sensory hairs vary in length along the papilla and there is a corresponding variation in mechanical tuning of the sensory hairs [7,10] and best frequency of neural responses [24,25]. A causal relationship has therefore been suggested. Also in the mammalian organ of Corti there is a correlation between the length of outer hair cell stereocilia [ 151 and sensory hair stiffness [22], indicating a general contribution of sensory hair mechanical properties to the tuning of the integrated organ. Also, there are indications that inner and outer hair cells differ in their coupling to the tectorial membrane [ 11, inner hair cells being more loosely coupled. The physiological response of the inner hair cell should then be influenced by the mechanics of its stereocilia. The present results show that the physiological response characteristics of a hair cell can be dictated by the architecture of its sensory hair bundle. These filter characteristics are passive, acquired during evolution. There is indication that the mechanical characteristics of the hair cell can also be actively controlled by processes akin to those that operate in muscle contraction. This is dealt with in an accompanying publication [4].
Acknowledgements This work was supported by grants from the Swedish Medical Research (04X-02461), the Soderberg’s Foundation and the Tysta Skolan Foundation.
Council
References 1 Dallos, P. (1981): Cochlear physiology. Annu. Rev. Physiol. 32, 153- 190. 2 Dohlman, G.F. (1980): Critical review of the concept of cupula function. Acta Otolaryngol. Suppl. 376, l-30. 3 Flock, A. and Cheung, H. (1977): Actin filaments in sensory hairs of the inner ear receptor cells. J. Cell Biol. 75, 339-343. 4 Flock, A. and Orman, S. (1983): Active control of sensory hair mechanics implied by susceptibility to media that induce contraction in a muscle fibre. Hearing Res. 11, 261-266. 5 Flock, A., Flock, B. and Murray, E. (1977): Studies on the sensory hairs of receptor cells in the inner ear. Acta Otolaryngol. 83, 85-91. 6 Flock, A., Bretscher, A. and Weber, K. (1982): Immunohistochemical localization of several cytoskeletal proteins in inner ear sensory and supporting cells. Hearing Res. 6, 75-89. 7 Frischkopf, L.S., De Rosier, D.J. and Engelman, E.H. (1982): Motion of basilar papilla and hair cell stereocilia in the excised cochlea of the alligator lizard: relation to frequency analysis. Sot. Neurosci. Abstr. 8, 40.
260
8 Goldberg, J.M. and Fernandez, C. (197 1): Physiology of peripheral neurons innervating semicircular canals of the squirrel monkey. I. Resting discharge and response to constant angular accelerations. J. Neurophysiol. 34, 635-684. 9 Hillman, D.E. (1974): Cupular structure and its receptor relationship. Brain Behav. Evol. 10, 52-68. IO Holton, T. and Hudspeth, A.J. (1982): Motion of hair cell stereocilia in the auditory receptor organ of the alligator lizard. Sot. Neurosci. Abstr. 8, 40. 11 Honrubia, V., Sitko, S., I&mm, J., Betts, W. and Schwartz, I. (1981): Physiological and anatomical characteristics of primary vestibular afferent neurons in the bullfrog. Int. J. Neurosci. 15, 197-206. 12 Hudspeth, A.J. and Jacobs, R. (1979): Stereocilia mediate transduction in vertebrate hair cells. Proc. Natl. Acad. Sci. U.S.A. 76, 1506- 1509. 13 Lewis, E.R. and Li, C.W. (1975): Hair cell types and distributions in the otolithic and auditory organs of the bullfrog. Brain Res. 83, 35-50. 14 Lim, D.J. (1979): Fine morphology of the otoconial membrane and its relationship to the sensory epithelium. SEM 1979 III, IIT, pp. 929-938. SEM Inc., AMF O’Hare, IL 60666. 15 Lim, D.J. (1980): Cochlear anatomy related to cochlear micromechanics. A review. J. Acoust. Sot. Am. 67, 1686-1695. 16 Lowenstein, O., Osborne, M.P. and Wersall, J. (1964): Structure and innervation of the sensory epithelia of the labyrinth in the Thornback ray (Raja clauara). Proc. Roy. Sot. Lond. Ser. B, 160, l-12. 