- Email: [email protected]
Longitudinal endolymph movements induced by perilymphatic injections
Hearing Research 123 (1998) 137^147
Longitudinal endolymph movements induced by perilymphatic injections Alec N. Salt *, John E. DeMott Department of Otolaryngology, Box 8115, Washington University School of Medicine, 517 South Euclid Avenue, St. Louis, MO 63110, USA Received 30 January 1998; revised 18 May 1998; accepted 23 May 1998
Abstract Endolymph movements and endocochlear potential (EP) changes were measured during disturbances of perilymphatic pressure, induced by injecting artificial perilymph into scala tympani (ST) or scala vestibuli (SV) of the guinea pig cochlea. Injections were performed either with or without an outlet made in the opposite perilymphatic scala. Injections into ST without an outlet induced large pressure changes but virtually no endolymph movement or EP change. Injection at the same rate into ST with an outlet in SV produced smaller pressure changes which were accompanied by a basally-directed displacement of endolymph and significant EP changes. The magnitude of endolymph displacements and EP changes varied as a function of injection rate. Injections into SV, either with or without an outlet in ST, produced apically-directed endolymph displacement and EP changes. For the SV injections without an outlet, the cochlear aqueduct and round window are likely to provide an outlet and compliance, permitting flow along the perilymphatic scalae to occur even when no ST outlet was provided. We conclude that endolymph movements are not dependent on the absolute pressure of the perilymph, but instead occur when small, sustained pressure gradients are present across the cochlear partition, corresponding to times when perilymph flow is induced. This study demonstrates that in the normal, sealed cochlea, endolymph and EP are insensitive to fluid injections into ST, but are sensitive to fluid injections into SV. Endolymph movements are therefore unlikely to be generated by cerebrospinal fluid pressure fluctuations (such as those produced by respiration, posture changes, coughing, sneezing, etc) which are transmitted to ST by the cochlear aqueduct. z 1998 Published by Elsevier Science B.V. All rights reserved. Key words: Endolymph; Perilymph; Flow; Endolymphatic hydrops; Perilymph ¢stula; Pressure
1. Introduction The biophysical relationships between volume, pressure and £ow of cochlear £uids are not well established. The use of marker ions as volume and £ow tracers has demonstrated that in the normal cochlea the rate of endolymph £ow is extremely low (Salt et al., 1986 ; Sykova et al., 1987; Salt and Thalmann, 1989), and does not make a signi¢cant contribution to the turnover of the major endolymph constituents (Salt and Thalmann, 1988). However, when volume disturbances are induced, longitudinal endolymph movements do play a role in the restoration of normal volume. Salt and DeMott (1997a) demonstrated that when small volumes were injected into endolymph, basally-directed * Corresponding author. Tel.: +1 (314) 362-7560; Fax: +1 (314) 362-7522; E-mail: [email protected]
£ow was induced, consistent with numerous studies which have demonstrated £ow towards the endolymphatic sac when markers were injected into endolymph in volume (Guild, 1927 ; Lundquist et al., 1964). In contrast, osmotic reduction of endolymph volume results in an apically-directed endolymph movement (Salt and DeMott, 1995), together with anatomical changes of the endolymphatic sac consistent with it performing a secretory, rather than resorptive role (Takumida et al., 1989). The relationships between the hydrostatic pressures of endolymph, perilymph and cerebrospinal £uid (CSF) have been investigated extensively. Many studies have shown that in the normal cochlea there is no measurable pressure di¡erence between endolymph and perilymph (Long and Morizono, 1987 ; Andrews et al., 1991 ; Bohmer, 1993). Both perilymph and endolymph resting pressures are dominated by cerebrospinal £uid
0378-5955 / 98 / $19.00 ß 1998 Published by Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 5 9 5 5 ( 9 8 ) 0 0 1 0 6 - 3
HEARES 3070 24-8-98
138
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147
pressure, transmitted to the cochlea by the cochlear aqueduct which opens into the basal turn of ST (Carlborg et al., 1992; Marchbanks and Reid, 1990). No pressure di¡erences between endolymph and perilymph were detected during a variety of manipulations, including posture changes, osmotic dehydration and cochlear perforations (Bohmer, 1993). These studies demonstrate that the boundary between endolymph and perilymph is highly compliant. Takeuchi et al. (1991) reported that identical pressure increases were measured in endolymph and perilymph during injection of 0.5^2 Wl of arti¢cial endolymph into scala media over a 2-min period. Similar observations were made by Salt and DeMott (1997b) in experiments where endolymph and perilymph pressures were monitored simultaneously during injections into endolymph or perilymph at rates less than 70 nl/min. These results con¢rm that part or all of the membranous labyrinth is extremely compliant, so that pressure changes throughout the labyrinth are equivalent with either endolymphatic or perilymphatic injection. In our earlier experiments, we showed that endolymph £ow was generated by endolymphatic injections (Salt and DeMott, 1997a). One issue raised by this study was whether the induced £ow could be related to pressure di¡erences between the inner ear £uids and the cerebrospinal £uid in the cranium. During any injection into the labyrinth, pressure gradients will be present across both the cochlear aqueduct and the endolymphatic sac. The possibility arose that longitudinal endolymphatic £ows could be related to pressure gradients between endolymph and the cranium, across the endolymphatic sac. The present study was designed to investigate this possibility by quantifying longitudinal endolymphatic movements during injections into the perilymphatic spaces, a procedure which also induces pressure gradients between endolymph and CSF across the endolymphatic sac. These studies were intended to further clarify the relationships between volume, £ow and pressure in the ear.
placement of electrodes, animals were immobilized with gallamine triethiodide (2^3 mg/kg) and arti¢cially ventilated, maintaining an end-tidal CO2 level close to 5%. This study required the use of 18 animals. 2.2. Endolymph £ow and area measurement Endolymph £ow and area changes were measured in the second turn using tetramethylammonium (TMA ) as a marker ion, as described in prior publications (Salt and DeMott, 1995, 1997a). Three electrodes were inserted into endolymph, separated along the lateral wall of the cochlea by approximately 0.5 mm, as illustrated in Fig. 1. The middle electrode of the three contained 160 mM TMACl, from which TMA was iontophoresed into endolymph by a current of 50 nA. This small current increased EP by approximately 0.2 mV, which was barely detectable relative to the usual recording noise. The other two electrodes, one placed apically and one placed basally to the site of iontophoresis, were each double-barreled TMA -selective electrodes, used to monitor the spread of TMA in the apical and basal directions. The TMA -selective electrodes were manufactured and calibrated in accordance with our previously published methods (Salt and Vora, 1991). In brief, double-barreled glass electrodes were pulled, and one barrel was silanized by exposure to dimethyldichlorosilane vapor (Sigma, St. Louis, MO, USA). The non-silanized, potential barrel of the electrode was ¢lled with 500 mM NaCl. The ion-selective barrel was ¢lled with 500 mM KCl and a small column of TMA -selec-
2. Methods 2.1. Animal preparation Pigmented guinea pigs weighing 300^500 g were anesthetized with a combination of sodium pentobarbital (25 mg/kg i.p.) and Innovar-vet (0.35 ml/kg i.m.). Body temperature was maintained at 38^39³C with a thermistor-controlled heating pad. The left external jugular vein was cannulated and sodium pentobarbital was given intravenously as required to maintain deep anesthesia. A tracheal cannulation was performed. The head was secured in a rigid head-holder and the right cochlea was exposed by a ventrolateral approach. Following the
Fig. 1. Schematic of electrode placement for the measurement of endolymph £ow during pressure manipulations of the perilymph. The cochlea is shown uncoiled, with the three scalae indicated (SV: scala vestibuli; SM: scala media; ST: scala tympani). The ductus reuniens (DR) and cochlear aqueduct (CA) are also indicated. Endolymph £ow was measured in turn II using tetramethylammonium (TMA) as a volume marker, detected by two TMA-selective electrodes placed approximately 0.5 mm on each side of the iontophoresis site. The distance between the electrodes has been exaggerated for clarity.
