- Email: [email protected]
A reversible ischemia model in gerbil cochlea
ELSEVIER
Hearing Research 92 (1995) 30-37
A reversible ischemia model in gerbil cochlea Tianying Ren a,b,*, Nadine J. Brown a, Minsheng Zhang a, Alfred L. Nuttall a, Josef M. Miller
a
a Kresge Hearing Research Institute, Department of Otolaryngology, University of Michigan Medical School, 1301 East Ann Street, Ann Arbor, MI 48109-0506. USA b Department ofOtolaryngology, Xian Medical University, Xian, China
Received 8 April 1995; revised 25 August 1995; accepted 6 September 1995
Abstract
A completely reversible cochlear-ischemia animal model was developed, and an initial study of ischemia/reperfusion-induced cochlear function change is presented. The bulla of the anesthetized gerbil was opened through a ventral approach and the anterior inferior cerebellar artery and its branches were exposed. Cochlear blood flow (CBF) from the basal turn of the cochlea was monitored with a laser Doppler flowmeter. An electrically isolated microclamp was used to occlude the labyrinthine artery (LA). During LA clamping, the cubic distortion product (DP) was continuously recorded. The LA was repeatedly clamped for different durations in all animals, and CBF consistently showed full recovery after clamp release. No obvious change in vessel diameter or flow pattern was observed under a stereomicroscope. Mean blood pressure did not show significant change during clamping. Immediately upon LA clamping, CBF decreased rapidly nearly to zero. After clamp release, CBF demonstrated an immediate rapid increase, followed by a secondary gradual recovery to baseline. CBF recovery patterns were clamp duration-related. Within a few seconds of occlusion, DP decreased and reached a minimum of approximately 24% of the initial level in less than 30 s. Following reperfusion of the cochlea, DP gradually increased, decreased again, then slowly recovered. Time delay between CBF reperfusion and the first increase of DP was proportional to clamping duration, and the increased amplitudes demonstrated a negative relationship to clamp duration. We assume that the first decrease in DP during clamping was caused by ischemia in the cochlea; the second decrease, during the cochlear reperfusion, could be a form of reperfusion-induced change in cochlear function. This ischemia/reperfusion model in gerbil cochlea demonstrates excellent repeatability and reversibility. Since DP and other measurements can be used to dynamically monitor cochlear or hair cell functions, this model is useful in studies of auditory physiology and pathophysiology. Keywords: Cochlear blood flow; Laser Doppler flowmetry; Ischemia/reperfusion injury; Cochlea; Otoacoustic emission
1. I n t r o d u c t i o n Vascular disorders in the cochlea may be major contributing factors in a number of inner ear diseases (House, 1975; Short et al., 1985; Miller et al., 1986; Cole and Jahrdoerfer, 1988; Yamasoba et al., 1993). Particularly in sudden deafness, a vascular compromise has been proposed, and subsequently various agents such as vasodilating drugs and carbon dioxide are used to improve inner ear circulation, with some positive results (Huitcrantz, 1988). Although there have been numerous attempts to define the relationship between impairment of cochlear microcircula-
* Corresponding author. Tel.: (313) 764-6116; Fax: (313) 764-0014; E-mail: [email protected]. 0378-5955/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0378-5955(95)00192- 1
tion and these disorders, evidence for a vascular etiology is only indirect, and appropriate experimental animal models are needed for further investigation in this field. Toward understanding the mechanism of ischemia-induced damage in the cochlea, animal models of inner ear vascular disturbances have been of interest to the basic auditory researcher and to otologists for many decades (Kimura, t986). A common method used in the development of the animal model for sudden deafness of vascular origin is to damage or occlude the major supplying artery to the cochlea. When the anterior inferior cerebellar artery (AICA) is occluded close to the basilar artery, cat cochlea may or may not show pathologic alterations (Bernstein and Silverstein, 1966). When the common cochlear artery is obliterated in guinea pig ears, the most severe morphologi-
T. Ren et al. / Hearing Research 92 (1995) 3 0 - 3 7