17 O’Leary, D.P., Dunn, R. and Honrubia. V. (1974): Functional and anatomical correlation of afferent responses from the isolated semicircular canal. Nature (London) 25 1, 225-227. 18 Orman, S. and Flock, A. (1981a): Micromechanics of the hair cell stereociliary bundles in the frog crista ampullaris. Ass. Res. Otolaryngol. Abstr. 4. 30-31. 19 Precht, W., Llinas, R. and Clarke, M. (1971): Physiological responses of frog vestibular fibers to horizontal angular rotation. Exp. Brain Res. 13, 378-407. 20 Retzius, G. (1884): Das Gehororgan der Wirbeltiere. I. Fische und Amphibien. Sanson and Wallin. Stockholm. 21 Schultze, M. (1858): Ueber dei Endigungsweise der Hornerven im Labyrinth. Arch. Anat. Physiol. Wiss. Med. 343-381. 22 Strelioff, D. and Flock, A. (1982): Mechanical properties of hair bundles of receptor ceils in the guinea pig cochlea. Sot. Neurosci. Abstr. 8, 40. 23 Tilney, L.G., De Rosier, D.J. and Mulroy. M.J. (1980): The organization of actin filaments in the stereocilia of cochlear hair cells. J. Cell Biol. 86, 244259. 24 Turner, R.G., Muraski, A. and Nielsen, D.W. (1981): Cilium length: Influence on neural tonotopic organization. Science 213, 1519- 1521. 25 Weiss, F.E., Peake, W.T., Ling, A. and Holton, T. (1978): Which structures determine frequency selectivity and tonotopic organization of vertebrate cochlear nerve fibres? Evidence from alligator lizard. In: Evoked Electrical Activity in the Auditory Nervous System, pp. 91-l 12. Editors: R. Naunton and C. Fernandez. Academic Press, New York. 26 Werslll, J. (1956): Studies on the structure and innervation of the sensory epithelium of the cristae ampullares in the guinea pig. Acta Otolaryngol. Suppl. 126. l-85. 27 Wever, E.G. (1967): The tectorial membrane of the lizard ear. Types of structure. J. Morphol. 122, 307-320.
11(1983) 249-260
249
Micromechanical properties of sensory hairs on receptor cells of the inner ear * Ake Flock and Steven Orman Department
**
of Physiology II, Karolinska Instituter, 104 01 Stockholm, Sweden (Received
28 February
1983; accepted
16 March
1983)
The aim of the investigation was to obtain quantitative measures of mechanical filter properties of hair cell sensory hairs which were not restricted to move by auxiliary structures. The specimen used was the crista ampullaris of the frog, dissected free and positioned in a fluid-filled chamber where it could be viewed with differential interference contrast optics. The sensory hairs were displaced by a brief jet of frog Ringer’s solution from a specially constructed microsyringe apparatus. The velocity of the jet could be stepwise controlled and was determined by measuring the speed of motion of 3 gm plastic beads propelled with the jet across the microscopic field. The stereocilia displacements were recorded on 16 mm film, and both the angle of deflection and the time for the return to the resting position were measured on the film. It was found that after displacement the sensory hairs returned to the upright position by elastic properties in the hinge region at their insertion point in the cuticular plate. Sensory hairs differed in their speed of return so that some had fast time constants, others quite slow ones. This was correlated to a difference in structural development of the stereocilia, fast sensory hairs having thick and tall stereocilia, slow ones having thin and short stereocilia. The various bundle types were identified in the scanning electron microscope and their distribution on the crista was mapped. This was found to match the distribution of nerve fibres with different functional properties. It is concluded that sensory hair cells can differ in their mechanical filter properties as a result of the structural arrangement of their stereocilia and in accordance with functional demands. Key words:
inner ear; stereocilia;
frequency
selectivity.
