Ionic coupling among cells in the organ of Corti

Hearing Research, 57 (1992) 175-194 0 1992 Elsevier Science Publishers B.V. All rights reserved

HEARES

175 0378-5955/92/$05.00

01673

Ionic coupling among cells in the organ of Corti Jozef J. Zwislocki, Norma B. Slepecky, Lisa K. Cefaratti and Robert L. Smith Institute for Sensory Research, Syracuse University, Syracuse, New York, U.S.A. (Received

26 March

1991; accepted

23 August

1991)

Gap junctions have been demonstrated morphologically among the supporting cells of the mammalian organ of Corti but, in contradistinction to reptiles, evidence for their existence between the supporting cells and hair cells is equivocal. The literature is ambiguous with respect to electrical coupling and dye coupling among the supporting cells, and no coupling of either kind has been demonstrated for the hair cells. We found strong coupling of both kinds among the supporting cells in the cochleas of live Mongolian gerbils and a less stable coupling between the supporting cells and the outer hair cells. The electrical coupling was established by recording alternating receptor potentials in the hair cells and following their decrement in the population of Hensen’s,cells; the dye coupling, by injecting Lucifer yellow electrophoretically into the hair cells or the supporting cells and investigating its spread to the neighboring cells. The electrical recordings were made by means of microelectrodes filled with either 1.5 or 3 M KCI or 1 M LiCl with 6% Lucifer yellow, the latter used for dye injection. The electrode resistances ranged from about 20 to 60 MR in the first instance, and from about 50 to 110 MR, in the second. The electrodes were inserted into the organ of Corti through Scala media according to the method of Dallos, Santos-Sacchi and Flock (1982) modified by us. The alternating potential in Hensen’s cells was usually larger than in the outer tunnel of Corti and remained practically constant up to the outer margin of the Hensen’s-cell population. Its phase was the same as in the outer hair cells. When the dye was injected into a Hensen’s cell, it always spread to neighboring Hensen’s cells and often to Deiter’s cells. Dye injected into outer hair cells (identified according to anatomical and physiological criteria) also spread to Deiter’s and Hensen’s cells and, usually, to other outer hair cells. Stained cells were identified in surface preparations and, on two occasions, in serial sections from plastic embedded cochleas. Cochlea;

Hair cells; Supporting

cells; Electrical

potentials;

Gap junctions

Introduction

Ionic coupling among cells in the organ of Corti must affect the biological homeostasis of the organ as well as the cochlear electromechanics, especially, in view of the bidirectional coupling between the mechanical vibration and the electrical fields it generates. In addition, knowledge of the types of cells coupled is relevant to the interpretation of cochlear electrical recordings (e.g. Dallos, 1983, 1984). The literature describing the morphological evidence for gap junctions in the mammalian organ of Corti, thought to be essential for electrical coupling between cells, is unanimous in documenting such junctions among the supporting cells (Gulley and Reese, 1976; Iurato et al., 1976; Nadol, 1978; see for review Oesterle and Dallos, 1989) but not between the supporting cells and hair cells. Iurato and his coworkers found no morphological evidence for such junctions. Gulley and Reese saw particles at appositions between hair cells and supporting cells, which resembled those found in association with gap junctions in the retina

Correspondence to: Jozef J. Zwislocki, Institute for Sensory Syracuse University, Syracuse, NY 13244.5290, U.S.A.

Research,

(Raviola and Gilula, 1973). However, they found no other structural evidence for gap junctions in their freeze-fracture preparations. Nadol found structures suggesting gap junctions at juxtapositions between hair cells and supporting cells in thin sections of human and cat preparations of the organ of Corti examined by transmission electron microscopy. However, the intermembrane space was filled with dense material not usually found in such junctions. In lizards, gap junctions between the hair cells and the supporting cells were free of this material (Nadol et al., 1976; Miller and Beck 1990). Below this phylogenetic level, gap junctions have been found by Hama (1980) on hair cells in the saccular macula of goldfish. The literature concerning functional evidence for gap junctions among the supporting cells in the mammalian organ of Corti is equivocal. On the basis of intracellular and intercellular recordings of alternating potentials and electrical studies, Oesterle and Dallos (1989) concluded that the supporting cells communicate with each other electrically through direct contacts but that this is unlikely for their interaction with the hair cells. Their conclusions are based to a large extent on the finding that, in guinea pigs, intracellular alternating potentials of all the supporting cells in the organ of Corti are similar and smaller than or equal to the

176

corresponding potentials in the intercellular space, especially, the outer tunnel of Corti. They mention, however, that the relatively small alternating potentials of the supporting cells do not by themselves prove the absence of direct electrical coupling between these cells and the hair cells. One method for demonstrating gap junctions relies on dye transfer from an injected cell to neighboring cells. Using this method, Santos-Sacchi and Dallas (1983) failed initially to find gap junctions in either live or in vitro preparations of guinea-pig organs of Corti. Subsequently, Santos-Sacchi (1986) was able to find dye coupling among Hensen’s (HCs) and Deiter’s cells (DCs) in preparations maintained in tissue-culture medium by means of fluorescein or 6-carboxy-fluorescein. Lucifer yellow produced equivocal results. The author speculated that Lucifer yellow may have precipitated in the potassium-rich cytoplasm. To the best of our knowledge, no dye transfer has yet been found among the cells of the mammalian organ of Corti in vivo. Our studies of the ionic coupling in Mongolian gerbils were prompted by our observation that the intracellular alternating potentials of HCs of these animals were almost always larger than in the intercellular space, especially, in the outer tunnel of Corti. The reverse was encountered only occasionally, usually in run-down preparations in which both the alternating cochlear potentials and the endolymphatic potential (EP) were substantially reduced relative to what appeared to be healthy preparations. If the alternating potential of HCs is a reflection of the receptor potential of the outer hair cells (OHCs; see for review Oesterle and Dallas, 1989), it is difficult to see how it could exceed the corresponding potential in the intercellular space separating the supporting cells from the OHCs unless an ionic pathway shunted the intercellular space. Since HCs are separated from the OHCs by DCs, the pathway would have to lead through them. According to Oesterle and Dallos, supporting-cell alternating potentials that exceed those in the intercellular space are an important criterion for the existence of gap junctions. in contrast to the experiments of Oesterle and Dal10s (1989) performed on regions of the guinea-pig cochlea having best frequencies (BFs) below 1 kHz, our experiments were performed on the second turn of the Mongolian-gerbil cochlea, at a location corresponding to a BF around 2 kHz. This is the most easily accessible location in our gerbil preparation.