cal lesions occur in the sensory cells of the cochlea, saccule, and posterior ampulla, (Kimura and Perlman, 1958a, b). In order to avoid surgical injury to the auditory nerve and increased pressure in the brain, Randolf et al. (1990) used microforceps to clamp AICA and induce acute transient local blood flow impairment in the guinea pig cochlea. They reported that compression of the cerebellar arteries resulted in variable effects on the laser Doppler flowmeter (LDF) signal from the cochlea. The AICA network supplying the rat cochlea has been described by Seidman and Quirk (1992). In our previous study (Ren et al., 1993, 1994), we found that clamping AICA of young adult guinea pigs did not result in a stable or constant decrease but, rather, caused a time-dependent dynamic response in the cochlear blood flow (CBF). A/CA clamping and release caused very regular and repeatable transient biphasic exponential changes in CBF. According to our model-based analysis of CBF responses to AICA clamping, the AICA contributes only about 45% of CBF to the cochlea; 55% of CBF must come from alternative supplying vessels (Ren et al., 1993). AICA clamping is, therefore, not a suitable model for investigation of ischemia effects in the guinea pig cochlea. However, Asai et al. (1993) found that photothrombotic occlusion of the AICA causes inner ear ischemia to various degrees, with or without alterations of the compound action potentials. Photochemical vascular occlusion is based on the production of oxygen radicals by photo-activated rose bengal. The differences between mechanical and photochemical A/CA occlusion were probably contributed by additional microvascular damage caused by oxygen radicals and singlets. These toxic molecules may have traveled with flowing blood to downstream vessels, causing cochlear microvasculature damage. Since the gerbil is one of the most commonly used animals for basic hearing research (Hellstrom and Schmiedt, 1990), a cochlear ischemia and reperfusion model was developed in this animal in the current study. The labyrinthine artery (LA) was exposed ventrally and repeatedly occluded with a custom-made microclamp. Even after multiple LA occlusions, LDF-monitored CBF and continuously recorded cubic distortion product (DP) of otoacoustic emission showed a full recovery. The current study demonstrates that LA clamping in the gerbil is an ideal model for study of ischemia and reperfusion injury in the cochlea.
2. Methods
2.1. Surgical approach The 13 healthy young Mongolian gerbils (45-65 g) used in this study were housed in American Association for Accreditation of Laboratory Animal Care-approved facilities. Experimental protocols were approved by the
31
>) ,..TJ '\
l
lT~
.'
Fig. 1. Diagram of the microclamp used for occlusion of the LA.
University of Michigan Committee on Use and Care of Animals. After tranquilization with pentobarbital sodium (30 m g / k g , i.p.), analgesia was provided by fentanyl (0.32 m g / k g , i.m.). Anesthesia was maintained by a supplementary half-dose of fentanyl every 45 min and half-dose of pentobarbital sodium every 60 min. After the animal's head was firmly fixed in a headholder, a tracheotomy was performed and a ventilation tube inserted into the trachea to insure free breathing. The left carotid artery was cannulated and connected to a pressure transducer (P23KL, Statham Instruments, Oxnard, CA) for continuous blood pressure (BP) measurement with a custom-made electrical circuit. Rectal temperature was maintained at 38 + I°C with a servo-regulated heating blanket. The bulla was exposed and opened through a ventral approach. After the stapedial artery was ligated and cut, the ventral and medial bulla wall was gently removed and the AICA and its branches were observed through the dura.
2.2. Microclamp and LA clamping For LA occlusion, a microclamp was designed based on our previous development (Ren et al., 1993) (Fig. 1). The blades of a No. 5 microsurgical forceps were separated and the two tips sharpened to approximately 5 /zm wide and 3 ;zm thick. To form the microclamp tip, the ends of the two tips were re-connected with a narrow piece of steel film. The resulting micro-clamp tip was firmly fixed to a plastic handle with a screw. The distal end of a camera shutter release cable was also attached to the handle at the screw. A short metal arm was soldered perpendicular to the steel film spring joining the clamp blades. Pressing the cable button advanced the end of the camera shutter cable to push against the spring arm, thus compressing the clamp tips. Electrical isolation between the clamp tip and the camera shutter release cable was achieved by two plastic washers. The entire microclamp system was held in a micromanipulator and positioned so that the LA was between the clamp tips. Clamping and release of the vessel
T. Ren et a l . / Hearing Research 92 (1995) 3 0 - 3 7
32
were achieved by depressing and releasing the cable button. When the system was in clamp position, the tension from the steel film spring was sufficient to completely occlude the vessel and overcome mechanical resistance from surface tension of fluid around the clamp tip but gentle enough not to damage the LA even after repeated occlusions. Since there is a great variation in vascular
anatomical characteristics within the surgical field, the LA was identified by CBF response to the vessel clamping.