Introduction The sensory cells of the inner ear operate as mechanoreceptors through sensory hairs which extend like microlevers from their surface. Each cell is equipped with a large number of rod-shaped so-called stereocilia collected into a bundle. In the vestibular system they are joined by a single kinocilium occupying one side of the bundle [26], but it appears that the stereocilia are the excitable structures [ 121.
A preliminary account of this work was presented for Research in Otolaryngology [ 181. ** Present address: Department of Neurophysiology, U.S.A. l
0378-5955/83/%03.00
at the 1981 Midwinter
0 1983 Elsevier Science Publishers
University
B.V.
Meeting
of Wisconsin,
of the Association
Madison,
WI 53706,
250
In most inner-ear sense organs the sensory hairs are coupled at their tips to auxiliary structures which provide the mechanical input. In such cases the time pattern of motion of the sensory hairs will reflect the motion pattern of the auxiliary structure more or less faithfully depending on the degree of coupling. The coupling varies from one type of sense organ to another, and can also vary in different regions within one and the same organ. The most striking example is perhaps the basilar papilla in the lizard which is divided into two adjacent regions; in one of these the sensory hairs are tightly coupled to a tectorial membrane, in the other region the tectorial membrane is missing and the sensory hairs are free-standing [27]. In the case of free loosely coupled sensory hairs the mechanical response properties of the individual sensory hairs would be expected to influence the transduction of mechanical stimuli into biological signals as seen in sensory cell and nerve fibre responses. The aim of this work has been to investigate the micromechanical properties inherent in the sensory hairs of individual receptor cells unrestrained by auxiliary structures. The organ selected was the crista ampullaris of the semicircular canal in the frog. The sensory epithelium is draped over a thin connective tissue ridge spanning the floor of the ampulla. The roof of the ampulla can be opened, the cupula gently removed, and the crista can be viewed in profile with the sensory hairs accessible for micromanipulation [5]. Individual sensory hairs can be observed with differential interference contrast optics and their motion pattern in response to calibrated stimuli can be filmed and analyzed. An abundant literature describes regional variations in hair bundle structure in vestibular sensory epithelia [13,14,16]. In this study we have noted structural variations among the hair bundles investigated mechanically. Scanning electron microscopy was therefore used to characterize different types of hair bundles. This allowed the study of a possible correlation between mechanical properties and sensory hair structure, a correlation that was indeed found to exist [ 181. Furthermore, it was found by mapping the regional distribution of different types of hair bundles over the sensory epithelium that the distribution of hair bundles with fast and slow time constants matched the distribution of nerve fibres characterized by their electrophysiological response [ 111. Although dealing with a vestibular end organ the general question of sensory hairs as mechanical filters is pertinent in cochlear physiology, hence the submission of this work to Hearing Research. A second reason is that this is the first in a series of papers on sensory hair micromechanics that leads to work on the mammalian organ of Corti [22].
Methods The preparation
The preparation used has been briefly described elsewhere [S]. Frogs of the species Rana temporaria were decapitated and the posterior semicircular canal exposed by a ventral approach. The canal was gently tied off distal to the ampulla to
251
Fig. 1.The excised crista ampullaris is mounted between two platinum holders (h) so that it can be viewed in profile. The sensory hairs that arise from the sensory epithelium (se) are deflected by a jet from a micropipette (mp).