Methods Surgery and physiological

procedure

Initial surgery was as described by Zwislocki et al. (1988). Mongolian gerbils weighing between 60 and

75 g were anesthetized with sodium pentobarbital (35 mg/kg) and tracheotomized, but no artificial respiration was applied. Soft tissue over the ventrolateral bulla and at the ear canal was removed and the bulla opened ventrolaterally so as to expose the whole cochlea but leave the middle-ear apparatus intact. A silver-wire electrode was placed in the round-window antrum and fixed to the bone with dental cement. It served for monitoring cochlear microphonics (CM) as an index of the overall state of the cochlea and of sound transmission through the middle ear. A miniature earphone (Knowles, ED-1932) with an acoustically corrected transfer characteristic and a miniature microphone (Knowles, EA-1934) were coupled to the bony ear canal by means of a conical teflon adapter and a sealing plastic compound. The sound pressure generated by the system in the ear canal was constant within f3 dB between 0.25 and 10 kHz. The acoustic signals delivered to the ear canal consisted of sinusoids swept logarithmically along the frequency scale from about 0.18 to 18 kHz. Every sweep lasted 5 s and was followed by a pause of about 1 s, when automatically triggered. Manual triggering was also possible. Part of the bone of the first and second turns of the cochlea was stained with Alcian blue to improve the visibility of microelectrodes during their insertion into the cochlea (the dye did not penetrate the bone). One to three holes were drilled in the cochlear bony capsule and into the spiral ligament of the second cochlear turn over Scala media, about 200 pm apically from the interturn groove, as shown in Fig. 1. Precise location of the holes was made possible by the use of a motorized low-speed drill developed especially for this purpose and mounted on the same micromanipulator as used for the microelectrodes. The micromanipulator was under three-axial hydraulic control (Narishige, MO103). As a result of a special guide tube, the wobble at the tip of the drill bit was under 5 pm and could not be discerned under the 40X magnification. The drilling produced a cochlear round-window (RW) output equivalent to a mid-frequency sinusoid of only about 60 dB. An electrical heating coil prevented vapor condensation on the guide tube or the bit. The drill is shown in its lateral aspect in Fig. 2. Drilling into the tough spiral ligament greatly facilitated insertion of fine microelectrodes into the cochlea. Because of the resilience of the spiral ligament, the hole in it did not remain entirely open but sealed itself partially so that, during microelectrode insertion, tissue was encountered. The endolymphatic potential was not encountered until stria vascularis was traversed, and its magnitude was consistent with preceding findings (e.g. Schmiedt and Zwislocki. 1977). These observations suggest that drilling into the ligament did not adversely affect intracochlear conditions.

vidual fine-adjustments of the angle were made so as to find the organ of Corti and its outer tunnel. With some practice, finding these structures proved to be reasonably easy without visualization of the organ of Corti. As monitored by means of cochlear direct and alternating potentials, the electrodes encountered the organ of Corti at a distance of about 100 pm from the slightly pressed-in stria vascularis and the outer tunnel about 50 to 80 ,um further in. Cells were penetrated with the help of negative-capacitance ringing. I-Es and other supporting cells had resting potentials between -60 and - 100 mV, and OHC, between - 60 and - 86 mV, in agreement with the findings of Dallos and his associates (e.g. Oesterle and Dallos, 1989). The OHCs had the highest amplitudes of alternating potentials and were surrounded with an intercellular zone of approxi-

Fig. 1. Cochlear capsule as seen during the experiments. The three arrowheads point to three holes drilled in the second turn over Scala media. Through the rightmost hole in the bone it is possible to discern a small hole in the spiral ligament. Part of the capsule is stained with Alcian blue to improve the visibility of microelectrodes during their insertion into the cochlea.

Microelectrodes with a short taper were used. They were pulled on the horizontal Brown-Flaming puller from l-mm Borosii Omega-Dot tubing. They were filled with 1.5- to 3-M KC1 for recording of cochlear potentials and with a millipore-fiItered solution of 6% Lucifer yellow in 1-M LiCl for dye marking in addition to the recording of the potentials. The KC1 electrodes had resistances between 20 and 60 Ma, the LiCl ones, between 50 and 110 MO. The resistance was monitored with the help of positive current injection and the active-bridge circuitry provided on the WPI KS-700 amplifier. Comparisons between electrode resistances measured in the intercellular fluid and inside cells provided us with estimates of cell-membrane resistance (Oesterle and Dallos, 1989). The electrodes were held in a WPI holder with a silver-chloride interface, and the same interface was used in the ground electrode embedded in a cheek muscle to cancel out electrodeinterface potentials. The electrodes were placed and advanced with the help of the three-axial, hydraulic micromanipulator mentioned above. They entered the cochlea nearly perpendicularly to the cochlear bone surface, as indicated schematically in Fig. 3, but indi-

Fig. 2. Side view of the precision drill used to make the holes in the cochlear capsule. The drill is seen mounted on a three-axial hydraulic micromanipulator. Note the narrow stationary tube marked by the arrowhead at the bottom of the drill - it stabilizes the drill bit by means of a teflon bushing near its end. The tip of the bit can be seen protruding from the tube. A small electric coil near the other end of the tube heats the tube and the bit to prevent water condensation.