HEARES 3070 24-8-98
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147
tive exchanger was drawn into the tip. TMA -selective ion exchanger consisted of a 5% solution of potassium tetrakis-(4-chlorophenyl) borate in 2-nitrophenyloctylether (Fluka, Ronkonkoma, NY, USA). Electrodes were calibrated before and after use in a series of standard solutions at 38³C. The standards contained 0, 80, 160, 640 or 1280 WM TMACl in a background of 150 mM KCl. A non-linear function was ¢tted to the calibration data in accordance with the documented response characteristics for electrodes in low ion concentrations in the presence of a background interfering ion (Salt and Vora, 1991). For the 30 electrodes used in this study we found a mean slope of 79.8 mV/decade TMA change (S.D. 13.2) and a linear response detection limit of 169 WM (S.D. 108). Iontophoresis pipettes were made from single-barreled glass electrodes with internal ¢ber, beveled to a tip diameter of 2^3 Wm. Electrode tips were ¢lled to the shoulder with 0.5% agarose gel, to prevent volume displacement of the electrolyte. Electrodes were then back-¢lled with 160 mM TMACl and stored in the same solution to allow the electrolyte to equilibrate with the gel. An optically-coupled iontophoresis unit (World Precision Instruments, Sarasota, FL, USA ; Model 260) was used to eject TMA into endolymph. Flow and area changes of endolymph were derived using a mathematical model to interpret the two simultaneously-recorded time courses of TMA concentration. The model combined the e¡ects of ion di¡usion, longitudinal £ow and clearance to describe tracer dispersal along scala media as a function of time. The model was used to calculate the marker concentrations at the measurement sites and to track the recorded curves by sequentially determining the £ow and area values which best ¢t the recorded data in each 10-s interval, as detailed in previous publications (Salt and DeMott, 1995). In this manner, £ow and area time courses which produced the recorded curves were established. The method has been validated and shown to be accurate by £ow measurements in vitro (Salt and DeMott, 1997a). 2.3. Injections into perilymph Injections into perilymph were performed at rates of 0.75^10 Wl/min with an infusion pump (WPI model UMC4 micro-pump ¢tted with a 25-Wl or 50-Wl Hamilton gas-tight syringe). The pump connected directly to a glass pipette or 32-gauge needle with tip diameter of 100^200 Wm. The arti¢cial perilymph which was injected contained (in mM) NaCl (125), KCl (3.5), CaCl2 (1.3), NaHCO3 (25), MgCl2 (1.2), NaH2 PO4 (0.75) and glucose (5). After bubbling with 5% CO2 , the pH of this solution was close to 7.4. Injections were performed in the basal turn of either ST (as in Fig. 1) or SV in animals, with or without making an outlet for £uid e¥ux. In those experiments where an outlet was made, this was in the opposite scala of the basal turn,
139
i.e. SV for ST injections and vice versa. All injections were performed for at least a 5-min period. In the majority of experiments, two injections were performed in each animal, ¢rst without an outlet and then subsequently after the outlet had been made. To control for a possible sequence e¡ect, and because an outlet once made is di¤cult to re-seal, some experiments were entirely performed either with or without an outlet. No sequence e¡ects were apparent so all data were pooled according to injection condition. 2.4. Endolymph pressure measurement In separate experiments, endolymph pressure changes induced by injections into perilymph were recorded using a WPI model 900, servo-nulling micro pressure system. The recording pipette contained 3 M KCl and the tip was beveled to a diameter of approximately 5 Wm using a Narishige EG-40 pipette grinder. Pressure measurements were made in endolymph of the second cochlear turn. In all experiments the endocochlear potential (EP) was recorded from the reference barrel of both ion-selective electrodes inserted into scala media or from the pressure recording pipette. A microcomputer sampled and stored all potentials at 10-s intervals. Statistical signi¢cance was determined using Sigmastat software (Jandel Corp., San Rafael, CA, USA) by t-tests or rank sum tests as appropriate. This study was performed under NIH grant #DC 01368 titled `Inner Ear Fluid Interactions'. The care and use of the guinea pigs used for this study were approved by the Animal Studies Committee of Washington University, St. Louis, MO, Approved protocol number 95069. 3. Results 3.1. Injections into scala tympani A total of 22 injections were performed in 12 cochleas, 8 with no outlet and 14 in which an outlet was made in SV. The EP values measured before injection into ST were 80.6 mV (S.D. 6.6; n = 8) when there was no outlet and 75.6 mV (S.D. 8.13; n = 14) when there was an outlet in SV. The baseline endolymph £ow rate before injection averaged 30.067 mm/min (S.D. 0.074; n = 8) without an outlet and 30.012 mm/min (S.D. 0.030; n = 14) with an outlet in SV. The two groups were not signi¢cantly di¡erent with regard to baseline EP or endolymph £ow rate. An example of the £ow-marker time courses recorded simultaneously from two TMA-selective electrodes positioned in endolymph during injections into ST is shown in Fig. 2. The iontophoresis of TMA into
HEARES 3070 24-8-98
140
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147
Fig. 2. Data recorded during an experiment using tetramethylammonium (TMA) as a volume and £ow marker. At time zero, iontophoresis of TMA into endolymph commences, producing the rising curves of TMA concentration at the two recording electrodes. In this experiment, the apical electrode was at a distance of 0.375 mm from the injection site and the basal electrode was at a distance of 0.425 mm. Endocochlear potential (EP) recorded from the second turn is shown in the lower panel. At 20 min, an injection is performed into ST without an outlet present, which produces only minor changes in the rising TMA curves and in the EP. The same injection is then repeated after an outlet is made in SV. Under these conditions the TMA concentration at the basal electrode is increased and at the apical electrode is decreased, consistent with a basally-directed displacement of endolymph. Injection-induced EP changes were also larger after the outlet was made.
endolymph commences at time zero, which results in a progressive concentration increase at the locations of the two recording electrodes. The initial injection into ST was performed without making an outlet in the cochlea. This injection produced only minor changes in the rising TMA curve and had little e¡ect on the EP. The same injection was subsequently repeated after an outlet for £uid e¥ux was made in the basal turn of SV. Under these conditions, an increase in TMA concentration was seen at the basal electrode and a decrease at the apical electrode, indicating a basally-directed displacement of endolymph. EP de£ections were also greater when the injection occurred with an outlet present. The endolymph £ow and area changes derived by mathematical analysis of these tracer time courses are shown in Fig. 3. In addition to the endolymph £ow rate, the cumulative e¡ects of £ow, i.e the distance moved by endolymph is given. Flow toward the base is represented by positive values and £ow toward the apex as negative values. In the calculation of injection-induced displacement, the prevailing baseline £ow rate (in this example, 0.077 mm/min toward the apex) was subtracted. From this analysis it is apparent that the injection performed without an outlet gener-
ated minor £uctuations in both £ow and area, but no systematic changes. In contrast, after the outlet was made in scala vestibuli, £ow and area changes were greater. At the onset of the injection there was a transient movement of endolymph toward the basal turn, with a small increase in endolymph cross-section. A summary of a similar analysis of injections at 1.5 Wl/ min is given in Fig. 4. For 5 injections without an outlet present, the short-term e¡ects of injection into the perilymphatic space on £ow, area and EP are small, although there may be subsequent slowly-occurring changes which are less easily quanti¢ed by our present techniques. Injections often give rise to slow endolymph movements, although these varied from animal to animal and are seen here as an increasing variation of distance with time. In addition, the area of scala media tended to decline slowly in the period following injection into perilymph. When an outlet was present in SV, resulting in a perfusion of the perilymphatic space, larger changes of endolymph £ow, distance and EP were observed. For 6 injections into scala tympani at 1.5 Wl/ min, endolymph in the second turn was consistently displaced towards the base by an average of 0.16 mm (S.D. 0.02, n = 6) which was signi¢cantly greater than
HEARES 3070 24-8-98
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147
141
Fig. 3. Analysis of the data from Fig. 2 in which the £ow and area changes which gave rise to the recorded marker time courses were derived by a mathematical model. Upper panel: Endolymph £ow rate. Positive values indicate basally-directed £ow and negative values indicate apically-directed £ow. Middle panel: Injection-induced distance moved by endolymph, calculated by integrating the £ow rate above baseline over time. Positive values indicate basally-directed displacement. Lower panel: Scala media area changes as a percentage of the baseline pre-treatment area value. In this experiment, the injection into ST was ¢rst performed without an outlet in the cochlea and then after an outlet was made in the basal turn of SV.
the displacement of 30.002 (S.D. 0.05, n = 5) observed with no outlet present (t-test, P 6 0.0001). The endolymph displacement induced by perfusion onset was not a transient phenomenon, but was sustained throughout the injection period. The EP changes associated with injection were also greater when an outlet was present, corresponding to those which are welldocumented to be associated with perfusion. The EP showed an initial negative de£ection at the onset, followed by a slow increase during the injection so that EP commonly exceeded the pre-injection value. An additional positive de£ection was observed when injection ceased. The combined amplitude changes (the di¡erence
between the lowest and highest value during the injection period) averaged 3.02 mV (S.D. 1.57, n = 6) when an outlet was present, which was signi¢cantly greater than the mean change of 0.53 mV (S.D. 0.54, n = 5) without an outlet (t-test, P 6 0.0001). The distances moved by endolymph measured with and without and outlet over a range of injection rates are compared in Fig. 5. Only low injection rates could be evaluated without an outlet as higher rates caused substantial reductions in EP (data not presented here), which occurs when the pressure elevation disturbs blood £ow (Nakashima and Ito, 1981). Without an outlet, endolymphatic movements were minimal at all in-
HEARES 3070 24-8-98
142
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147
jection rates tested. In contrast, the degree of endolymph movement induced by injection with an outlet varied as a function of the perfusion rate. At a rate of 5 Wl/min, the rate widely used for perfusion by many groups, including our own, longitudinal endolymph movements averaged 0.37 mm (S.D. 0.14,
n = 6). Although this distance may seem small, when it is considered that there is only approximately 8 mm of scala media apical to the second turn measurement site, the displacement corresponds to almost 5% of the endolymph volume being displaced by perfusion at this rate. A ¢tted straight line through the di¡erent rates
Fig. 4. Summary of endolymphatic changes induced by 5 ST injections at 1.5 Wl/min when no outlet was present (left side) and 6 ST injections when an outlet was present in the basal turn of scala vestibuli (right side). Endolymph movements and EP changes are small when no outlet exists but are signi¢cantly larger when an outlet is present.