2.3. Laser Doppler flowmetry (LDF) For LDF measurement of CBF, the cochlear mucosa on the first turn was gently removed with a small cotton
1!1
20O
A
BP
0 ~ 10.0 .~ 7.5 ~
5.0t 2.5
~
CBF
0.0 4O
DP go 0 ~ 180
o0t
....... Phase
-180 0
2
4
6
8
10
12
16
14
18
20
Time (min) 2O0
B
~/~BP 10.0
~ C B F
5° i 2.5
0.0
go
~
DP
0 180
gO0 g~ o
:
~
~
-
Phase
-90 -180 20
22
24
26
28
30
32
34
36
38
40
Time (rain)
Fig. 2. Response of BP, CBF and D P O A E in one animal to a series of L A clampings ranging in duration from 1 to 30 sec (A), 45 to 90 sec (B), 180 to 300 sec (C). A clear CBF decrease and increase is initiated by L A clamping and release. Complete CBF recovery occurs after multiple clamps. DP amplitude and phase show a dynamic response to the LA clamping. Formation of a clot in the fine catheter tip at about 14 min reduced the recording of BP oscillations in this animal. The catheter was flushed with heparinized saline at about 39 min.
T. Ren et a l . / Hearing Research 92 (1995) 3 0 - 3 7
pledget. A laser probe (PF 303, Perimed, Stockholm, Sweden) was held securely in a micromanipulator, and the probe tip was positioned against the lateral wall of the basal cochlear turn such that no motion occurred between the probe and cochlea. A light petroleum jelly was used between the laser probe tip and the surface of the cochlea to assure a constant optical coupling of the laser light to the cochlea and to prevent the accumulation of fluid and extraneous blood cells beneath the probe tip. After the probe tip was properly positioned, CBF was monitored with a Periflux PF2B laser Doppler flowmeter (Perimend, Stockholm, Sweden). The time constant was set to 0.2 s and the frequency cutoff to 12 kHz.
derive a reference signal ( 2 f l - f 2 , 5.2 kHz) from two primary tones. DP amplitude and relative phase to reference signal were continuously measured.
2.5. Experimental protocol For producing cochlear ischemia and reperfusion, the microclamp was used to briefly occlude the LA; BP, CBF and DP were simultaneously monitored. Each animal received 1, 2.5, 5, 10, 15, 30, 45, 60, 90, 180, and 300 s LA occlusions. The interval between occlusions was long enough to allow complete recovery of CBF and DP.
2.6. Data collection and process
2.4. Distortion product otoacoustic emission (DP)
BP, CBF and DP amplitude and phase were continuously collected using a computer-based chart recorder throughout the experiment. Sampling frequency of the data collection system was 2 or 4 Hz. CBF baseline (BL) was determined over 1 min immediately before each occlusion. The level of CBF measured with the LDF is expressed in arbitrary units based on voltage output from the instrument. BP and DP amplitude and phase are presented in absolute units: mm Hg, dB SPL, and degree.
Primary tones (fl and f2) were generated by two oscillators (Hewlett-Packard 3594A sweeping local oscillator and Fordham audiogenerator AG-260). Two primary tones were independently attenuated and applied to the test ear through separate earphone (Sony MDR-A22L). The frequency of f] was 6.6 kHz and of f2 8 kHz. The intensities of fl and f2 were 55 dB SPL. The output of each earphone was coupled to a 20-ga hypodermic needle and transmitted to a low-noise microphone (Etymotic Research ER-10B, Elk Grove Village, IL). The sound system was connected to the external ear canal by a plastic microphone tip and was calibrated in situ. In order to measure DP, the microphone output was introduced to a lock-in amplifier (Stanford Research Systems, Palo Alto, CA). A signal generator was used to
.I-
E E
2OO 150 100 5O
33
3. Results All animals tolerated the surgical procedure well. BP was 113.4 _+ 10.7 mm Hg (n = 13) immediately before the
C BP
0 10.0
¢U
~
5.0 2.5
C
B
F
0.0
413 ~
DP
20
go 0
g6o ~
-16O 2 ~
,
,
,
40
42
44
46
48
j
~
4
~
~
i
i
i
50
52
54
T i m e (rain)
Fig. 2. (continued).