17x

mately 0 potential. The physiological criteria for deciding that an OHC was impaled were: 1) distance from the outermost HC in excess of 50 pm; 2) an approximately 0 potential surrounding the cell; 3) alternatingpotential amplitude distinctly larger than in the intercellular space and in HCs. To find an OHC, we proceeded as follows. As soon as the EP potential and CM at the tip of the electrode started to decrease, signaling contact with the organ of Corti, negative-capacitance ringing was applied. More often than not, the potential decreased to strongly negative values: indicating penetration of a HC. The response phase changed by about 180 O. The electrode was further advanced, making sure that the alternating and direct potentials at the electrode tip preserved approximately their initial val-

ues recorded after penetration of the first HC. This was usually the case until the outer tunnel was encountered. Occasionally, the values of these potentials decreased because of partial clogging of the microelectrode. They were usually restored after a short burst of negative-capacitance ringing. When zero direct potential was encountered at a distance in excess of 50 pm, usually accompanied by a reduced alternating potential, and the intracellular HC conditions were not restored after negative-capacitance ringing, we assumed that the electrode tip was inside the outer tunnel, and started to search for OHCs. The electrode was advanced slowly and the direct and alternating potentials carefully monitored. A change of the first toward negativity and an increased amplitude of the second were

Fig. 3. Midmodiolar section of the second turn of the gerbil cochlea at the location of our experiments. The hole in the bony capsule and SFGral ligarnent and the location of the microelectrode are shown schematically. Approximately horizontal orientation of the cochlea. as during the experiments.

179

regarded as possible signals of contact with an OHC. A burst of negative-capacitance ringing often led to the penetration of the cell. A relatively high amplitude of the alternating potential was taken as confirmation that the penetrated cell was an OHC. Usually, OHCs could only be held for periods of time shorter than 1 minute, in contradistinction to HCs which could be held almost indefinitely. In all preparations, magnitude transfer functions, and in some, phase transfer functions of the alternating potentials were recorded with the help of the logarithmic frequency sweeps. The potentials were filtered in a Lock-in Amplifier with a ~-HZ bandwidth, displayed on the screen of a storage oscilloscope and photographed whenever a permanent record appeared desirable. In the manual trigger mode, it was possible to produce multiple oscilloscope traces selectively for direct quantitative comparisons among traces obtained under diverse conditions. The magnitude was displayed linearly and the frequency logarithmically with the help of a home-built frequency-to-voltage converter that controlled the horizontal deflection on the oscilloscope screen. The phase along the log-frequency scale was displayed as a sine function defined by the equation y = Y sin [wAt(f) + a(f)], where Y is the amplitude, w = 2af, f, sound frequency, At(f), the wave time delay in the cochlea, and a(f), the phase shift relative to the phase of the voltage input to the earphone. It should be noted that both At(f) and positive a(f) signify phase lags which can be compensated for by decreasing the frequency, f. In addition, an increased At(f) increases the slope of the phase function 0 = [wAt(f) + a(f)], whereas an increased a(f) produces a parallel shift. The obtained functions, y = F(f), indicated that both At(f) and cu(f) were either approximately constant or varied slowly with f. In dye marking experiments, the dye was injected electrophoretically into impaled cells by means of a square wave with a 5-nA amplitude and a 16-Hz fundamental frequency. The injection was always stopped before the magnitude of the resting potential decreased to one half. In some experiments sound was delivered to the cochlea at 40 dB SPL during dye injection. Usually, only one cell was injected per hole. On some occasions, however, two electrode tracks were made at two different angles in the plane of the basilar membrane so as to obtain nonoverlapping staining regions. It was observed that dye injection did not usually affect the resting and alternating potentials of the cells injected, in agreement with the experience of Oesterle and Dallos (1989) concerning horseradishperoxidase cell marking. Histological procedures

At the end of the physiological part of an experiment, the stapes was extracted from the oval window

and the cochlea fixed with 4% paraformaldehyde in 0.1-M sodium phosphate buffer, perfused through a small hole in Scala tympani of the basal turn at a rate of 9.2 pl/min for 15 min. The cochlea was then extracted and stored refrigerated in the same solution for one to several days prior to dissection. It then was washed in 0.1 M phosphate buffer, and the thin bone over the front face was removed. The lateral wall containing the stria vascularis and spiral ligament was taken off this region. The turn containing the injected cell(s) was removed from the modiolus and placed in buffer on a Nikon Diaphot inverted microscope. The specimen was observed with phase-contrast optics or fluorescence microscopy with a filter set for fluoresceine. In most cases, the specimens were photographed with a 35 mm camera through a 2X adapter and a 10X glycerine objective. In some cases, the specimens were further dissected and photographed through a 40X glycerine objective. Black and white negatives of triX400 film were developed in HC-110 (dilution B for 6.5 min). For serial-section analysis, the piece of the cochlea used for surface preparation was washed in 0.1 M phosphate buffer, dehydrated through 80% ethanol, embedded in Historesin, and cured in a vacuum for several hours at room temperature. Sections were cut with a glass knife on a JB-4 microtome, at 10 pm thickness. They were placed on slides on top of drops of immersion oil to flatten, immersed in the oil, and coverslipped. Slides were observed with phase or fluorescence microscopy using a Zeiss Standard microscope and a 40X oil objective. Photographs were made as for the surface preparations.

Results Alternating Potentials All the potentials were displayed on the screen of a

storage oscilloscope in the form of transfer functions (constant input), the magnitude vertically on a linear scale, and the frequency horizontally on a log scale. The BFs ranged from 1.5 to 2 kHz, and the main lobes of the functions extended to about one octave below and one octave above BF. Amplitudes of the alternating potentials of an OHC recorded intracellularly and extracellularly are compared in the upper panel of Fig. 4. The lower panel shows the intracellular potentials of a HC recorded on the same electrode track. Note that the intracellular OHC potential is substantially larger than the extracellular one, and that the HC potential is about equal to the latter. The intracellular potentials of a HC (highest trace) are compared to the outer-tunnel potentials (lowest trace) and to CM, all obtained on the same electrode track, in Fig. 5. The noisiness of the traces is due to the low SPL of about 30 dB at which

180

Fig. 5. Cochlear alternating-potential transfer functions plotted as in Fig. 4. HC - Hensen’s cell, intracellular; CM - scala media; OT - outer tunnel. Note that the magnitude is the highest in HC and the lowest in the outer tunnel: also, that the CM magnitude is lower than the HC magnitude and that the CM has a somewhat lower BF than HC. These relationships are typical. SPL = 30 dB; BF, about 1.8 kHz; EP = 60 mV; RP = - 87 mV. GZ17Y.