HEARES 3070 24-8-98
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147
tested indicates an endolymph movement of 0.071 mm for each Wl/min injection rate (correlation coe¤cient 0.75). The slope was similar (0.062 mm per Wl/min) with the single datum at 10 Wl/min excluded, although the correlation coe¤cient was lower (0.525). We found there was no apparent correlation between EP changes and the measured area changes of scala media. In contrast, the EP changes did correlate to some extent with the distance moved by endolymph (R2 = 0.458), as shown in Fig. 6. It can therefore be assumed in qualitative terms, that the induction of EP changes by `control' perfusions may suggest that longitudinal endolymph movements are being induced. However, since EP increase or decrease may occur with other aspects of the perfusion, such as the chemical composition of the medium, an EP change alone should not be interpreted as a de¢nite indicator of the existence of endolymph £ow. 3.2. Injections into scala vestibuli The results of injections into SV, with or without an outlet in ST, are shown in Fig. 7. Under the condition with no outlet, e¡ects of injection on both EP and endolymph movement were larger than those which were observed with ST injection. Under these conditions EP showed a mean increase of 1.2 mV (S.D. 3.53, n = 3) and endolymph showed an apically directed displacement of 0.10 mm (S.D. 0.21, n = 3). Making an outlet in ST marginally increased distance and EP changes, although not signi¢cantly so. The mean injection-induced EP increase with an outlet in ST was 2.8 mV (S.D. 4.3, n = 3) and the mean distance endolymph
Fig. 5. Summary of endolymphatic movements generated by injection at varying rates into ST without an outlet (open symbols; n = 8) or with an outlet in SV (closed symbols; n = 14). Injection with an outlet (equivalent to perilymphatic perfusion) induces longitudinal displacement of endolymph, which becomes greater at higher injection rates. For the no outlet condition, the number of experiments (given in parentheses) at each injection rate was 0.75 (2), 1.5 (5), 3 (1). For injections with an outlet the numbers at each injection rate were 1.5 (6), 3 (1), 5 (6), 10 (1).
143
Fig. 6. Relationship between EP change and longitudinal endolymph movement for ST injections with no outlet (open symbols) and with an outlet in SV (closed symbols).
was displaced towards the apex was 0.12 mm (S.D. 0.07, n = 3). The major di¡erence with SV injection was that the direction of endolymph movements and EP changes induced by injection were oppositely-directed to those seen with ST injections, with injections into SV generating apically-directed endolymph movements. The EP de£ections were positive at the start and negative at the end of injection, which was opposite those with ST injection, although the slow rise in EP occurring throughout the injection period was comparable for both ST and SV injections. 3.3. Injection-induced pressure changes In 3 separate experiments, the pressure increase in endolymph associated with injections either with or without outlets was measured. Fig. 8 shows the pressure increase recorded in endolymph during an injection into ST. In the initial situation without an outlet, there is a resting pressure of the perilymph, due to the intercommunication with CSF through the cochlear aqueduct. In this situation, injection at 1.5 Wl/min produces substantial pressure change throughout the inner ear. Pressure was increased by an average of 3.23 mm Hg (S.D. 0.73, n = 3). The EP disturbance associated with this pressure increase was minimal, averaging 0.3 mV (S.D. 0.28, n = 3) which was comparable to the measurements shown earlier in £ow-measurement experiments. The act of making an outlet in SV releases the resting pressure of the perilymph, with pressure falling by an average of 1.51 mmHg (S.D. 0.76, n = 3). In this condition, injections into ST at 1.5 Wl/min induced signi¢cantly smaller endolymph pressure increases, which averaged 0.58 mmHg (S.D. 0.47, n = 3; t = test, P = 0.012). In contrast, substantially larger disturbances of EP were observed when an outlet was present, which averaged 4.13 mV (S.D. 1.88, n = 3) and which were consistent with the data presented earlier.
HEARES 3070 24-8-98
144
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147
Fig. 7. Summary of endolymphatic changes induced by 3 SV injections at 1.5 Wl/min when no outlet was present (left side) and 3 injections when an outlet was present in the basal turn of ST (right side). Endolymph movements and EP changes were induced under both conditions. When no outlet was present, it is likely that the round window and cochlear aqueduct provided both compliance and an outlet from ST.
4. Discussion The major ¢nding of this study is that under certain conditions injection into the perilymphatic spaces induces longitudinal endolymph movements and associated
EP changes. Contrary to our expectations, these changes are not closely related to the pressure increase in the cochlear £uids with respect to the cranium, since injections into ST under sealed cochlear conditions generate large pressure increases in endolymph but induce
HEARES 3070 24-8-98
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147
Fig. 8. Fluid pressure measured in cochlear endolymph during injection into ST at 1.5 Wl/min, ¢rst without, then with an outlet in SV. Injections produce a large pressure increase when no outlet is present. Resting pressure is reduced and injection-induced pressure increases are smaller when an outlet is present. In contrast, simultaneously-measured EP changes are small when no outlet is present and are larger when an outlet is present.
negligible endolymph movement and EP change. On the other hand, injections into ST with an outlet in SV induce smaller measured pressure changes, but larger endolymph movements and EP changes. Our interpretation of these ¢ndings is that the longitudinal endolymph movements and EP changes are likely generated by extremely small pressure di¡erences across the cochlear partition, which occur when £ow occurs along the perilymphatic scalae. For injection into ST of the sealed cochlea, there is little compliance or outlet from the SV. We assume that the primary outlet for the injected solution will be the cochlear aqueduct so that £ow will be from the basal turn of ST toward the cochlear aqueduct and there will be virtually no £ow apical to the injection site in ST or in SV. As a result, it can be assumed that the large elevated pressure resulting from the injection is distributed uniformly throughout the cochlea without a signi¢cant pressure gradient across the cochlear partition. In contrast, when an outlet is made in SV the injected £uid will £ow apically in ST and basally in SV towards the outlet site. In order for this £ow to occur there must be a small pressure gradient between the injection and outlet sites. It follows that a pressure gradient must exist across the cochlear partition which is presumably su¤cient to displace the structure, resulting
145
in EP changes and longitudinal movements. This explanation is also consistent with our observations with SV injections. In the case of injections into SV without an outlet, the round window membrane and cochlear aqueduct provide sites of compliance and an outlet, respectively, so that injection will result in an apically directed £ow in SV and a basally directed £ow in ST. This will again result in a small pressure di¡erence across the cochlear partition, but in the opposite direction to the ST injection condition with an outlet. Exactly the same amount of £ow through the perilymphatic spaces will occur when an outlet is made in the base of ST, so it is not surprising that making the outlet has little e¡ect on the EP and endolymph £ow measured with SV injection. Our results are thus consistent with the view that procedures which generate small, sustained pressure di¡erentials across the cochlear partition may mechanically induce a disturbance of endolymph volume. One example, as demonstrated here, is cochlear perfusion, which increasingly displaces endolymph longitudinally as perfusion rate increases. This perfusion-induced endolymph displacement is not a transient phenomenon, but is sustained throughout the injection period. This contrasts with the impression, gained from EP recordings, that the mechanical e¡ects of perfusion are transient, seen primarily at the start and end of perfusion. Based on the degree of endolymph displacement, we expect the mechanical e¡ects of perfusion remain throughout the entire perfusion period. These studies raise the intriguing question of whether longitudinal endolymph movements are pathological, only occurring under abnormal conditions (such as cochlear perfusion), or whether longitudinal movements are a normal occurrence in an intact cochlea. The CSF undergoes substantial pressure £uctuations induced by respiration, posture changes, coughing, sneezing, etc., which are transmitted to the cochlea through the cochlear aqueduct (Marchbanks et al., 1987 ; Carlborg et al., 1992). However, as these £uctuations are applied to ST it would be unlikely, based on our ¢ndings, that they would induce endolymph movements or EP changes in the normal cochlea. In addition, Gopen et al. (1997) have demonstrated that the cochlear aqueduct and round window membrane act in combination as a low-pass ¢lter so that CSF-transmitted pressure £uctuations will be attenuated as they enter the normal cochlea. In contrast, when the otic capsule is perforated in SV, as in our experiments or as in a clinical case of a perilymph ¢stula near the stapes footplate, it is then likely that the endolymphatic system could be rendered more sensitive to mechanical disturbances by the ongoing, normal £uctuations of CSF pressure. The observation that pressures applied to SV may induce endolymphatic movements, raises the possibility that slow or sustained movements of the stapes could