prevent excessive motion of the cupula during further dissection. The canal was cut distal to the tie and proximal to the ampulla. The segment was transferred to a chamber where, in reflected light, the roof of the ampulla was opened, the cupula removed and the crista freed from surrounding tissue. The dissected crista is held by a pair of platinum wires positioned in a Ringer’s solution-filled chamber, so that the edge of the crista can be viewed in profile by interference contrast optics according to Nomarski. The long working distance of the 40 X water-immersion lens (1.6 mm) allows placement of a miniature pipette beneath the objective lens close to the ridge of the crista (Fig. 1). Hair defection system The micropipette is coupled by a pressure-tight lucite holder to a micro-syringe (Hamilton, 25 ~1) and the micro-syringe plunger is coupled to a step motor microdrive (Transvertex, Sweden). The whole assembly is attached to a micromanipulator which is used to position the micropipette tip in close approximation to a stereociliary bundle and bring it into the same plane of focus. The micropipette tip was carefully cut to 50 pm under a 40 x dissecting microscope, gently scratching it with a diamond glass cutter and slightly pushing the unwanted portion. This produced a smooth, flat opening in the tip. The tip was then bent, using a heated platinum wire, to an angle of 45”. This bend placed the micropipette tip parallel to the surface of the experimental chamber and directed the stream across the plane of focus during the experiment. The microdrive was externally triggered by a digital pulse generator (Devices, U.K.) so that a predetermined number of advances (steps) of the microdrive armature could be initiated by pushing a button on a hand-held control unit. A single forward step of the microdrive depressed the syringe plunger 2 pm, this delivered 0.83 nl to the micropipette, ejected as a jet directed at stereocilia. The volume ejected could be varied by preselecting the number of steps from 1 to 10, usually delivered with 5 ms interval.
252
Visualization and calibration of flow
Since our objective was to deliver a calibrated pulse of fluid to the hairs, careful attention was paid to determining the shape of the fluid stream emerging from the micropipette and the fluid stream’s velocity profile. The velocity profile of the jet was determined, by measuring the motion of l-3 pm plastic beads (Affi-Gel; Richmond, Calif.) propelled with the jet across the microscopic field. The micropipette was filled with frog’s Ringer’s solution containing a small quantity of the plastic beads. These were sufficiently small to flow freely from the pipette with the jet of fluid. The movement of the beads across the microscopic field was recorded on 16 mm film for each jet intensity (1 to 10 steps on the microdrive). The position of individual beads was plotted frame by frame to visualize their trajectories as illustrated in Fig. 2 for three beads exiting almost simultaneously from the pipette opening. Flow velocities could thus be determined and were found to decrease with distance from the opening. The jet was essentially laminar and turbulence was negligible. During several calibration experiments the flow profile was simultaneously visualized by adding 1% sucrose to the Ringer’s/plastic bead solution. Using a dark Nomarski field the ejected Ringer’s sucrose solution appeared as a bright glow, revealing the shape of the ejected flow. The beads and the ejected fluid were found to travel in company. It was ascertained in this fashion that micropipettes selected for use in hair deflection experiments provided jets of appropriate width and velocity at the point where a selected sensory hair was to be positioned. Measurement
of hair bundle deflection
The motion pattern of sensory hairs in response to a fluid jet was recorded on 16 mm film running at 25 or 50 frames/s. A light-emitting diode synchronized to the pulse generator produced a light flash on the edge of the film at the initiation of a fluid squirt. The filmed sequence of hair deflections was analyzed by projecting single frames
1
,
.
2
. + . . - v . . 1. . ‘.
:
;
r...
y.*=...,
.
.
.
.
Fig. 2. Three plastic beads (0 v H) were ejected simultaneously from the micropipette opening indicated at the left. Their trajectories were filmed and plotted frame by frame. At arrow 1 the velocity is 240 pm/s. at arrow 2 it is 120 pm/s.