Fig. 4. Cochlear alternating-potential transfer functions plotted automatically over a log- frequency scale. The vertical magnitude coordinates are linear. In the upper panel, top curve - OHC, intracellular; lower curve - extracellular. In the lower panel, HC, intracellular. Note the similarity of patterns. The HC magnitudes are about equal to the extracellular OHC magnitudes. SPL = 40 dB; BF = 1.8 kHz; EP = 82 mV; RP = - 85 mV (OHC), - 100 mV (HC). Animal: GZ-80.

they were generated. .Note that both the CM and the outer-tunnel potentials have lower magnitudes than the HC potentials and that the CM has a somewhat lower BF than HCs. These relationships as well as the shapes of the transfer functions are typical of our gerbil population. The expected decrement of HC alternating potentials from the outer tunnel to the outer margin of the organ of Corti is illustrated in Fig. 6. The separate transfer-function traces represent samples taken along an electrode track from the outer margin to the outer tunnel of the organ of Corti. The highest trace was obtained at the margin, the second lowest, at the tunnel - an inversion of what was expected. The probable reason for the inversion was a progressive obstruction of the electrode. We observed that a burst of negative-capacitance ringing restored approximately the value obtained in the first cell that was impaled. With this corrective maneuver, we were not able to ascertain any HC-potential difference between the outer tunnel and the margin. The lowest trace in Fig. 6 shows the relatively low potential found in the outer

tunnel. It demonstrates that the uniformity of the HC potentials was not provided by the intercellular fluid. Typical phase relationships are shown in Figs. 7, 8 and 9. The first shows by means of the sine-phase curves (see methods) that the OHC and outer-tunnel

Fig. 6. Cochlear alternating-potential transfer functions plotted as in Fig. 4. The upper set of curves was recorded as the electrode was advanced intracellularly through the population of HCs toward the outer tunnel, the lowest curve was recorded in the outer tunnel. The small differences among the HC curves do not represent a systematic potential gradient but are due to resistance variation of the electrode. The lower-frequency peak is an artifact due to an unusual middle-ear transfer function with a strong peak at the same frequency. SPL = 40 dB: BF, about 1.8 kHz; EP = 54 mV; RP = -90 mV. GZ-87.

Fig. 7. Phase transfer functions recorded as sine phase functions (see: methods) in an OHC (upper panel) and in the outer tunnel (lower panel). The thin curves were recorded at 40 dB SPL, the thick ones, at 100 dB. Note the phase reversal at 100 dB and the near phase identity in the OHC and the tunnel. BF, about 1.8 kHz; EP = 82 mV; RP = - 73 mV.GZ-139.

Fig. 9. Phase transfer functions plotted as in Fig. 7. The thin curve was recorded in a HC, the thick one, in the endolymphatic space. Note the phase reversal between the two curves and that the HC amplitude is slightly larger than the CM amplitude. SPL = 40 dB; BF, about 1.8 kHz; EP = 67 mV, RP = - 93 mV. GZ-179.

potentials are essentially in phase; the second, that the same is true for the HCs and the outer tunnel; the third, that CM is in phase opposition with the HCs. These results are consistent with those of Oesterle and Dallos (1989) and with the well known phase reversal on penetration of the organ of Corti from the endolymphatic space. The finding that the alternating potentials of the HCs are almost always larger than the alternating potentials in the outer-tunnel is sufficient by itself for us to conclude that ionic pathways exist between the OHCs and HCs, which shunt the outer-tunnel space. Since, morphologically, the HCs can communicate with the OHC only through DCs, the latter must be in-

Fig. 8. Phase transfer functions plotted as in Fig. 7. The thin curve was recorded in a HC, the thick one in the outer tunnel. Note the near phase identity between the two curves and that the HC amplitude is substantially larger than the outer-tunnel amplitude. SPL = 40 dB; BF, about 1.8 kHz; EP = 67 mV; RP = - 88 mV. GZ-179.

Fig. 10. Fluorescence photomicrograph of Lucifer-yellow distribution among Hensen’s cells after injection into one cell. Note the elongated pattern of dye spread. The two arrowheads on the right indicate two stained nuclei that are in focus. The arrowhead on the left points to the location of the greatest dye concentration, probably the injection site. Above the stained region, three rows of OHC nuclei are faintly visible through autofluorescence. Sound was left on during dye injection at 40 dB SPL. EP = 50 mV; RP = - 70 mV; electrode resistance, ER = 70 MR; injection duration, T = 2.75 min. GZ-153.

182

eluded in the pathway. The finding that the whole population of HCs between the outer margin of the organ of Corti and its outer tunnel is nearly at the same alternating potential indicates that the ionic channels among the cells have low resistances by comparison to the cell-membrane resistances. By injecting electrical current into HCs, we found resistances in the order of 20 M0, in rough agreement with the measurements of Oesterle and Dallos (1989). Because of the apparently close electrical coupling among HCs, this value does not refer to a single cell but rather to a large aggregate of HCs. To be sure, the uniformity of the HC alternating potentials cannot be produced by CM, which is in phase opposition to the alternating potentials of the

Fig. 12. Typical, roughly circular pattern of dye spread among HCs, which tends to occur when one HC is injected in the absence of ;ound. Superposition photomicrograph. EP = 60 mV; RP = - 71 mV; ER = 64 MR; T = 1 min. GZ-157.

xgan of Corti, and which we have found to have a different BF than these potentials and to usually be smaller than the HC potentials. Also the CM in Scala :ympani is known to be smaller than in Scala media IWeiss, Peake and Sohmer, 19771, further indicating :hat it cannot have such an effect. Fig. 11. Fluorescence photomicrograph of stained HCs in another preparation. Lucifer yellow was injected into one HC in the presence of 40 dB SPL sound. Clearly. the dye is present in several cells, with the predominant spread in the radial cochlear direction. OHCs are faintly visible above the stained HCs through autofluorescence. EP = 5.5mV: RP = - 73 mV: ER = 55 MR; T = 3 min. GZ-156.