HEARES 3070 24-8-98
146
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147
give rise to endolymphatic movements. Elicitation of the middle ear re£ex by loud sound, swallowing, vocalization, etc. is known to induce substantial stapes movement, as do changes of pressure across the tympanic membrane. Substantial movements of the tympanic membrane, and hence the stapes, may be induced by barometric pressure changes or the application of air pressure to the external canal, which is routinely used in impedance tympanometry. They can also be generated by middle ear pressure changes, such as those produced by the Valsalva manoeuver (swallowing while blowing gently with the mouth and nose closed) or in some cases produced by forceful sni¤ng. With elicitation of the acoustic re£ex there may be substantial displacement of the round window membrane (unpublished observation) illustrating the considerable degrees of stapes and £uid movements involved. Pressure measurements from the ear canal during the elicitation of the acoustic re£ex by Marchbanks et al. (1987) show the time course to be complex, varying with di¡erent individuals and under di¡erent experimental conditions. The possibility that transient or sustained movements of the middle ear structures may induce endolymph movements may also play a part in a number of middle ear manipulations which have been used as treatments for Meniere's disease. Montandon et al. (1989) reported that the insertion of tympanostomy tubes alleviated Meniere's symptoms. This ¢nding was supported by a study which showed signi¢cantly less development of endolymphatic hydrops in guinea pigs following ablation of the endolymphatic sac when the tympanic membrane was also perforated (Kimura and Hutta, 1997). Atmospheric and middle ear pressure £uctuations have also been reported to signi¢cantly reduce the development of hydrops in guinea pigs (Sakikawa and Kimura, 1997) and to be bene¢cial as a treatment for Meniere's disease (Densert, 1987 ; Densert et al., 1997). Our studies demonstrate that pressure changes applied to SV a¡ect the endolymphatic system causing signi¢cant endolymph movements. Further investigation is needed to determine whether endolymphatic movements are induced by movements of middle ear structures and whether middle ear structures could play some role in the regulation of endolymph volume. The mechanism by which endolymph is displaced longitudinally during injections into perilymph also remains uncertain. The ¢rst possibility is that the movement arises from the mechanical properties of the organ of Corti and Reissner's membrane. If both structures behave as elastic membranes then a pressure applied to one compartment could result in a pressure increase in the center compartment and potentially a longitudinal £ow towards a site of outlet. However, this type of model, even with two boundaries of unequal elasticity, does not account for our observation that the £ow direction di¡ers for injections applied to ST or SV, with
apically- directed £ow generated by SV injection. It has been reported by Hao and Khanna (1996) that in response to low frequency stimulation, Reissner's membrane and the organ of Corti show di¡erent amplitude and phase characteristics of vibration. In the situation where Reissner's membrane and organ of Corti displacements are unequal, then longitudinal endolymph £ow must take place. However, no study has yet reported the displacement of cochlear structures induced by sustained pressures, or whether the magnitude of any di¡erences would be su¤cient to account for the longitudinal endolymph movements we have observed here. The above reasoning assumes that the longitudinal displacements of endolymph are a mechanical response of the endolymphatic system for which a constant volume is assumed. It remains equally possible that the movement we observe may re£ect a change in total endolymph volume. Ion currents, transport processes and water movements may change very rapidly when the organ of Corti or Reissner's membrane is displaced. This could involve changes in current through the hair cells (Zidanic and Brownell, 1990), changes in transport by stria vascularis (Wangemann et al., 1995), or change in current through other cell types such as those of the outer sulcus (Marcus et al., 1997). The potential interactions between the many different structures bounding endolymph may lead to complex relationships between endolymph volume, endolymph concentration and EP which are only now beginning to characterize. Acknowledgments Supported by research grant number DC 01368 from the National Institute on Deafness and Other Communication Disorders, National Institutes of Health. References Andrews, J.C., Bohmer, A., Ho¡man, L., 1991. The measurement and manipulation of endolymphatic pressure in experimental endolymphatic hydrops. Laryngoscope 101, 661^668. Bohmer, A., 1993. Hydrostatic pressure in the inner ear £uid compartments and its e¡ects on inner ear function. Acta Otolaryngol. Suppl. 507, 5^24. Carlborg, B.I., Konradsson, K.S., Carlborg, A.H., Farmer, J.C., Jr., Densert, O., 1992. Pressure transfer between the perilymph and the cerebrospinal £uid compartments in cats. Am. J. Otol. 13, 41^48. Densert, B., 1987. E¡ects of overpressure on hearing function in Meniere's Disease. Acta Otolaryngol. 103, 32^42. Densert, B., Densert, O., Arlinger, S., Sass, K., Odkvist, L., 1997. Immediate e¡ects of middle ear pressure changes on the electrocochleographic recordings in patients with Meniere's disease: a clinical placebo-controlled study. Am. J. Otol. 18, 726^733. Gopen, Q., Rosowski, J.J., Merchant, S.N., 1997. Anatomy of the human cochlear aqueduct with functional implications. Hear. Res. 107, 9^22.
HEARES 3070 24-8-98
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147 Guild, S.R., 1927. The circulation of the endolymph. Am. J. Anat. 39, 57^81. Hao, L.F., Khanna, S.M., 1996. Reissner's membrane vibration in the apical turn of a living guinea pig cochlea. Hear. Res. 99, 176^189. Kimura, R.S., Hutta, J., 1997. Inhibition of experimentally induced hydrops by middle ear ventilation. Eur. Arch. Otorhinolaryngol. 254, 213^218. Long, C.H., Morizono, T., 1987. Hydrostatic pressure measurements of endolymph and perilymph in a guinea pig model of endolymphatic hydrops. Otolaryngol. Head Neck Surg. 96, 83^95. Lundquist, P.G., Kimura, R., Wersall, J., 1964. Experiments in endolymph circulation. Acta Otolaryngol. 188, 194^201. Marchbanks, R.J., Reid, A., Martin, A.M., Brightwell, A.P., Bateman, D., 1987. The e¡ect of raised intracranial pressure on intracochlear £uid pressure: three case studies. Br. J. Audiol. 21, 127^ 130. Marchbanks, R.J., Reid, A., 1990. Cochlear and cerebrospinal £uid pressure: their inter-relationship and control mechanisms. Br. J. Audiol. 24, 179^187. Marcus, D.C., Chiba, T., Wangemann, P., 1997. Vibrating probe measurements demonstrate active cation reabsorption in inner ear epithelia. J. Gen. Physiol. 110, 43a. Montandon, P., Guillemin, P., Hausler, R., 1989. Treatment of Meniere's disease by minor surgical procedures: sacculotomy, cochleosacculotomy, transtympanic ventilation tubes. In: Nadol, J.B., Kugler, Ghedini (Eds.), Meniere's Disease, 503^508. Nakashima, T., Ito, A., 1981. E¡ect of increased perilymphatic pressure on endocochlear potential. Ann. Otol. 90, 264^266. Sakikawa, Y., Kimura, R.S., 1997. Middle ear overpressure treatment of endolymphatic hydrops in guinea pigs. Otorhinolaryngology 59, 84^90. Salt, A.N., Thalmann, R., Marcus, D.C., Bohne, B.A., 1986. Direct measurement of longitudinal endolymph £ow rate in the guinea pig cochlea. Hear. Res. 23, 141^151.
147
Salt, A.N., Thalmann, R., 1988. Interpretation of endolymph £ow results. Hear. Res. 33, 279^281. Salt, A.N., Thalmann, R., 1989. Rate of longitudinal £ow of cochlear endolymph. In: Nadol, J.B. (Ed.), Meniere's Disease. Kugler, Amsterdam, pp. 69^73. Salt, A.N., Vora, A.R., 1991. Calibration of ion-selective microelectrodes for use with high levels of interfering ions. J. Neurosci. Methods 38, 233^237. Salt, A.N., DeMott, J.E., 1995. Endolymph volume changes during osmotic dehydration measured by two marker techniques. Hear. Res. 90, 12^23. Salt, A.N., DeMott, J.E., 1997a. Longitudinal endolymph £ow associated with acute volume increase in the guinea pig cochlea. Hear. Res. 107, 29^40. Salt, A.N., DeMott, J.E., 1997b. Undetectable pressure increase during induction of acute endolymphatic hydrops by microinjection. Assoc. Res. Otolaryngol. (Abstr.) 20, 12. Sykova, E., Syka, J., Johnstone, B.M., Yates, G.K., 1987. Longitudinal £ow of endolymph measured by distribution of tetraethylammonium and choline in scala media. Hear. Res. 28, 161^171. Takeuchi, S., Takeda, T., Saito, H., 1991. Pressure relationship between perilymph and endolymph associated with endolymphatic infusion. Ann. Otol. Rhinol. Laryngol. 100, 244^248. Takumida, M., Bagger-Sjoback, D., Rask-Andersen, H., 1989. The endolymphatic sac and inner ear homeostasis I: E¡ect of glycerol on the endolymphatic sac with or without colchicine pretreatment. Hear. Res. 40, 1^16. Wangemann, P., Liu, J., Marcus, D.C., 1995. Ion transport mechanisms responsible for K secretion and the transepithelial voltage across marginal cells of stria vascularis in vitro. Hear. Res. 84, 19^ 29. Zidanic, M., Brownell, W.E., 1990. Fine structure of the intracochlear potential ¢eld. I. The silent current. Biophys. J. 57, 1253^1268.