~
P
h
56
a
s
e
i
i
58
6O
T. Ren et a l . / Hearing Research 92 (1995) 3 0 - 3 7
34
first occlusion and was maintained above 90 mm Hg throughout all experiments. A series of BP, CBF, and DP magnitude and phase responses to LA clamping of different durations in one animal is shown in Fig. 2A-C. Although irregular BP fluctuations are present, none are obviously clamp-related. The LA clamping initiates an abrupt CBF decrease, which is identical for different clamping durations. CBF reaches the minimum in a few seconds and maintains this level throughout the clamping duration. Upon clamp release, CBF increases immediately to a level above the baseline. The mean of original CBF decrease amplitude across all animals is 83.2 _+ 3.7% BL (n = 13). Amplitude and time delay of CBF recovery following the LA reperfusion appears clamping-duration related. The CBF pattern of response to 1 s LA clamp (Fig. 2A) is different from the response pattern for longer duration clamping, since the time required for full CBF decrease is longer than 1 s. A non-flow phenomenon was not observed during the LA reperfusion. Complete CBF recovery occurred after each clamping. The time required for CBF recovery seems to have a positive relationship to the clamping duration. A sharp negative and positive BP change in Fig. 2B at about 39 min was caused by flushing the catheter. Mean BP values before the flush show no significant difference from the value after the flush, although the baseline is more noisy, suggesting that even a flat BP trace probably indicates a correct or nearly correct mean BP. The DP amplitude shows a series of dynamic responses to the various clamp durations, except for the 2 s clamp which causes no obvious change in DP. A clamp of 5 s provokes only a brief DP decrease. DP responses to l0 and 15 s ischemia define the transition from 5 to 30 s or longer clamping. During the first 30 s clamp time, DP amplitude gradually decreases, reaches a minimum, and maintains this level throughout the clamp time. Even for 5 min LA clamping (Fig. 1C), no DP amplitude increase is observed
120"
\\\\ ~
6O 40 20
""e (s)
""
"
~o~
Fig. 3. The first 5 s of CBF responses to LA clamping of different durations from 1 animal, showing a typical exponential decrease.
180 8 o ~
120
~
~
:
"-
100
6O
~oo
40 ~ 2O
~o0 Fig. 4. Typical CBF response to the first 200 s of reperfusion plotted against clamping duration. After a short clamping, such as 5, 10, 15 or 30 s, CBF increases immediately and shows a brief overshoot. For longer duration clamping ( > 60 s), CBF overshoot reaches the maximum in approximately 150 s after reperfusion and then gradually recovers.
during the cochlear ischemia. The DP amplitude decrease across all animals is 76.3 -I- 5.2 BL (n = 13). On clamp release, a brief DP increase occurs following CBF increase. Amplitude of this DP increase is negatively correlated with clamping duration. This peak is not detectable from DP curves in response to > 60 s clamps. The second peak of DP amplitude can be identified from these responses to > 30 s clamps. Amplitude of the second DP peak seems to have a negative relationship to clamping duration. The third increase of DP amplitude is observed from responses to 30 s clamps. Amplitude of the third DP increase peak appears to be clamping-duration related. Although obvious overshoot of DP amplitude occurs after the LA reperfusion, no overshoot is observed before clamp release. The DP pattern in response to cochlear reperfusion is different from the CBF response pattern. A full recovery of DP is observed for all clamps. Full recovery time shows a positive relationship to clamping duration. DP phase shows changes that are consistent with DP magnitude. The first 5 s of CBF responses to LA clamping of different durations from one animal are plotted in Fig. 3. Despite different total clamping times, CBF shows a very similar decline. CBF changes show a typical exponential decrease with a time constant of approximately 0.83 s, which provides a measurement of cochlear vascular mechanics. The similarity of CBF decreases in the figure demonstrate excellent reliability for LA clamping-induced ischemia in the gerbil cochlea. CBF response to the first 200 s of reperfusion is plotted against different clamping durations (Fig. 4). After a short clamping, such as 5, 10, 15 or 30 s, CBF increases immediately and reaches an overshoot maximum in less than 2.5 s during LA reperfusion. The CBF overshoot completely recovers in less than 1 min. With clamping
T. Ren et a l . / Hearing Research 92 (1995) 3 0 - 3 7
140 " > 120 100 o_,,
80
~"
60
~
x
g
EJ
(
Fig. 5. DP amplitude during the first 200 s of reperfusion plotmd against clamping duration. Time delay of recovered DP exhibits a positive relationship to clamping duration. DP response patterns also appears to be clamping duration-related. Rough surface at long clamping duration range ( > 30 s) suggests fluctuations of DP amplitude.