Dye coupling

Cells of the organ of Corti were injected with Lucifer yellow in 20 gerbils. The dye was injected systematically into HCs and OHCs but, incidentally, also into

183

DCs and pillar cells. In all 20 animals, dye-filled cells were identified in surface preparations in HCs. In 9, it was also found in DCs, and in 6, in OHCs. In a small number of preparations, it also invaded the pillar cells and the supporting cells at the IHC. In most preparations, it was possible to identify the stained cells in surface preparation unambiguously. The overlap between OHCs and DCs presented the greatest difficulty. Since the presence of the dye in the OHCs was crucial for the confirmation of our electrical-potential results indicating ionic communication between them and the supporting ceils, a substantial effort was devoted to separating the stained OHCs from the stained DCs. In the first set of experiments, Lucifer-yellow was injected into HCs. Since these cells are easy to find, penetrate and hold, no particular difficulties were encountered. Each dye injection lasted for over a minute, and if the monitored resting potential became gradu-

Fig:. 13. Massive dye spread to many HCs and on the right. A HC at the margin of the organ milddle arrowhead - DCs; lowest arrowhead EP =

ally less negative, the process was terminated before the potential reached half its original value. Three examples of the staining results are shown in Figs. 10, 11, and 12. The photomicrograph of Fig. 10 was obtained with a fluorescence filter and shows the stained HCs. The three rows of the OHCs are faintly visible because of autofluorescence and serve as a landmark for the location of the stained HCs. Clearly, the dye injected into one cell diffused to its neighbors, and three stained cell nuclei are clearly discernible. A similar pattern can be seen in Fig. 11. Here, the predominent dye spread is clearly in the radial direction, as indicated by the position of OHCs that can just be discerned at the top of the photograph. In Fig. 12, the same region of the organ of Corti was photographed twice with the same focus, once with the fluorescence filter and once with phase-contrast optics, and the two negatives were superimposed. Again it is

DCs, as seen in a fluorescence photomicrograph on the left and a superposition photomicrogl *aph of Corti was injected (brightest spot toward the bottom of the photomicrographs). P - pillar cells; - HC with greatest dye concentration; top arrowhead - nerve fiber crossing the tunnel of Corti. 22 mV, RP = - SS mV; ER = 30 MO; T = 2 min. GZ-162.

184

clear that the dye invaded several HCs, and the same was true for all the other preparations where Lucifer yellow was injected into a HC. No dye was found in the OHC under these conditions. We should mention that the dye injection that produced the staining patterns of Figs. 10 and 11 occurred during sound delivery to the ear at a SPL of 40 dB, whereas the pattern in Fig. 12 was obtained with an injection made after the sound had been turned off. Note that the latter pattern is much more symmetrical than the preceding two. Although variabili~ was substantial, we gained the impression that the patterns obtained in the presence of acoustic stimuli tended to be more elongated in the radial cochlear direction than the patterns obtained in their absence. The patterns of Figs. 10 to 12 were obtained in the presence of rather low values of the endolymphatic potential (EP), ranging from 50 to 60 mV. However, no systematic effect of the potential was observed throughout the course of our experiments. In one preparation, dye injected into a HC near the margin of the organ of Corti invaded a Iarge number of

Fig. 15. An injected DC shown in a superposition photomicrograph (40x objective). Note that the stained cell partially envelopes the base of an OHC (arrowhead) and that the dye is visible in many other cells (not in focus). EP = 81 mV, RP = - 79 mV; ER = 70 MO; T = 2 min. GZ-163.

Fig. 14. Same specimen as in Fig. 13 but with the focus near the reticular lamina to show the dye in the phalangeal processes of DCs. P - pillars-of-Corti region: 01, 02, 03 - unstained OHCs of the first, second and third rows; D - phalangeal process; H - HC region;. The arrowhead points to the stained cytoplasmic region of a phalangeal process.

DCs and stained their nuclei and cell bodies, including their phalangeal processes, as shown in Figs. 13 and 14. The left panel of the first figure shows the stained nuclei of DCs, the right paneI, a superposition of the same fluorescence negative with its phase-contrast counterpart. The area of the injected HC at the bottom and the massive spread of the dye to the area of DCs and OHCs at the top are clearly visible. The second of the two figures shows the same specimen at a different plane of focus, with the stained phalangeal processes of the DCs at the reticular lamina. Note that the OHCs were not stained, as can be concluded from the dark areas designated as Or, O,, and 0, at the location of these cells. Although no effect of the EP on the dye spread among HCs was noticed, it should be mentioned that the massive dye spread to DCs seen in this figure took place in the presence of a very low EP of only 22 mV, signaling a pronounced cochlear deterioration. Santos-Sacchi and Dallos (1983) found a stronger dye coupling among Hensen’s cells in deteriorated preparations. Nevertheless, Fig. 15 shows that dye coupling exists among DCs and between them and HCs even when the EP is high (81 mV1. The dye was

185

inje cted into a DC, according to the distance and phy’siological criteria. The superimposed fluorescencepha se contrast photomicrograph shows the dye in sev-

era1 HCs and DCs surrounding the bases of the 01 KS of the third row. No dye can be discerned in the OHCs.

photomicrograph on the left Fig. 16. An injected OHC, shown in a fluorescence :entration of the dye in the cell’s nucleus and its spread to a number of HCs spaces (compare char acteristic pattern of OHCs with clearly apparent intercellular ER = 100 MR; T = 1.6 min.

and in a superposition photomicrograph on the right. Not1e the (several stained nuclei visible in the left panel). Note alsoJ the with the following three figures). EP = 54 mV; RP = -63 mV, GZ-158.

Fig. 17. A massive spread of Lucifer yellow to many OHCs, DCs and HCs, shown in a fluorescence photomicrograph in the top panel and in a rposition photomicrograph in the bottom panel. According to physiological and distance criteria, an OHC was injected but the greatest dye dens ,ity is in HCs. The well-defined regions of the second and third rows of the OHCs are indicated by arrowheads. Note the many stained ntJclei in thle upper panel, and the characteristic pattern of the stained OHCs with well defined intercellular spaces in the lower panel. EP = 80 mV; RP=-71mV;ER=80MR;T=1.5min.GZ-174.

187

Fig. 38. Another example of wide dye spread to many OHCs and supporting cells, shown in a superposition photomicrograph. According to physic3logical and distance criteria an 0% was injected. Arrowheads marked 0,, 0, and O3 indicate same of the stained OHCs of the th.ree rows. The arrowhead marked P points to one of the stained outer pillars of Corti. The dye is also present in the inner pillars. In the upper ri,ght come r, an IHC supporting cell, stained during a separate microelectrode penetration can be seen. EP = 70 mV; RP = - 73 mV, ER = 1IO M n: T = 3 min. GZ-178.