HEARES 3070 24-8-98
Longitudinal endolymph movements induced by perilymphatic injections Alec N. Salt *, John E. DeMott Department of Otolaryngology, Box 8115, Washington University School of Medicine, 517 South Euclid Avenue, St. Louis, MO 63110, USA Received 30 January 1998; revised 18 May 1998; accepted 23 May 1998
Abstract Endolymph movements and endocochlear potential (EP) changes were measured during disturbances of perilymphatic pressure, induced by injecting artificial perilymph into scala tympani (ST) or scala vestibuli (SV) of the guinea pig cochlea. Injections were performed either with or without an outlet made in the opposite perilymphatic scala. Injections into ST without an outlet induced large pressure changes but virtually no endolymph movement or EP change. Injection at the same rate into ST with an outlet in SV produced smaller pressure changes which were accompanied by a basally-directed displacement of endolymph and significant EP changes. The magnitude of endolymph displacements and EP changes varied as a function of injection rate. Injections into SV, either with or without an outlet in ST, produced apically-directed endolymph displacement and EP changes. For the SV injections without an outlet, the cochlear aqueduct and round window are likely to provide an outlet and compliance, permitting flow along the perilymphatic scalae to occur even when no ST outlet was provided. We conclude that endolymph movements are not dependent on the absolute pressure of the perilymph, but instead occur when small, sustained pressure gradients are present across the cochlear partition, corresponding to times when perilymph flow is induced. This study demonstrates that in the normal, sealed cochlea, endolymph and EP are insensitive to fluid injections into ST, but are sensitive to fluid injections into SV. Endolymph movements are therefore unlikely to be generated by cerebrospinal fluid pressure fluctuations (such as those produced by respiration, posture changes, coughing, sneezing, etc) which are transmitted to ST by the cochlear aqueduct. z 1998 Published by Elsevier Science B.V. All rights reserved. Key words: Endolymph; Perilymph; Flow; Endolymphatic hydrops; Perilymph ¢stula; Pressure
1. Introduction The biophysical relationships between volume, pressure and £ow of cochlear £uids are not well established. The use of marker ions as volume and £ow tracers has demonstrated that in the normal cochlea the rate of endolymph £ow is extremely low (Salt et al., 1986 ; Sykova et al., 1987; Salt and Thalmann, 1989), and does not make a signi¢cant contribution to the turnover of the major endolymph constituents (Salt and Thalmann, 1988). However, when volume disturbances are induced, longitudinal endolymph movements do play a role in the restoration of normal volume. Salt and DeMott (1997a) demonstrated that when small volumes were injected into endolymph, basally-directed * Corresponding author. Tel.: +1 (314) 362-7560; Fax: +1 (314) 362-7522; E-mail: [email protected]
£ow was induced, consistent with numerous studies which have demonstrated £ow towards the endolymphatic sac when markers were injected into endolymph in volume (Guild, 1927 ; Lundquist et al., 1964). In contrast, osmotic reduction of endolymph volume results in an apically-directed endolymph movement (Salt and DeMott, 1995), together with anatomical changes of the endolymphatic sac consistent with it performing a secretory, rather than resorptive role (Takumida et al., 1989). The relationships between the hydrostatic pressures of endolymph, perilymph and cerebrospinal £uid (CSF) have been investigated extensively. Many studies have shown that in the normal cochlea there is no measurable pressure di¡erence between endolymph and perilymph (Long and Morizono, 1987 ; Andrews et al., 1991 ; Bohmer, 1993). Both perilymph and endolymph resting pressures are dominated by cerebrospinal £uid
0378-5955 / 98 / $19.00 ß 1998 Published by Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 5 9 5 5 ( 9 8 ) 0 0 1 0 6 - 3
HEARES 3070 24-8-98
138
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147
pressure, transmitted to the cochlea by the cochlear aqueduct which opens into the basal turn of ST (Carlborg et al., 1992; Marchbanks and Reid, 1990). No pressure di¡erences between endolymph and perilymph were detected during a variety of manipulations, including posture changes, osmotic dehydration and cochlear perforations (Bohmer, 1993). These studies demonstrate that the boundary between endolymph and perilymph is highly compliant. Takeuchi et al. (1991) reported that identical pressure increases were measured in endolymph and perilymph during injection of 0.5^2 Wl of arti¢cial endolymph into scala media over a 2-min period. Similar observations were made by Salt and DeMott (1997b) in experiments where endolymph and perilymph pressures were monitored simultaneously during injections into endolymph or perilymph at rates less than 70 nl/min. These results con¢rm that part or all of the membranous labyrinth is extremely compliant, so that pressure changes throughout the labyrinth are equivalent with either endolymphatic or perilymphatic injection. In our earlier experiments, we showed that endolymph £ow was generated by endolymphatic injections (Salt and DeMott, 1997a). One issue raised by this study was whether the induced £ow could be related to pressure di¡erences between the inner ear £uids and the cerebrospinal £uid in the cranium. During any injection into the labyrinth, pressure gradients will be present across both the cochlear aqueduct and the endolymphatic sac. The possibility arose that longitudinal endolymphatic £ows could be related to pressure gradients between endolymph and the cranium, across the endolymphatic sac. The present study was designed to investigate this possibility by quantifying longitudinal endolymphatic movements during injections into the perilymphatic spaces, a procedure which also induces pressure gradients between endolymph and CSF across the endolymphatic sac. These studies were intended to further clarify the relationships between volume, £ow and pressure in the ear.
placement of electrodes, animals were immobilized with gallamine triethiodide (2^3 mg/kg) and arti¢cially ventilated, maintaining an end-tidal CO2 level close to 5%. This study required the use of 18 animals. 2.2. Endolymph £ow and area measurement Endolymph £ow and area changes were measured in the second turn using tetramethylammonium (TMA ) as a marker ion, as described in prior publications (Salt and DeMott, 1995, 1997a). Three electrodes were inserted into endolymph, separated along the lateral wall of the cochlea by approximately 0.5 mm, as illustrated in Fig. 1. The middle electrode of the three contained 160 mM TMACl, from which TMA was iontophoresed into endolymph by a current of 50 nA. This small current increased EP by approximately 0.2 mV, which was barely detectable relative to the usual recording noise. The other two electrodes, one placed apically and one placed basally to the site of iontophoresis, were each double-barreled TMA -selective electrodes, used to monitor the spread of TMA in the apical and basal directions. The TMA -selective electrodes were manufactured and calibrated in accordance with our previously published methods (Salt and Vora, 1991). In brief, double-barreled glass electrodes were pulled, and one barrel was silanized by exposure to dimethyldichlorosilane vapor (Sigma, St. Louis, MO, USA). The non-silanized, potential barrel of the electrode was ¢lled with 500 mM NaCl. The ion-selective barrel was ¢lled with 500 mM KCl and a small column of TMA -selec-
2. Methods 2.1. Animal preparation Pigmented guinea pigs weighing 300^500 g were anesthetized with a combination of sodium pentobarbital (25 mg/kg i.p.) and Innovar-vet (0.35 ml/kg i.m.). Body temperature was maintained at 38^39³C with a thermistor-controlled heating pad. The left external jugular vein was cannulated and sodium pentobarbital was given intravenously as required to maintain deep anesthesia. A tracheal cannulation was performed. The head was secured in a rigid head-holder and the right cochlea was exposed by a ventrolateral approach. Following the
Fig. 1. Schematic of electrode placement for the measurement of endolymph £ow during pressure manipulations of the perilymph. The cochlea is shown uncoiled, with the three scalae indicated (SV: scala vestibuli; SM: scala media; ST: scala tympani). The ductus reuniens (DR) and cochlear aqueduct (CA) are also indicated. Endolymph £ow was measured in turn II using tetramethylammonium (TMA) as a volume marker, detected by two TMA-selective electrodes placed approximately 0.5 mm on each side of the iontophoresis site. The distance between the electrodes has been exaggerated for clarity.