253
on the screen of a viewer (Recordak Magnaprint Reader, Kodak). A transparent graded angle protractor was fixed to the front of the screen so that the protractor appeared superimposed over the image of the hair. For the sensory hair being examined, the protractor was aligned prior to the squirt (the so-called resting position) so that the bundle axis lay at 0”. This axis was easily determined since the Nomarski optics produced a distinct straight-edge shadow along the center of each bundle. Measurement of the angle of deflection then proceeded frame by frame, plotting hair motion at 40 or 20 ms intervals. Scanning electron microscopy The ampulla of the posterior semicircular canal was exposed and fixed in situ in a fixative composed as follows: 2% glutaraldehyde + 4% paraformaldehyde + 2 mM MgCl, in 0.1 M sodium phosphate buffer at pH 6.2. After 30 min the ampulla was separated and immersed in the fixative for another 30 min and was then post-fixed for 1 h in 2% 0~0, in the same buffer and rinsed in distilled water. The crista was dissected free as described above, treated by the OTO-method [ 141 and critical point dried. The specimen was covered with gold by sputter coating (Baltzer) and examined in a Cambridge Stereoscan scanning electron microscope. Results General features The stimulus is defined as the fluid pulse directed at a selected sensory hair from the micropipette, and the response-is the resulting deflection as measured from the film recording. In calibration tests with plastic beads it was found that the velocity of the stimulus jet for a given volume ejected was highly reproducible between repeated pulses. This was confirmed in the preparation with the sensory hairs as indicators of fluid displacement. All hairs within the jet deflect in unison and then return to the resting position. Because the jet expands with distance from the pipette, hairs at the periphery of the jet may exhibit a vertical component of deflection taking their tip out of focus. Such behaviour was avoided by vertical alignment of the center of the pipette to the same focal plane as a selected hair bundle. This procedure ensured that the sensory hair was located in the most laminar part of the
400
Fig. 3. The angle of deflection
800 msec
of a sensory
hair bundle
subjected
twice to identical
stimuli.
254
-
1 step
-
3 step
-
5 step
msec
Fig. 4. Plot of the deflection
of a sensory
hair in response
to three different
velocities
of the jet
flow. For a 50 pm wide pipette the ejected fluid jet was wide enough to easily aim at and fully envelop a hair bundle selected for observation. Fig. 3 shows the time course of deflection of a hair displaced twice by repeated fluid pulses. The accuracy of the measurement of deflection is about 0.5’ and the reproducibility is of approximately the same order. In response to the fluid displacement the hair bundle deflects as a stiff rod hinged at the base. All the stereocilia join the motion as if connected to one another. In Fig. 4 a family of deflection curves for three different pulse velocities are plotted. The angle of deflection reaches a maximum and then returns to the original position. For increasing velocities the angle of deflection increases as seen in Fig. 5. For large displacements there is a
2
3
4
5
step
Fig. 5. The angle of maximum deflection was measured The mean and the standard deviation are plotted.
for three velocities
of the Jet in 7 sensory
hairs.
255
tendency for the hair to return with a higher velocity than at small displacements as seen from the slope of the return phase. Hence, the elastic restoring force increases with deflection of the hair. Time course of sensory hair motion The time course of the return phase varied for different sensory hair bundles as illustrated with the examples given in Fig. 6: (A) Some bundles returned rapidly to the resting position. (B) Other bundles returned more slowly to the resting position. They also tended to reach maximum excursion later. (C) A third type of response was a large initial deflection followed by a slow return. When the time constants for different bundles (displaced by equal velocity pulses) are compared one finds a clear separation into two groups, one below and one above 760 ms. Among the faster time constants there is no clearcut division but a smooth transition from a fast group centered about 80 ms and a population of intermediate time constants. Correlation of sensory hair stiffness to architecture of stereociiium bundle Bundles with slow and fast return can be distinguished in the microscope during the experiment and may be found side by side in the same preparation. Also, slow return can be seen in a bundle placed between two fast ones. Therefore, the observed variability does not depend on differences between preparations or differences in the position of a particular bundle in the jet. Rather, we have observed a correlation between the time constant of a given bundle and its structure in terms of length,
25
400
600
msec Fig. 6. The time pattern of motion of three different hair bundles.