Fig. 1Y. 5itill another example of massive dye spread to OHCs and supporting cells. According to distance and physiological criteria, an 0 IHC: was injec:ted. The arrowheads on the left indicate the three rows of OHCs; the arrowhead on the right points to a stained phalangeal proc ‘ess of a Deit er’s cell. Note the massive staining of Hensen’s cells and the presence of the dye in the pillars of Corti. EP = 75 mV: RP = - 73 mV: ER = 60 MO; T = 2 min. GZ-169.

189

In a second set of experiments, we attempted to impale OHCs for the purpose of direct staining. This was achieved in 6 preparations. With the exception of one, the dye always invaded neighboring OHCs, DCs and HCs, irrespective of the EP. Sometimes, its spread was quite extensive. Examples of these results are shown in Figs. 16, 17 and 18. A single stained OHC is dispIayed in Fig. 16, The panel on the left shows its fluorescence photomicrograph, the panel on the right, a superposition photomicrograph of the fluorescing cell with the corresponding phase-contrast image (same focus). Note that the dye was concentrated in the cell

nucleus and also diffused to HCs. Several stained HC nuclei are apparent in the left panel. Similar photomicrographs are shown in Fig. 17 for another preparation - fluorescence photomicrograph on the top, and superposition photomicrograph on the bottom. Note that many nuclei at the location of the OHC were stained. According to the superposition photomicrograph, some of the stained nuclei belonged to OHCs. Interestingly, the highest dye concentration appears in HCs although, according to physiological indices (distance beyond HCs, cell surrounded with 0 potential, alternating-potential amplitude higher than in HCs) an OHC

Fig. 20. Fluorescence photomicrograph of a IO-pm thick radial section of the organ of Corti embedded in plastic. The same specimen as in Fig. 19. Arrowheads on the left point to some stained HC nuclei, those in the middle, to stained nuclei of DCs, and those at the top to their stained phalangeal processes. The arrowheads furthest to the right point out a stained outer pillar of Corti, and its nucleus. Because of their strong autofluorescence it is not clear if any stain is left in the OHCs that can be seen between the reticular lamina and the bodies of DCs.

was injected. We observed more generally that HCs tended to accumulate more dye than the other cells of the organ of Corti.

Another example of multiple OHC staining is shown in Fig. 18 by means of a superposition photomicrograph. The highest dye concentration spreads radially

Fig, 21. Superposition photomicrographs of two adjacent regions of the same for 3 min into the intercellular space of the outer tunnel of Corti (Potential intracellularly through HCs. The arrowhead points out a small dye spill in a found in OHCs or any supporting cells in direct contact with the tunnel fluid. dye spread to several other HCs. EP = 75 mV; RP = -

organ of Corti. In the region on the left, Lucifer yellow was injected = 0 mV; ER = 30 MR). To do this, the micropipette was advanced HC, probably caused by negative-capacitance ringing. No stain was In the region on the right, the dye was injected into a HC. Note the 90 mV: ER = 20 Md): T = 2.5 min. GZ- 1x0.

191

from HCs to the third row of OHCs. Outside this area, the dye spreads over many HCs and OHCs. It can even be seen in the pillars of Corti. Note a stained supporting cell at IHCs in isolation in the upper right corner of the photograph. It was stained separately on a different electrode track. Still another example of massive dye spread from one OHC (according to physiological indices) to numerous neighboring OHCs, DCs and HCs and even to the pillars of Corti is shown in Fig. 19. The upper panel displays the fluorescence photomicrograph, the lower, the superposition photomicrograph. Again, the superposition indicates that the dye invaded the OHCs. Surprisingly, when this specimen was processed for analysis of 10 pm serial sections, we found no dye in the OHCs. One section is shown in Fig. 20. The highest dye concentration appears in DCs, including their phalangeal processes, but especially in their nuclei. The dye is also visible in the nuclei of HCs and in the body and nucleus of the outer pillar. The OHCs are visible through their autofluorescence resulting from the process of fixation for histology. No dye concentration can be ascertained in their nuclei above this background level. Since the dye appeared clearly in the OHCs in the surface preparation of the same specimen, we suspect that it diffused out during the chemical processing required for serial sectioning. Such diffusion has been observed in the past in other systems (retina; Engbretson and Solessio, 1990, personal communication). To test the rather remote possibility that Lucifer yellow entered the OHCs from intercellular space, we injected the dye into the intercellular space of the outer tunnel for 3 minutes, the longest time ever used by us for intracellular dye injections. The result is shown in Fig. 21 in the left panel. No dye is detectable in the outer tunnel, DCs or OHCs. A trace of dye is present in the middle of the HC population. It most probably stems from a small spill caused by negativecapacitance ringing used for cell penetration. Before entering the tunnel, the electrode went intracellularly through the interposed population of HCs as a means of locating the tunnel with the help of direct-potential change from strongly negative to 0. The right panel shows the result of dye injection into a HC through another hole in the same cochlear turn. The injection lasted for 2.5 min, and the dye presence in several HCs is unmistakable.

Discussion

and Conclusions

Our results on the distribution of alternating electrical potentials in the organ of Corti of Mongolian gerbils differ partially from the corresponding findings

obtained by Oesterle and Dallos (1989) on guinea pigs. The important difference concerns the balance between the alternating potentials measured in HCs and in the intercellular space, especially the outer tunnel of Corti. They found the amplitude of the potential to be smaller in these cells than in the tunnel; we found in the preponderant number of preparations the opposite to be true. Oesterle and Dallos’s finding is compatible with the absence of direct ionic coupling between the OHCs and the supporting cells but does not exclude it. Our results appear to require such a coupling. It does not seem plausible to conclude that the coupling exists in the gerbils but not in guinea pigs, both being rodents. We are more inclined to think that it exists in both species but does not show up unambiguously in the potential distribution of the guinea-pig organ of Corti. In this respect, it should be pointed out that the potential difference between HCs and the outer tunnel found by Oesterle and Dallos was rather small, and we would expect a larger one if the current generating the HC potential had to flow from the OHCs through the outer tunnel and, then, through the HC membrane not containing any specialization. On the other hand, the potential differences we found could have been contaminated by partial blockage of the recording micropipettes. If such blockage existed during the recording from the outer tunnel but not from the cells, the tunnel potential would have been artifactually lowered. Experience indicates, however, that micropipettes tend to become blocked inside the cells rather than in the intercellular space, and we tried to keep our micropipettes unobstructed by negative-capacitance ringing and resistance monitoring. In addition, piercing the wall of a cell with a micropipette can cause shunting of the membrane potential. Therefore, we think that the most likely artifacts would tend to lower the HC potentials relative to those of the tunnel, rather than raise them. It has been suggested that our finding of a greater alternating potential in HCs than in the outer tunnel could be due to capacitive coupling between OHCs and DCs. The potential induced in DCs would be transmitted to HCs. We doubt that this possibility is realistic. The coupling capacitance would have to be very large to effectively transmit low frequency potentials in view of a relatively low effective resistance of DCs coupled through gap junctions to HCs. Such a large capacitance would increase prohibitively the time constant of OHCs. If the capacitance were not very large, we would have found phase differences between HC potentials and those of OHCs at low frequencies, which we did not. Our results agree at least qualitatively with those of Oesterle and Dallos (1989) and also Santos-Sacchi and Dallos (1983) and Santos-Sacchi (1984) with respect to ionic coupling among the supporting cells, especially