duration increase, reperfusion-caused CBF increase rate becomes slower. For longer duration clamping ( > 60 s), CBF overshoot reaches the maximum at approximately 150 s during reperfusion. Full recovery time for these CBF overshoots is longer than 4 min. A long-lasting CBF overshoot occurs at approximately 150 s reperfusion time following a long-duration clamping ( > 60 s). In summary, Fig. 4 shows that CBF changes during cochlear reperfusion are reperfusion time- and ischemia duration-related. In order to show ischemia and reperfusion effects on DP, DP amplitude response during the first 200 s of reperfusion time from one animal is presented against the clamping duration in Fig. 5. DP response patterns appear clamping duration-related. For short duration clamping ( < 30 s), DP fully recovers in less than 60 s reperfusion time with no oscillation. The rough surface shown in the figure for long clamping duration range ( > 30 s) suggests fluctuations of DP amplitude. Overshoots and decreased amplitudes after overshoots seem to be clamping-duration related. DP full recovery time is positively related to clamping duration. These data show that following different durations of ischemia, cochlear reperfusion results in complicated multiphase DP responses. Compared to data in Fig. 4, DP amplitude responses are quite different from CBF response during the reperfusion.
4. Discussion
4.1. Reversibility and repeatability of the current animal model Reversibility is a basic requirement for a cochlear ischemia animal model for study of ischemia/reperfusion effects and other cochlear physiological and pathophysio-
35
logical mechanisms. Before LDF was introduced for the study of cochlear microcirculation (Miller et al., 1983), it was difficult, if not impossible, to continuously monitor CBF. Without the real time dynamic CBF measurement provided by LDF, it was impossible to directly study reversibility of the cochlear ischemia model. Although LDF has been used in a few studies of cochlear ischemia, no data are available which clearly indicate complete reversibility. Variation in the anatomy of the feeding vessels of the cochlea plus methods used to manipulate cochlear supplying vessels probably were responsible for incomplete reversibility in these models. To demonstrate LDF application in the cochlea, Miller et al. (1983) and Short et al. (1985) found that local pressure on the contents of the internal auditory meatus reduced the cochlear LDF signal to 15-30% BL in a reversible pattern. This preparation causes local pressure on the auditory nerve, with potential damaging and as well as stimulating effects. Therefore, it is inadequate as an ischemia model. Using a microclamp to mechanically occlude the AICA, Randolf et al. (1990) found a variable CBF decrease during the clamping and partial recovery after the release of the clamping. Clamp-caused mechanical damage may have contributed to their experimental results, since Ren et al. (1993, 1994) showed a complete CBF recovery following a large overshoot ( > 160% BL) after AICA clamping. Although Seidman et al. (1991) had investigated superoxide dismutase and allopurinol effects on ischemia and reperfusion in the rat cochlea, there was a little information on the CBF dynamic response to the clamping of the AICA. Since photochemically induced thrombosis was introduced to develop a cochlear ischemic model (Umenura et al., 1990; Asai et al., 1993), a great effort has been made to improve reversibility in this model. Because of difficulties in controlling thrombosis formation and resolution, full reversibility has not yet been achieved. Moreover, direct cochlear sensory cell damage caused by photochemically induced free radicals make interpretation of experimental results difficult. Data presented in Figs. 2 and 4 clearly show a full recovery following dynamic responses to different duration LA clamping, indicating complete reversibility of the current model. Good repeatability is illustrated also in Fig. 3. However, proper tension on the microclamp tip is essential: great enough to completely clamp the LA but soft enough to avoid any vascular damage. Quantitative calibration of the microclamp was not feasible because of the varying effects of vascular tension, fluid surface tension, viscosity-caused friction, and the non-linearity of the clamp spring and vascular wall mechanics. Therefore a practical approach was utilized to achieve repeatable non-invasive clamping in each animal. Tension on the clamp was increased until the LA blood flow completely stopped, as observed under a stereomicroscope. Then, the distance advanced was marked and fixed by a screw on the distal end of the camera release cable. This approach improved the efficiency of the ischemia/re-