800
1000
256
number and diameter of the stereocilia within the bundle and the relative length of the kinocilium as compared to the stereocilia. Response types A, B and C of Fig. 6 correlate to sensory hairs with the general structural characteristics seen in Fig. 7: (A) The sensory hair bundle is about 5 pm wide at the base and tapers gradually
8A
B
C
Fig. 7. Microphotographs of sensory hair bundles development of stereocilia are seen. Fig. 8. Scanning electron matching Fig. 7.
micrographs
observed
of hair bundles
in the experimental
showing differences
microscope.
in architecture
Differences
in
of the stereocilia
251
towards the tip. The stereocilia within the bundle can be discerned and reach to the tip of the bundle. (B) This type of bundle is somewhat more slender with a width at the base of about 4 pm, and there is a gradual taper to the tip. The stereocilia are more difficult to discern and seem to end below the tip of the hair. (C) The sensory hair bundles have a small basal bundle of stereocilia and a tall solitary kinocilium. On the basis of these results scanning electron microscopy was performed on the crista ampullaris in order to search for and examine in closer detail the corresponding types of sensory hairs. Scanning electron micrographs of such sensory hairs are shown in Fig. 8 and are described below: (A) The stereocilia are massive with a width of about 0.4 pm and insert in a broad cuticular plate. The tallest stereocilia accompany the kinocilium up to just beneath the tip. (B) The total width of the stereociliary bundle is smaller and the diameter of the individual stereocilia is less, measuring about 0.3 pm. The tallest stereocilia end about halfway to the tip of the kinocilium. (C) This type of bundle has a solitary kinocilium joined at its base by a group of short stereocilia having a diameter of about 0.15 pm. The cuticular plate has a correspondingly small diameter. Regional distribution of different types of bundles The distribution of hair bundles type A, B and C is shown in Fig. 9. Fast response type A bundles are present in the central part of the crista only. Medium fast type B bundles dominate the lateral planum region and are also scattered in the center. Slow type C bundles are present mainly along the edges of the sensory epithelium with a few found higher up on the ridge centrally in the crista.
Fig. 9. Distribution of sensory hairs type A, B and C.
258
Discussion The stiffness of vestibular sensory hairs was first noted by Schultze [21] and Retzius [20] but was forgotten, probably because microdissection of fresh tissue was abandoned when fixation, embedding and sectioning techniques were introduced around the turn of the century. It is now known that fixation of the tissue results in loss of sensory hair stiffness unless special fixation techniques are used [5]. The stereocilium stiffness relies on an inside core of actin filaments [3] probably cross-linked by the protein fimbrin [6]. A similar structure of the core has been reported in the lizard basilar papilla where a paracrystalline packing of the actin filaments was noted [23]. Such a core implies nonelastic properties, and the elastic component responsible for the return of the sensory hair has to be looked for elsewhere. Restoring force could be supplied by the network of fibrillar material that intertwines the ciliary bundles. A more important site may be at the tapered neck of the stereocilia where the actin filaments are tightly joined and penetrate as rootlets into the cuticular plate, a tight mesh of actin filaments beneath the apical membrane. The present work shows that the micromechanical behaviour of a sensory hair is closely related to the architecture of the stereocilia of which it is composed. Ranging from fast to slow the sensory hairs would in physiological terms be described as dynamic rapidly adapting velocity sensitive units, or static slowly adapting displacement sensitive units. This differentiation of mechanical response patterns may derive ‘passively’ from the structural differentiation of the sensory hairs or it could rely on act:ve mechanisms related to those which cause ciliary beat or muscle contraction. Invoked in the latter case is a dependence on membrane related ionic mechanisms. The mechanical properties or the sensory hairs described here do not, however, appear to depend critically on an intact membrane [4] and are therefore more likely to be built into hair structure than being derived from active mechanisms. The different sensory hair response types would