HEARES 3070 24-8-98
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147
tive exchanger was drawn into the tip. TMA -selective ion exchanger consisted of a 5% solution of potassium tetrakis-(4-chlorophenyl) borate in 2-nitrophenyloctylether (Fluka, Ronkonkoma, NY, USA). Electrodes were calibrated before and after use in a series of standard solutions at 38³C. The standards contained 0, 80, 160, 640 or 1280 WM TMACl in a background of 150 mM KCl. A non-linear function was ¢tted to the calibration data in accordance with the documented response characteristics for electrodes in low ion concentrations in the presence of a background interfering ion (Salt and Vora, 1991). For the 30 electrodes used in this study we found a mean slope of 79.8 mV/decade TMA change (S.D. 13.2) and a linear response detection limit of 169 WM (S.D. 108). Iontophoresis pipettes were made from single-barreled glass electrodes with internal ¢ber, beveled to a tip diameter of 2^3 Wm. Electrode tips were ¢lled to the shoulder with 0.5% agarose gel, to prevent volume displacement of the electrolyte. Electrodes were then back-¢lled with 160 mM TMACl and stored in the same solution to allow the electrolyte to equilibrate with the gel. An optically-coupled iontophoresis unit (World Precision Instruments, Sarasota, FL, USA ; Model 260) was used to eject TMA into endolymph. Flow and area changes of endolymph were derived using a mathematical model to interpret the two simultaneously-recorded time courses of TMA concentration. The model combined the e¡ects of ion di¡usion, longitudinal £ow and clearance to describe tracer dispersal along scala media as a function of time. The model was used to calculate the marker concentrations at the measurement sites and to track the recorded curves by sequentially determining the £ow and area values which best ¢t the recorded data in each 10-s interval, as detailed in previous publications (Salt and DeMott, 1995). In this manner, £ow and area time courses which produced the recorded curves were established. The method has been validated and shown to be accurate by £ow measurements in vitro (Salt and DeMott, 1997a). 2.3. Injections into perilymph Injections into perilymph were performed at rates of 0.75^10 Wl/min with an infusion pump (WPI model UMC4 micro-pump ¢tted with a 25-Wl or 50-Wl Hamilton gas-tight syringe). The pump connected directly to a glass pipette or 32-gauge needle with tip diameter of 100^200 Wm. The arti¢cial perilymph which was injected contained (in mM) NaCl (125), KCl (3.5), CaCl2 (1.3), NaHCO3 (25), MgCl2 (1.2), NaH2 PO4 (0.75) and glucose (5). After bubbling with 5% CO2 , the pH of this solution was close to 7.4. Injections were performed in the basal turn of either ST (as in Fig. 1) or SV in animals, with or without making an outlet for £uid e¥ux. In those experiments where an outlet was made, this was in the opposite scala of the basal turn,
139
i.e. SV for ST injections and vice versa. All injections were performed for at least a 5-min period. In the majority of experiments, two injections were performed in each animal, ¢rst without an outlet and then subsequently after the outlet had been made. To control for a possible sequence e¡ect, and because an outlet once made is di¤cult to re-seal, some experiments were entirely performed either with or without an outlet. No sequence e¡ects were apparent so all data were pooled according to injection condition. 2.4. Endolymph pressure measurement In separate experiments, endolymph pressure changes induced by injections into perilymph were recorded using a WPI model 900, servo-nulling micro pressure system. The recording pipette contained 3 M KCl and the tip was beveled to a diameter of approximately 5 Wm using a Narishige EG-40 pipette grinder. Pressure measurements were made in endolymph of the second cochlear turn. In all experiments the endocochlear potential (EP) was recorded from the reference barrel of both ion-selective electrodes inserted into scala media or from the pressure recording pipette. A microcomputer sampled and stored all potentials at 10-s intervals. Statistical signi¢cance was determined using Sigmastat software (Jandel Corp., San Rafael, CA, USA) by t-tests or rank sum tests as appropriate. This study was performed under NIH grant #DC 01368 titled `Inner Ear Fluid Interactions'. The care and use of the guinea pigs used for this study were approved by the Animal Studies Committee of Washington University, St. Louis, MO, Approved protocol number 95069. 3. Results 3.1. Injections into scala tympani A total of 22 injections were performed in 12 cochleas, 8 with no outlet and 14 in which an outlet was made in SV. The EP values measured before injection into ST were 80.6 mV (S.D. 6.6; n = 8) when there was no outlet and 75.6 mV (S.D. 8.13; n = 14) when there was an outlet in SV. The baseline endolymph £ow rate before injection averaged 30.067 mm/min (S.D. 0.074; n = 8) without an outlet and 30.012 mm/min (S.D. 0.030; n = 14) with an outlet in SV. The two groups were not signi¢cantly di¡erent with regard to baseline EP or endolymph £ow rate. An example of the £ow-marker time courses recorded simultaneously from two TMA-selective electrodes positioned in endolymph during injections into ST is shown in Fig. 2. The iontophoresis of TMA into
HEARES 3070 24-8-98
140
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147
Fig. 2. Data recorded during an experiment using tetramethylammonium (TMA) as a volume and £ow marker. At time zero, iontophoresis of TMA into endolymph commences, producing the rising curves of TMA concentration at the two recording electrodes. In this experiment, the apical electrode was at a distance of 0.375 mm from the injection site and the basal electrode was at a distance of 0.425 mm. Endocochlear potential (EP) recorded from the second turn is shown in the lower panel. At 20 min, an injection is performed into ST without an outlet present, which produces only minor changes in the rising TMA curves and in the EP. The same injection is then repeated after an outlet is made in SV. Under these conditions the TMA concentration at the basal electrode is increased and at the apical electrode is decreased, consistent with a basally-directed displacement of endolymph. Injection-induced EP changes were also larger after the outlet was made.
endolymph commences at time zero, which results in a progressive concentration increase at the locations of the two recording electrodes. The initial injection into ST was performed without making an outlet in the cochlea. This injection produced only minor changes in the rising TMA curve and had little e¡ect on the EP. The same injection was subsequently repeated after an outlet for £uid e¥ux was made in the basal turn of SV. Under these conditions, an increase in TMA concentration was seen at the basal electrode and a decrease at the apical electrode, indicating a basally-directed displacement of endolymph. EP de£ections were also greater when the injection occurred with an outlet present. The endolymph £ow and area changes derived by mathematical analysis of these tracer time courses are shown in Fig. 3. In addition to the endolymph £ow rate, the cumulative e¡ects of £ow, i.e the distance moved by endolymph is given. Flow toward the base is represented by positive values and £ow toward the apex as negative values. In the calculation of injection-induced displacement, the prevailing baseline £ow rate (in this example, 0.077 mm/min toward the apex) was subtracted. From this analysis it is apparent that the injection performed without an outlet gener-
ated minor £uctuations in both £ow and area, but no systematic changes. In contrast, after the outlet was made in scala vestibuli, £ow and area changes were greater. At the onset of the injection there was a transient movement of endolymph toward the basal turn, with a small increase in endolymph cross-section. A summary of a similar analysis of injections at 1.5 Wl/ min is given in Fig. 4. For 5 injections without an outlet present, the short-term e¡ects of injection into the perilymphatic space on £ow, area and EP are small, although there may be subsequent slowly-occurring changes which are less easily quanti¢ed by our present techniques. Injections often give rise to slow endolymph movements, although these varied from animal to animal and are seen here as an increasing variation of distance with time. In addition, the area of scala media tended to decline slowly in the period following injection into perilymph. When an outlet was present in SV, resulting in a perfusion of the perilymphatic space, larger changes of endolymph £ow, distance and EP were observed. For 6 injections into scala tympani at 1.5 Wl/ min, endolymph in the second turn was consistently displaced towards the base by an average of 0.16 mm (S.D. 0.02, n = 6) which was signi¢cantly greater than
HEARES 3070 24-8-98
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147
141
Fig. 3. Analysis of the data from Fig. 2 in which the £ow and area changes which gave rise to the recorded marker time courses were derived by a mathematical model. Upper panel: Endolymph £ow rate. Positive values indicate basally-directed £ow and negative values indicate apically-directed £ow. Middle panel: Injection-induced distance moved by endolymph, calculated by integrating the £ow rate above baseline over time. Positive values indicate basally-directed displacement. Lower panel: Scala media area changes as a percentage of the baseline pre-treatment area value. In this experiment, the injection into ST was ¢rst performed without an outlet in the cochlea and then after an outlet was made in the basal turn of SV.
the displacement of 30.002 (S.D. 0.05, n = 5) observed with no outlet present (t-test, P 6 0.0001). The endolymph displacement induced by perfusion onset was not a transient phenomenon, but was sustained throughout the injection period. The EP changes associated with injection were also greater when an outlet was present, corresponding to those which are welldocumented to be associated with perfusion. The EP showed an initial negative de£ection at the onset, followed by a slow increase during the injection so that EP commonly exceeded the pre-injection value. An additional positive de£ection was observed when injection ceased. The combined amplitude changes (the di¡erence
between the lowest and highest value during the injection period) averaged 3.02 mV (S.D. 1.57, n = 6) when an outlet was present, which was signi¢cantly greater than the mean change of 0.53 mV (S.D. 0.54, n = 5) without an outlet (t-test, P 6 0.0001). The distances moved by endolymph measured with and without and outlet over a range of injection rates are compared in Fig. 5. Only low injection rates could be evaluated without an outlet as higher rates caused substantial reductions in EP (data not presented here), which occurs when the pressure elevation disturbs blood £ow (Nakashima and Ito, 1981). Without an outlet, endolymphatic movements were minimal at all in-
HEARES 3070 24-8-98
142
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147
jection rates tested. In contrast, the degree of endolymph movement induced by injection with an outlet varied as a function of the perfusion rate. At a rate of 5 Wl/min, the rate widely used for perfusion by many groups, including our own, longitudinal endolymph movements averaged 0.37 mm (S.D. 0.14,
n = 6). Although this distance may seem small, when it is considered that there is only approximately 8 mm of scala media apical to the second turn measurement site, the displacement corresponds to almost 5% of the endolymph volume being displaced by perfusion at this rate. A ¢tted straight line through the di¡erent rates
Fig. 4. Summary of endolymphatic changes induced by 5 ST injections at 1.5 Wl/min when no outlet was present (left side) and 6 ST injections when an outlet was present in the basal turn of scala vestibuli (right side). Endolymph movements and EP changes are small when no outlet exists but are signi¢cantly larger when an outlet is present.