36
T. Ren et al. / Hearing Research 92 (1995) 30-37
perfusion model, achieving full reversibility and good repeatability. 4.2. Controllable and quantitative ischemia in the cochlea A basic requirement for an ideal quantitative cochlear ischemia model is the ability to control and monitor the level and duration of ischemia. To the best of our knowledge, no cochlear ischemia model which fully meets these criteria has been reported in the literature. The dynamic and relatively quantitative properties of L D F have made it the most commonly used method to measure CBF. Instant change on L A clamping or release (Figs. 2 and 3) demonstrates the good time resolution in the LDF-measured C B F data obtained from gerbil. The L D F time resolution and the speed o f the L A clamping and releasing made a clear o n / o f f marker on the C B F signal. Thus, a controllable and quantitative duration o f cochlear ischemia was achieved in the gerbil cochlea. However, the autoregulatory response of cochlear vessels made it difficult to control the level of ischemia in the cochlea (Ren et al., 1993, 1994). Development of a microclamp with the capability of partial L A clamping may be a way to control ischemia level in future experiments. 4.3. Continuous cochlear function measurement during ischemia / reperfusion There are several methods, such as the compound action potential, cochlear microphonic potential, endocochlear potential, auditory evoked brain-stem response, and the different otoacoustic emissions, for revealing ischemia-caused cochlear function changes. Since i s c h e m i a / r e p e r f u s i o n damage is a time-dependent phenomenon, cochlear function should be monitored continuously a n d dynamically. W h e n two continuous tones with appropriate parameters are introduced in the ear, cubic combination tone at frequency of 2 f l - f 2 ( f l < f 2 ) can be generated in the cochlea and measured in the external ear canal (Brown and Kemp, 1984). This combination tone is c o m m o n l y known as the cubic DP and has been widely used as an indication of the non-linearity of the cochlear active process and normal cochlear function. Because of its high sensitivity and adequate time resolution, cubic DP was used to assess cochlear function in this experiment. Regarding the ischemia effect on otoacoustic emissions, Schmiedt and A d a m s (1981) reported that DP in the gerbil can persist up to 2 h after death from anoxia. They found that the cubic difference tone ( 2 f i e f 2) decreased initially, then recovered to approximately the normal magnitude. Schmiedt (1986) later noted that the effect of anoxia on DP ( 2 f l - f 2) is dependent on the level of the stimuli. K e m p and Brown (1984) found gerbil cubic DP declines 5 - 1 0 min after death. Lonsbury-Martin et al. (1987) found that a stimulus level of 75 dB SPL produced DP that lasted longer after death than the 65 dB SPL evoked emissions.
Whitehead et al. (1992) also noted a similar characteristic in the rabbit ear. Rebillard and Lavigne-Rebillard (1992) observed that after hypoxia, cubic DP in the guinea pig often initially increased (4 dB) then decreased. Data presented in Figs. 2 and 5 clearly show constant DP decrease during ischemia and dynamic change upon reperfusion. Transient DP decrease during cochlear reperfusion may indicate reperfusion-induced changes in cochlear function, since this DP change cannot be interpreted simply as ischemia-caused changes. The difference between previously reported data and the current experimental results may be due to the different animal model and protocol variances, since anoxia must have caused complicated systematic changes and uncontrollable cochlear metabolic disturbances. Since the current controllable quantitative ischemia model has complete reversibility and good repeatability, it will be useful in studies o f ischemia/reperfusion-induced cochlear damage and other basic mechanisms of cochlear physiology or pathophysiology.
Acknowledgements Authors gratefully acknowledge thoughtful suggestions given by Dr. A l f Axelsson and editorial help by Carole Tice. This work was supported by NIH Grants DC 00105 and A G 08885.