192

HCs. However, the coupling inferred from our results appears to be stronger than the coupling indicated by theirs. We found the drop of the alternating potential between adjacent HCs to be too small to be measured. Santos-Sacchi and Dallos had found potential steps produced in one HC to be much reduced in an adjacent HC. However, Santos-Sacchi (1984) subsequently found a tighter coupling in an in-vitro preparation. Our dye-coupling experiments in situ agree with the in-vitro experiments of Santos-Sacchi (1986) with respect to dye coupling among the supporting cells. Both sets of experiments disagree with the earlier experiments of Santos-Sacchi and Dallos (1983) and show that the coupling does exist. However, our experiments are the first to demonstrate such a coupling in vivo and to indicate that it also exists between the OHCs and the supporting cells. This coupling is consistent with the amplitude of the alternating potential being higher in the HCs than in the outer tunnel. It is also consistent with some aspects of the morphological studies of Gulley and Reese (1976) and Nadol (1978) who did see membrane structures typical of gap junctions at appositions between mammalian hair cells and adjacent supporting cells. It is difficult to speculate why dye coupling between the OHC and the supporting cells was not demonstrated previously, especially, since we found it in well functioning cochleas with high CM and EP values as well as in deteriorated cochleas with low values of these potentials. A possible explanation is that gap junctions tend to be unstable (Peracchia and Peracchia, 1980a,b). More puzzling was the directionality of dye transfer we found. We were able to demonstrate dye spread to OHCs unequivocally only when an OHC was injected. When the dye was injected into HCs, the results were equivocal. Possibly, there is a direct current flowing from the OHCs through the DCs, HCs and CCs (Claudius cells) to the spiral ligament and stria vascularis. Such a current would close the loop for the current flowing from stria vascularis through the endolymph to the OHCs. According to our measurements, the pathway through gap junctions between the OHCs and the supporting cells and among the supporting cells has a lower resistance than the pathway through cell membranes and the intercellular spaces. It should also be mentioned that rectifying gap junctions are known to exist (Jaslove and Brink, 1987, for review). These gap junctions are voltage sensitive and allow the current to flow down the potential gradient (from less negative to more negative). In our experiments the OHCs had somewhat less negative resting potentials than did the supporting cells in any given preparation. Although, according to Oesterle and Dallos (1989), a strong statistical overlap exists between the resting potentials of the two populations, it is not clear if this overlap holds within preparations. Also, part of the

scatter may be due to membrane leakage produced by intracellular electrodes. Is it possible that our demonstration of dye coupling between the OHCs and the supporting cells was artifactual? We excluded the possibility that any of the cells in and around the outer tunnel picked up the dye from the extracellular space. But we were unable to demonstrate with certainty the transfer of dye to OHC’s when it was injected into a HC, or its presence in the OHCs in serial sections, even when it was injected directly into an OHC identified by physiological criteria. The presence of the dye in HCs after its injection into an OHC could be explained away by assuming that some of it was spilled during the intracellular traverse of the HC population by the electrode on its way to the OHCs. Such spills could have resulted from negativecapacitance ringing. However, this explanation does not appear convincing in view of the dye density found in the HCs (compare Figs. 17, 18 and 19 with Fig. 21). Neither can it explain the presence of the dye in more than one OHC after only one had been injected. Whether the staining of HCs was due to dye spills or not, staining of several OHCs required dye coupling between DCs and OHCs. We have already given a possible explanation for the absence of the dye in the serial sections of OHCs after it had been positively identified in the same specimen as a surface preparation. In this preparation, the dye was injected directly into an OHC according to physiological criteria. The same criteria led to direct staining of other OHCs, in particular, of the cell of Fig. 16, which was stained singly. We did not study systematically the possibility of ionic coupling between IHCs and their supporting cells. In none of our experiments did we detect any stain in the IHC. However, Goodman, Smith and Chamberlain (1982) found direct-potential changes produced by sound bursts to be larger in the supporting cells than in the intercellular space. This suggests direct ionic coupling between the IHCs and the supporting cells adjacent to them. Our results indicating the existence of direct ionic coupling between the OHCs and the supporting cells and among the supporting cells of the organ of Corti lead us to suggest a partial network diagram of the organ of Corti. It is shown in Fig. 22. The network applies to the receptor potentials generated in the OHCs. The IHCs and their supporting cells are not included for lack of unequivocal information. The horizontal thick lines indicate the borders of the organ of Corti at the basilar membrane and the reticular lamina, and the vertical one, the border of stria vascularis and the spiral ligament. The dashed lines schematize the borders between various cell types and the solid rectangles, the ionic-coupling impedances. Note that the pillars of Corti are coupled to OHCs and that the

193

coupling between HCs and CCs has not been investigated by us. From the point of view of the spiral ligament, the network contains three branches. The current in the lower two branches (CC and Scala tympani, ScT) is roughly in phase; the current in the upper branch (Scala media, SCM), in phase opposition. The sum of the currents and, with it, the voltage at the node in the spiral ligament should be 0. Such a cancellation in the vicinity of the node was found by Zidanic and Brownell (1989) and observed by us during CM recordings on the top of the spiral ligament. Partial cancellation was found along the spiral-ligament and stria-vascularis complex (Zwislocki, 1986). Theoretically, the voltage cancellation in the vicinity of the node can be obtained without the current path through the organ of Corti. However, such a path makes the measured current distribution in Scala tympani near the basilar membrane more plausible (Brownell, 1991, personal communication). Summarizing, our main conclusions for the Mongolian-gerbil cochlea are as follows: 1. The OHC alternating receptor potential usually produces an alternating potential of higher magnitude in HCs than in the outer tunnel without affecting the frequency pattern. 2. The phase of the alternating potential is practically the same in OHCs, HCs and the tunnel. 3. The phase of the alternating potential in HCs is opposite to the phase of CM in the endolymph, and its magnitude is usually greater than that of CM.