provide individual filtering characteristics for individual hair cells, if allowed to become expressed in the physiological situation. This will depend on the physical nature of the coupling between sensory hairs and the material of the cupula, a question which is under discussion [2,9,14]. In the vestibular system differences in firing patterns, particularly in adaptive behaviour, among nerve fibres innervating the crista ampullaris have been described by several authors [8,17,19]. There are two types of hair cells in the mammalian crista ampullaris which differ significantly in structure and innervation patterns [26], a fact that has been suggested as a basis for differences in nerve response properties. Lim [14] has noted that Type I and Type II hair cells in the mammalian vestibular system also have different bundle structures. Hair cells in the frog crista ampullaris are all Type II according to innervation patterns. but they can be subdivided into at least two groups according to sensory hair structure ([9], and the present study). We find that the different bundle types are differentially distributed over the crista, fast thick bundles in the centre, slow slender bundles on the two sides. Highly significant in this context is the work of Honrubia et al. [ 1 l] correlating anatomical and physiological characteristics of primary afferents in the frog crista ampullaris. They found that neurons with an evoked impulse discharge of
259
short time constant innervate the central part of the crista whereas neurons with slow time constant characteristics as a whole tended to innervate peripheral regions. The positive correlation between hair bundle mechanical response and nerve discharge properties does not prove a causative relationship but shows that receptors and neurons are tuned in harmony. The concept of sensory hairs acting as mechanical filters does have implications for other inner ear sense organs. For example, in the basilar papilla of the lizard only hair cells in a particular region are covered by a tectorial membrane, the remaining being free-standing in the endolymph [27]. The free-standing sensory hairs vary in length along the papilla and there is a corresponding variation in mechanical tuning of the sensory hairs [7,10] and best frequency of neural responses [24,25]. A causal relationship has therefore been suggested. Also in the mammalian organ of Corti there is a correlation between the length of outer hair cell stereocilia [ 151 and sensory hair stiffness [22], indicating a general contribution of sensory hair mechanical properties to the tuning of the integrated organ. Also, there are indications that inner and outer hair cells differ in their coupling to the tectorial membrane [ 11, inner hair cells being more loosely coupled. The physiological response of the inner hair cell should then be influenced by the mechanics of its stereocilia. The present results show that the physiological response characteristics of a hair cell can be dictated by the architecture of its sensory hair bundle. These filter characteristics are passive, acquired during evolution. There is indication that the mechanical characteristics of the hair cell can also be actively controlled by processes akin to those that operate in muscle contraction. This is dealt with in an accompanying publication [4].
Acknowledgements This work was supported by grants from the Swedish Medical Research (04X-02461), the Soderberg’s Foundation and the Tysta Skolan Foundation.
Council
References 1 Dallos, P. (1981): Cochlear physiology. Annu. Rev. Physiol. 32, 153- 190. 2 Dohlman, G.F. (1980): Critical review of the concept of cupula function. Acta Otolaryngol. Suppl. 376, l-30. 3 Flock, A. and Cheung, H. (1977): Actin filaments in sensory hairs of the inner ear receptor cells. J. Cell Biol. 75, 339-343. 4 Flock, A. and Orman, S. (1983): Active control of sensory hair mechanics implied by susceptibility to media that induce contraction in a muscle fibre. Hearing Res. 11, 261-266. 5 Flock, A., Flock, B. and Murray, E. (1977): Studies on the sensory hairs of receptor cells in the inner ear. Acta Otolaryngol. 83, 85-91. 6 Flock, A., Bretscher, A. and Weber, K. (1982): Immunohistochemical localization of several cytoskeletal proteins in inner ear sensory and supporting cells. Hearing Res. 6, 75-89. 7 Frischkopf, L.S., De Rosier, D.J. and Engelman, E.H. (1982): Motion of basilar papilla and hair cell stereocilia in the excised cochlea of the alligator lizard: relation to frequency analysis. Sot. Neurosci. Abstr. 8, 40.