References Asai, Y., Umemura, K., Kohno, Y., Uematsu, T. and Nakashima, M. (1993) An animal model for hearing disturbance due to inner ear ischemia: photochemically induced thrombotic occlusion of the rat anterior inferior cerebellar artery. Eur. Arch. Oto-rhino-laryngol.250, 292-296. Bernstein, J.M. and Silverstein, H. (1966) Anterior cerebellar and labyrinthine arteries: a study in the cat. Arch. Otolaryngol. 83, 422-435. Brown, A.M. and Kemp, D.T. (1984) Suppressibility of the 2 f l - f 2 stimulated acoustic emissions in gerbil and man. Hear. Res. 13, 29-37. Cole, R.R. and Jahrdoerfer, R.A. (1988) Sudden hearing loss: an update. Am. J. Otol. 9, 211-215. Hellstrom, L.I. and Schmiedt, R.A. (1990) Compound action potential input/output functions in young and quiet-aged gerbils. Hear. Res. 50, 163-174. House, W. (1975) Mrni~re's disease. Management and theory. Otolaryngol. Clin. N. Am. 8, 515-535. Hultcrantz, E. (1988) Clinical treatment of vascular inner ear diseases. Am. J. Otolaryngol. 8, 317-322. Kemp, D.T. and Brown, A.M. (1984) Ear canal acoustic and round window electrical correlates of 2 f l - f 2 distortion generated in the cochlea. Hear. Res. 13, 39-46. Kimura, R. (1986) Animal models of inner ear vascular disturbances. Am. J. Otolaryngol. 7, 130-139. Kimura, R. and Perlman, H.B. (1958a) Arterial obstruction of the labyrinth. 1. Cochlear changes. Ann. Otol. Rhinol. Laryngol. 67, 5-25.
T. Ren et a l . / Hearing Research 92 (1995) 3 0 - 3 7
Kimura, R. and Perlman, H.B. (1958b) Arterial obstruction of the labyrinth. 2. Vestibular changes. Ann. Otol. Rhinol. Laryngol. 67, 26-41. Lonsbury-Martin, B.L., Martin, G.K., Probst, R. and Coats, A.C. (1987) Acoustic distortion products in rabbit ear canal. I. Basic features and physiological vulnerability. Hear. Res. 28, 173-189. Miller, J.M., Hultcrantz, E., Short, S.O. and Nuttall, A.L. (1986) Pharmacological effects on cochlear blood flow measured with the laser Doppler technique. Scand. Audiol. Suppl. 26, 11-20. Miller, J.M., Marks, N.J. and Goodwin, P.C. (1983) Laser Doppler measurements of cochlear blood flow. Hear. Res. 11,385-394. Randolf, J.B., Hanpt, H. and Scheibe, F. (1990) Cochlear blood flow following temporary occlusion of the cerebellar arteries. Eur. Arch. Otorhinolaryngol. 247, 226-228. Rebillard, G. and Lavigne-Rebillard, M. (1992) Effect of reversible hypoxia on the compared time courses of endocochlear potential and 2 f l - f 2 distortion products. Hear. Res. 62, 142-148. Ren, T., Nuttall, A.L. and Miller, J.M. (1993) Contribution of the anterior inferior cerebellar artery to cochlear blood flow in guinea pig: a model-based analysis. Hear. Res. 71, 91-97. Ren, T., Avinash, G.B., Nuttall, A.L., Miller, J.M., Laurikainen, E.A. and Quirk, W.S. (1994) Dynamic response of cochlear blood flow to occlusion of anterior inferior cerebellar artery in guinea pigs. J. Appl. Physiol. 76, 212-217. Schmiedt, R.A. (1986) Effects of asphyxia on levels of ear canal emissions in gerbil. Abstr. Assoc. Res. Otolaryngol. 9, 112-113.
37
Schmiedt, R.A. and Adams, J.C. (1981) Stimulated acoustic emissions in the ear canal of the gerbil. Hear. Res. 5,295-305. Seidman, M.D., and Quirk, W.S. (1992) The anterior inferior cerebellar arterial network supplying the rat cochlea and its role in autoregulation of cochlear blood flow. Eur. Arch. Otorhinolaryngol. 249, 332335. Seidman, M.D., Quirk, W.S., Nuttall, A.L. and Schweitzer, V. (1991) The effects of superoxide dismutase and allopurinol on ischemia and reperfusion in the cochlea. Otolaryngol. Head. Neck. Surg. 105, 457-461. Short, S.O., Goodwin, P.C., Kaplam, J.N. and Miller, J.M. (1985) Measuring cochlea blood flow with laser Doppler spectroscopy. Otolaryngol. Head Neck Surg. 43, 786-793. Umenura, K., Kohno, Y., Uematsu, T. and Nakashima, M. (1990) A new model for photochemically induced thrombosis in the inner ear microcirculation and the use of hearing loss as measure for microcirculatory disorders. Eur. Arch. Otorhinolaryngol. 248, 105-108. Whitehead, M.L., Lonsbury-Martin, B.L. and Martin, G.K. (1992) Evidence for two discrete sources of 2 f l - f 2 distortion-product otoacoustic emission in rabbit. II. Differential physiological vulnerability. J. Acoust. Soc. Am. 92, 2662-2682. Yamasoba, T., Kikuchi, S., Higo, R., Ofchi, T. and Tokumaru, A. (1993) Sudden sensorineural hearing loss associated with slow blood flow of the verterbrobasilar system. Ann. Otol. Rhinol. Laryngol. 102, 873877.