4. The relationship under 1. is difficult to explain without assuming a direct current path from OHCs to HCs through DCs. 5. Lucifer yellow injected into a HC diffuses to other HCs and also to DCs. Our results are equivocal with respect to its diffusion to OHCs. 6. Lucifer yellow injected into an OHC diffuses to supporting cells and other OHCs. 7. Lucifer yellow injected into the intercellular space of the outer tunnel is not picked up by any cells. 8. The electrical coupling and the dye coupling we have found are concordant in suggesting the existence of direct ionic coupling between OHCs and the supporting cells and among the supporting cells. 9. Why Lucifer yellow, when injected into a HC, rarely, if ever, invaded OHCs, but did invade the supporting cells and other OHCs, when injected into an OHC, will require further investigation.

Acknowledgments

We thank R. Mitchell for the manufacture and maintenance of the precision drill, D. Arpajian for computer graphics and Drs. S.C. Chamberlain, G. Engbretson, E.M. Relkin and J. Santos-Sacchi and also E. Solessio for helpful comments and suggestions. Work supported by NIDCD Claude Pepper Grant 5 R01 DC 00074.

SCM

ScT

Fig. 22. Suggested partial gross resistive network basilar membrane; ScT - Scala tympani; PC Claudius’ cells; SL - spiral ligament;

of the organ of Corti in a radial section. SCM - Scala media; RL - reticular lamina; BM pillars of Corti; OHC - outer hair cells; DC - Deiter’s cells; HC - Hensen’s cells; CC SV - stria vascularis. The space under OHC schematizes the outer tunnel of Corti.

104

References Dallas, P. (1983) Some electrical circuit properties of the organ of Corti. 1. Analysis without reactive elements. Hear. Res. 12. 89119. Dallas, P. (1984) Some electrical properties of the organ of Corti. Il. Analysis including reactive elements. Hear. Res. 281-291. Dallas. P., Santos-Sacchi, J. and Flock. A. (1982) Intracellular recordings from cochlear outer hair cells. Science 218, 582-584. Gulley, R.L. and Reese, T.S. (1976) Intercellular junctions in the reticular lamina of the organ of Corti. J. Neurophysiol. 5,479-507. Hama, K. (1980) Fine structure of the afferent synapse and gap junctions on the sensory hair cell in the saccular macula of goldfish: a freeze-fracture study. J. Neurocytol. 9, 845-860. Iurato, S., Franke, K., Luciano, L., Wermbler, G., Pannese, E. and Reale, E. (1976) Intercellular junctions in the organ of Corti as revealed by freeze fracturing. Acta Otolaryngol. (Stockh.) X2, 57-69. Jaslove, S.W. and Brink, P.R. (1987) Electronic coupling in the nervous system. In: W.C. De Mello (Ed.), Cell-To-Cell Communication, Plenum Publishing Corp., New York. Miller, M.R. and Beck, J. (1990) Innervation of a lizard auditory organ having gap junctions between most hair cells - A serial transmission electron microscopy Study. J. Comp. Neural. 293, 223-23s. Nadol, J.B. Jr. (1978) Intercellular junctions in the organ of Corti. Ann. Otol. Rhinol. Laryngol. 87. 70-80. Nadol, J.B. Jr., Mulroy, J.J., Goodenough, D.H. and Weiss, T.F. (1076) Tight and gap junctions in a vertebrate inner ear. Am. J. Anat. 147, 281-302. Oesterle, E.C. and Dallas, P. (1989) Intracellular recordings from supporting cells in the guinea-pig cochlea: AC potentials. J. Acoust. Sot. Am. 86, 1013-1032.

Peracchia. C. and Peracchia. L.L. (1980a). Gap junction dynamics: reversible effects of divalent cations. J. Cell Biol. 87. 708-718. Peracchia, C. and Peracchia, L.L. tlY80b). Gap junction dynamics: reversible effects of hydrogen ions. J. Cell Biol. 87, 273-279. Raviola, E. and Gilula, N.B. (1973) Gap junctions between photoreceptor cells in the vertehrate retina. Proc. Nat. Acad. Sci. USA 70, 1677-1681. Santos-Sacchi, J. (1984) A reevaluation of cell coupling in the organ of Corti. Hear. Res. 14, 203-204. Santos-Sacchi, J. (1986) Dye coupling in the organ of Corti. Cell Tissue Res. 245, 525-529. Santos-Sacchi, J. and Dallas. P. (1983) Intercellular communication in the supporting cells of the organ of Corti. Hear. Res. 9, 317-326. Schmiedt, R.A. and Zwislockt, J.J. (1977) Comparison of soundtransmission and cochlear-microphonic characteristics in Mongolian gerbil and guinea pig. J. Acoust. Sot. Am. 61, 133-149. Weiss. T.F., Peake. W.T. and Sohmer. H.S. (1971) Intracochlear potential recorded with micropipettes. II. Responses in cochlear scalae to tones. J. Acoust. Sot. Am. SO, 587-601. Zidanic. M. and Brownell, W.E. (1989) Low-frequency modulation of the intracochlear potential field: An energy source for the cochlear amplifier. II Valsalva 54. 20-27. Zwislocki, J.J. (1986) Changes in cochlear frequency selectivity produced by tectorial-membrane manipulation. In: B.C.J. Moore and R.D. Patterson (Eds.), Auditory Frequency Selectivity A NATO Advanced Workshop, Plenum Press, London. Zwislocki, J.J., Chamberlain, S.C. and Slepecky, N.B. (1988) Tectorial membrane I: Static mechanical properties in vivo. Hear. Res. 33. 2077222.