260
8 Goldberg, J.M. and Fernandez, C. (197 1): Physiology of peripheral neurons innervating semicircular canals of the squirrel monkey. I. Resting discharge and response to constant angular accelerations. J. Neurophysiol. 34, 635-684. 9 Hillman, D.E. (1974): Cupular structure and its receptor relationship. Brain Behav. Evol. 10, 52-68. IO Holton, T. and Hudspeth, A.J. (1982): Motion of hair cell stereocilia in the auditory receptor organ of the alligator lizard. Sot. Neurosci. Abstr. 8, 40. 11 Honrubia, V., Sitko, S., I&mm, J., Betts, W. and Schwartz, I. (1981): Physiological and anatomical characteristics of primary vestibular afferent neurons in the bullfrog. Int. J. Neurosci. 15, 197-206. 12 Hudspeth, A.J. and Jacobs, R. (1979): Stereocilia mediate transduction in vertebrate hair cells. Proc. Natl. Acad. Sci. U.S.A. 76, 1506- 1509. 13 Lewis, E.R. and Li, C.W. (1975): Hair cell types and distributions in the otolithic and auditory organs of the bullfrog. Brain Res. 83, 35-50. 14 Lim, D.J. (1979): Fine morphology of the otoconial membrane and its relationship to the sensory epithelium. SEM 1979 III, IIT, pp. 929-938. SEM Inc., AMF O’Hare, IL 60666. 15 Lim, D.J. (1980): Cochlear anatomy related to cochlear micromechanics. A review. J. Acoust. Sot. Am. 67, 1686-1695. 16 Lowenstein, O., Osborne, M.P. and Wersall, J. (1964): Structure and innervation of the sensory epithelia of the labyrinth in the Thornback ray (Raja clauara). Proc. Roy. Sot. Lond. Ser. B, 160, l-12. 17 O’Leary, D.P., Dunn, R. and Honrubia. V. (1974): Functional and anatomical correlation of afferent responses from the isolated semicircular canal. Nature (London) 25 1, 225-227. 18 Orman, S. and Flock, A. (1981a): Micromechanics of the hair cell stereociliary bundles in the frog crista ampullaris. Ass. Res. Otolaryngol. Abstr. 4. 30-31. 19 Precht, W., Llinas, R. and Clarke, M. (1971): Physiological responses of frog vestibular fibers to horizontal angular rotation. Exp. Brain Res. 13, 378-407. 20 Retzius, G. (1884): Das Gehororgan der Wirbeltiere. I. Fische und Amphibien. Sanson and Wallin. Stockholm. 21 Schultze, M. (1858): Ueber dei Endigungsweise der Hornerven im Labyrinth. Arch. Anat. Physiol. Wiss. Med. 343-381. 22 Strelioff, D. and Flock, A. (1982): Mechanical properties of hair bundles of receptor ceils in the guinea pig cochlea. Sot. Neurosci. Abstr. 8, 40. 23 Tilney, L.G., De Rosier, D.J. and Mulroy. M.J. (1980): The organization of actin filaments in the stereocilia of cochlear hair cells. J. Cell Biol. 86, 244259. 24 Turner, R.G., Muraski, A. and Nielsen, D.W. (1981): Cilium length: Influence on neural tonotopic organization. Science 213, 1519- 1521. 25 Weiss, F.E., Peake, W.T., Ling, A. and Holton, T. (1978): Which structures determine frequency selectivity and tonotopic organization of vertebrate cochlear nerve fibres? Evidence from alligator lizard. In: Evoked Electrical Activity in the Auditory Nervous System, pp. 91-l 12. Editors: R. Naunton and C. Fernandez. Academic Press, New York. 26 Werslll, J. (1956): Studies on the structure and innervation of the sensory epithelium of the cristae ampullares in the guinea pig. Acta Otolaryngol. Suppl. 126. l-85. 27 Wever, E.G. (1967): The tectorial membrane of the lizard ear. Types of structure. J. Morphol. 122, 307-320.