HEARES 3070 24-8-98
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147
tested indicates an endolymph movement of 0.071 mm for each Wl/min injection rate (correlation coe¤cient 0.75). The slope was similar (0.062 mm per Wl/min) with the single datum at 10 Wl/min excluded, although the correlation coe¤cient was lower (0.525). We found there was no apparent correlation between EP changes and the measured area changes of scala media. In contrast, the EP changes did correlate to some extent with the distance moved by endolymph (R2 = 0.458), as shown in Fig. 6. It can therefore be assumed in qualitative terms, that the induction of EP changes by `control' perfusions may suggest that longitudinal endolymph movements are being induced. However, since EP increase or decrease may occur with other aspects of the perfusion, such as the chemical composition of the medium, an EP change alone should not be interpreted as a de¢nite indicator of the existence of endolymph £ow. 3.2. Injections into scala vestibuli The results of injections into SV, with or without an outlet in ST, are shown in Fig. 7. Under the condition with no outlet, e¡ects of injection on both EP and endolymph movement were larger than those which were observed with ST injection. Under these conditions EP showed a mean increase of 1.2 mV (S.D. 3.53, n = 3) and endolymph showed an apically directed displacement of 0.10 mm (S.D. 0.21, n = 3). Making an outlet in ST marginally increased distance and EP changes, although not signi¢cantly so. The mean injection-induced EP increase with an outlet in ST was 2.8 mV (S.D. 4.3, n = 3) and the mean distance endolymph
Fig. 5. Summary of endolymphatic movements generated by injection at varying rates into ST without an outlet (open symbols; n = 8) or with an outlet in SV (closed symbols; n = 14). Injection with an outlet (equivalent to perilymphatic perfusion) induces longitudinal displacement of endolymph, which becomes greater at higher injection rates. For the no outlet condition, the number of experiments (given in parentheses) at each injection rate was 0.75 (2), 1.5 (5), 3 (1). For injections with an outlet the numbers at each injection rate were 1.5 (6), 3 (1), 5 (6), 10 (1).
143
Fig. 6. Relationship between EP change and longitudinal endolymph movement for ST injections with no outlet (open symbols) and with an outlet in SV (closed symbols).
was displaced towards the apex was 0.12 mm (S.D. 0.07, n = 3). The major di¡erence with SV injection was that the direction of endolymph movements and EP changes induced by injection were oppositely-directed to those seen with ST injections, with injections into SV generating apically-directed endolymph movements. The EP de£ections were positive at the start and negative at the end of injection, which was opposite those with ST injection, although the slow rise in EP occurring throughout the injection period was comparable for both ST and SV injections. 3.3. Injection-induced pressure changes In 3 separate experiments, the pressure increase in endolymph associated with injections either with or without outlets was measured. Fig. 8 shows the pressure increase recorded in endolymph during an injection into ST. In the initial situation without an outlet, there is a resting pressure of the perilymph, due to the intercommunication with CSF through the cochlear aqueduct. In this situation, injection at 1.5 Wl/min produces substantial pressure change throughout the inner ear. Pressure was increased by an average of 3.23 mm Hg (S.D. 0.73, n = 3). The EP disturbance associated with this pressure increase was minimal, averaging 0.3 mV (S.D. 0.28, n = 3) which was comparable to the measurements shown earlier in £ow-measurement experiments. The act of making an outlet in SV releases the resting pressure of the perilymph, with pressure falling by an average of 1.51 mmHg (S.D. 0.76, n = 3). In this condition, injections into ST at 1.5 Wl/min induced signi¢cantly smaller endolymph pressure increases, which averaged 0.58 mmHg (S.D. 0.47, n = 3; t = test, P = 0.012). In contrast, substantially larger disturbances of EP were observed when an outlet was present, which averaged 4.13 mV (S.D. 1.88, n = 3) and which were consistent with the data presented earlier.
HEARES 3070 24-8-98
144
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147
Fig. 7. Summary of endolymphatic changes induced by 3 SV injections at 1.5 Wl/min when no outlet was present (left side) and 3 injections when an outlet was present in the basal turn of ST (right side). Endolymph movements and EP changes were induced under both conditions. When no outlet was present, it is likely that the round window and cochlear aqueduct provided both compliance and an outlet from ST.
4. Discussion The major ¢nding of this study is that under certain conditions injection into the perilymphatic spaces induces longitudinal endolymph movements and associated
EP changes. Contrary to our expectations, these changes are not closely related to the pressure increase in the cochlear £uids with respect to the cranium, since injections into ST under sealed cochlear conditions generate large pressure increases in endolymph but induce
HEARES 3070 24-8-98
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147
Fig. 8. Fluid pressure measured in cochlear endolymph during injection into ST at 1.5 Wl/min, ¢rst without, then with an outlet in SV. Injections produce a large pressure increase when no outlet is present. Resting pressure is reduced and injection-induced pressure increases are smaller when an outlet is present. In contrast, simultaneously-measured EP changes are small when no outlet is present and are larger when an outlet is present.
negligible endolymph movement and EP change. On the other hand, injections into ST with an outlet in SV induce smaller measured pressure changes, but larger endolymph movements and EP changes. Our interpretation of these ¢ndings is that the longitudinal endolymph movements and EP changes are likely generated by extremely small pressure di¡erences across the cochlear partition, which occur when £ow occurs along the perilymphatic scalae. For injection into ST of the sealed cochlea, there is little compliance or outlet from the SV. We assume that the primary outlet for the injected solution will be the cochlear aqueduct so that £ow will be from the basal turn of ST toward the cochlear aqueduct and there will be virtually no £ow apical to the injection site in ST or in SV. As a result, it can be assumed that the large elevated pressure resulting from the injection is distributed uniformly throughout the cochlea without a signi¢cant pressure gradient across the cochlear partition. In contrast, when an outlet is made in SV the injected £uid will £ow apically in ST and basally in SV towards the outlet site. In order for this £ow to occur there must be a small pressure gradient between the injection and outlet sites. It follows that a pressure gradient must exist across the cochlear partition which is presumably su¤cient to displace the structure, resulting
145
in EP changes and longitudinal movements. This explanation is also consistent with our observations with SV injections. In the case of injections into SV without an outlet, the round window membrane and cochlear aqueduct provide sites of compliance and an outlet, respectively, so that injection will result in an apically directed £ow in SV and a basally directed £ow in ST. This will again result in a small pressure di¡erence across the cochlear partition, but in the opposite direction to the ST injection condition with an outlet. Exactly the same amount of £ow through the perilymphatic spaces will occur when an outlet is made in the base of ST, so it is not surprising that making the outlet has little e¡ect on the EP and endolymph £ow measured with SV injection. Our results are thus consistent with the view that procedures which generate small, sustained pressure di¡erentials across the cochlear partition may mechanically induce a disturbance of endolymph volume. One example, as demonstrated here, is cochlear perfusion, which increasingly displaces endolymph longitudinally as perfusion rate increases. This perfusion-induced endolymph displacement is not a transient phenomenon, but is sustained throughout the injection period. This contrasts with the impression, gained from EP recordings, that the mechanical e¡ects of perfusion are transient, seen primarily at the start and end of perfusion. Based on the degree of endolymph displacement, we expect the mechanical e¡ects of perfusion remain throughout the entire perfusion period. These studies raise the intriguing question of whether longitudinal endolymph movements are pathological, only occurring under abnormal conditions (such as cochlear perfusion), or whether longitudinal movements are a normal occurrence in an intact cochlea. The CSF undergoes substantial pressure £uctuations induced by respiration, posture changes, coughing, sneezing, etc., which are transmitted to the cochlea through the cochlear aqueduct (Marchbanks et al., 1987 ; Carlborg et al., 1992). However, as these £uctuations are applied to ST it would be unlikely, based on our ¢ndings, that they would induce endolymph movements or EP changes in the normal cochlea. In addition, Gopen et al. (1997) have demonstrated that the cochlear aqueduct and round window membrane act in combination as a low-pass ¢lter so that CSF-transmitted pressure £uctuations will be attenuated as they enter the normal cochlea. In contrast, when the otic capsule is perforated in SV, as in our experiments or as in a clinical case of a perilymph ¢stula near the stapes footplate, it is then likely that the endolymphatic system could be rendered more sensitive to mechanical disturbances by the ongoing, normal £uctuations of CSF pressure. The observation that pressures applied to SV may induce endolymphatic movements, raises the possibility that slow or sustained movements of the stapes could
HEARES 3070 24-8-98
146
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147
give rise to endolymphatic movements. Elicitation of the middle ear re£ex by loud sound, swallowing, vocalization, etc. is known to induce substantial stapes movement, as do changes of pressure across the tympanic membrane. Substantial movements of the tympanic membrane, and hence the stapes, may be induced by barometric pressure changes or the application of air pressure to the external canal, which is routinely used in impedance tympanometry. They can also be generated by middle ear pressure changes, such as those produced by the Valsalva manoeuver (swallowing while blowing gently with the mouth and nose closed) or in some cases produced by forceful sni¤ng. With elicitation of the acoustic re£ex there may be substantial displacement of the round window membrane (unpublished observation) illustrating the considerable degrees of stapes and £uid movements involved. Pressure measurements from the ear canal during the elicitation of the acoustic re£ex by Marchbanks et al. (1987) show the time course to be complex, varying with di¡erent individuals and under di¡erent experimental conditions. The possibility that transient or sustained movements of the middle ear structures may induce endolymph movements may also play a part in a number of middle ear manipulations which have been used as treatments for Meniere's disease. Montandon et al. (1989) reported that the insertion of tympanostomy tubes alleviated Meniere's symptoms. This ¢nding was supported by a study which showed signi¢cantly less development of endolymphatic hydrops in guinea pigs following ablation of the endolymphatic sac when the tympanic membrane was also perforated (Kimura and Hutta, 1997). Atmospheric and middle ear pressure £uctuations have also been reported to signi¢cantly reduce the development of hydrops in guinea pigs (Sakikawa and Kimura, 1997) and to be bene¢cial as a treatment for Meniere's disease (Densert, 1987 ; Densert et al., 1997). Our studies demonstrate that pressure changes applied to SV a¡ect the endolymphatic system causing signi¢cant endolymph movements. Further investigation is needed to determine whether endolymphatic movements are induced by movements of middle ear structures and whether middle ear structures could play some role in the regulation of endolymph volume. The mechanism by which endolymph is displaced longitudinally during injections into perilymph also remains uncertain. The ¢rst possibility is that the movement arises from the mechanical properties of the organ of Corti and Reissner's membrane. If both structures behave as elastic membranes then a pressure applied to one compartment could result in a pressure increase in the center compartment and potentially a longitudinal £ow towards a site of outlet. However, this type of model, even with two boundaries of unequal elasticity, does not account for our observation that the £ow direction di¡ers for injections applied to ST or SV, with
apically- directed £ow generated by SV injection. It has been reported by Hao and Khanna (1996) that in response to low frequency stimulation, Reissner's membrane and the organ of Corti show di¡erent amplitude and phase characteristics of vibration. In the situation where Reissner's membrane and organ of Corti displacements are unequal, then longitudinal endolymph £ow must take place. However, no study has yet reported the displacement of cochlear structures induced by sustained pressures, or whether the magnitude of any di¡erences would be su¤cient to account for the longitudinal endolymph movements we have observed here. The above reasoning assumes that the longitudinal displacements of endolymph are a mechanical response of the endolymphatic system for which a constant volume is assumed. It remains equally possible that the movement we observe may re£ect a change in total endolymph volume. Ion currents, transport processes and water movements may change very rapidly when the organ of Corti or Reissner's membrane is displaced. This could involve changes in current through the hair cells (Zidanic and Brownell, 1990), changes in transport by stria vascularis (Wangemann et al., 1995), or change in current through other cell types such as those of the outer sulcus (Marcus et al., 1997). The potential interactions between the many different structures bounding endolymph may lead to complex relationships between endolymph volume, endolymph concentration and EP which are only now beginning to characterize. Acknowledgments Supported by research grant number DC 01368 from the National Institute on Deafness and Other Communication Disorders, National Institutes of Health. References Andrews, J.C., Bohmer, A., Ho¡man, L., 1991. The measurement and manipulation of endolymphatic pressure in experimental endolymphatic hydrops. Laryngoscope 101, 661^668. Bohmer, A., 1993. Hydrostatic pressure in the inner ear £uid compartments and its e¡ects on inner ear function. Acta Otolaryngol. Suppl. 507, 5^24. Carlborg, B.I., Konradsson, K.S., Carlborg, A.H., Farmer, J.C., Jr., Densert, O., 1992. Pressure transfer between the perilymph and the cerebrospinal £uid compartments in cats. Am. J. Otol. 13, 41^48. Densert, B., 1987. E¡ects of overpressure on hearing function in Meniere's Disease. Acta Otolaryngol. 103, 32^42. Densert, B., Densert, O., Arlinger, S., Sass, K., Odkvist, L., 1997. Immediate e¡ects of middle ear pressure changes on the electrocochleographic recordings in patients with Meniere's disease: a clinical placebo-controlled study. Am. J. Otol. 18, 726^733. Gopen, Q., Rosowski, J.J., Merchant, S.N., 1997. Anatomy of the human cochlear aqueduct with functional implications. Hear. Res. 107, 9^22.
HEARES 3070 24-8-98
A.N. Salt, J.E. DeMott / Hearing Research 123 (1998) 137^147 Guild, S.R., 1927. The circulation of the endolymph. Am. J. Anat. 39, 57^81. Hao, L.F., Khanna, S.M., 1996. Reissner's membrane vibration in the apical turn of a living guinea pig cochlea. Hear. Res. 99, 176^189. Kimura, R.S., Hutta, J., 1997. Inhibition of experimentally induced hydrops by middle ear ventilation. Eur. Arch. Otorhinolaryngol. 254, 213^218. Long, C.H., Morizono, T., 1987. Hydrostatic pressure measurements of endolymph and perilymph in a guinea pig model of endolymphatic hydrops. Otolaryngol. Head Neck Surg. 96, 83^95. Lundquist, P.G., Kimura, R., Wersall, J., 1964. Experiments in endolymph circulation. Acta Otolaryngol. 188, 194^201. Marchbanks, R.J., Reid, A., Martin, A.M., Brightwell, A.P., Bateman, D., 1987. The e¡ect of raised intracranial pressure on intracochlear £uid pressure: three case studies. Br. J. Audiol. 21, 127^ 130. Marchbanks, R.J., Reid, A., 1990. Cochlear and cerebrospinal £uid pressure: their inter-relationship and control mechanisms. Br. J. Audiol. 24, 179^187. Marcus, D.C., Chiba, T., Wangemann, P., 1997. Vibrating probe measurements demonstrate active cation reabsorption in inner ear epithelia. J. Gen. Physiol. 110, 43a. Montandon, P., Guillemin, P., Hausler, R., 1989. Treatment of Meniere's disease by minor surgical procedures: sacculotomy, cochleosacculotomy, transtympanic ventilation tubes. In: Nadol, J.B., Kugler, Ghedini (Eds.), Meniere's Disease, 503^508. Nakashima, T., Ito, A., 1981. E¡ect of increased perilymphatic pressure on endocochlear potential. Ann. Otol. 90, 264^266. Sakikawa, Y., Kimura, R.S., 1997. Middle ear overpressure treatment of endolymphatic hydrops in guinea pigs. Otorhinolaryngology 59, 84^90. Salt, A.N., Thalmann, R., Marcus, D.C., Bohne, B.A., 1986. Direct measurement of longitudinal endolymph £ow rate in the guinea pig cochlea. Hear. Res. 23, 141^151.
147
Salt, A.N., Thalmann, R., 1988. Interpretation of endolymph £ow results. Hear. Res. 33, 279^281. Salt, A.N., Thalmann, R., 1989. Rate of longitudinal £ow of cochlear endolymph. In: Nadol, J.B. (Ed.), Meniere's Disease. Kugler, Amsterdam, pp. 69^73. Salt, A.N., Vora, A.R., 1991. Calibration of ion-selective microelectrodes for use with high levels of interfering ions. J. Neurosci. Methods 38, 233^237. Salt, A.N., DeMott, J.E., 1995. Endolymph volume changes during osmotic dehydration measured by two marker techniques. Hear. Res. 90, 12^23. Salt, A.N., DeMott, J.E., 1997a. Longitudinal endolymph £ow associated with acute volume increase in the guinea pig cochlea. Hear. Res. 107, 29^40. Salt, A.N., DeMott, J.E., 1997b. Undetectable pressure increase during induction of acute endolymphatic hydrops by microinjection. Assoc. Res. Otolaryngol. (Abstr.) 20, 12. Sykova, E., Syka, J., Johnstone, B.M., Yates, G.K., 1987. Longitudinal £ow of endolymph measured by distribution of tetraethylammonium and choline in scala media. Hear. Res. 28, 161^171. Takeuchi, S., Takeda, T., Saito, H., 1991. Pressure relationship between perilymph and endolymph associated with endolymphatic infusion. Ann. Otol. Rhinol. Laryngol. 100, 244^248. Takumida, M., Bagger-Sjoback, D., Rask-Andersen, H., 1989. The endolymphatic sac and inner ear homeostasis I: E¡ect of glycerol on the endolymphatic sac with or without colchicine pretreatment. Hear. Res. 40, 1^16. Wangemann, P., Liu, J., Marcus, D.C., 1995. Ion transport mechanisms responsible for K secretion and the transepithelial voltage across marginal cells of stria vascularis in vitro. Hear. Res. 84, 19^ 29. Zidanic, M., Brownell, W.E., 1990. Fine structure of the intracochlear potential ¢eld. I. The silent current. Biophys. J. 57, 1253^1268.
HEARES 3070 24-8-98