Hearing Research 92 (1995) 30-37
A reversible ischemia model in gerbil cochlea Tianying Ren a,b,*, Nadine J. Brown a, Minsheng Zhang a, Alfred L. Nuttall a, Josef M. Miller
a
a Kresge Hearing Research Institute, Department of Otolaryngology, University of Michigan Medical School, 1301 East Ann Street, Ann Arbor, MI 48109-0506. USA b Department ofOtolaryngology, Xian Medical University, Xian, China
Received 8 April 1995; revised 25 August 1995; accepted 6 September 1995
Abstract
A completely reversible cochlear-ischemia animal model was developed, and an initial study of ischemia/reperfusion-induced cochlear function change is presented. The bulla of the anesthetized gerbil was opened through a ventral approach and the anterior inferior cerebellar artery and its branches were exposed. Cochlear blood flow (CBF) from the basal turn of the cochlea was monitored with a laser Doppler flowmeter. An electrically isolated microclamp was used to occlude the labyrinthine artery (LA). During LA clamping, the cubic distortion product (DP) was continuously recorded. The LA was repeatedly clamped for different durations in all animals, and CBF consistently showed full recovery after clamp release. No obvious change in vessel diameter or flow pattern was observed under a stereomicroscope. Mean blood pressure did not show significant change during clamping. Immediately upon LA clamping, CBF decreased rapidly nearly to zero. After clamp release, CBF demonstrated an immediate rapid increase, followed by a secondary gradual recovery to baseline. CBF recovery patterns were clamp duration-related. Within a few seconds of occlusion, DP decreased and reached a minimum of approximately 24% of the initial level in less than 30 s. Following reperfusion of the cochlea, DP gradually increased, decreased again, then slowly recovered. Time delay between CBF reperfusion and the first increase of DP was proportional to clamping duration, and the increased amplitudes demonstrated a negative relationship to clamp duration. We assume that the first decrease in DP during clamping was caused by ischemia in the cochlea; the second decrease, during the cochlear reperfusion, could be a form of reperfusion-induced change in cochlear function. This ischemia/reperfusion model in gerbil cochlea demonstrates excellent repeatability and reversibility. Since DP and other measurements can be used to dynamically monitor cochlear or hair cell functions, this model is useful in studies of auditory physiology and pathophysiology. Keywords: Cochlear blood flow; Laser Doppler flowmetry; Ischemia/reperfusion injury; Cochlea; Otoacoustic emission
1. I n t r o d u c t i o n Vascular disorders in the cochlea may be major contributing factors in a number of inner ear diseases (House, 1975; Short et al., 1985; Miller et al., 1986; Cole and Jahrdoerfer, 1988; Yamasoba et al., 1993). Particularly in sudden deafness, a vascular compromise has been proposed, and subsequently various agents such as vasodilating drugs and carbon dioxide are used to improve inner ear circulation, with some positive results (Huitcrantz, 1988). Although there have been numerous attempts to define the relationship between impairment of cochlear microcircula-
* Corresponding author. Tel.: (313) 764-6116; Fax: (313) 764-0014; E-mail: [email protected]. 0378-5955/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0378-5955(95)00192- 1
tion and these disorders, evidence for a vascular etiology is only indirect, and appropriate experimental animal models are needed for further investigation in this field. Toward understanding the mechanism of ischemia-induced damage in the cochlea, animal models of inner ear vascular disturbances have been of interest to the basic auditory researcher and to otologists for many decades (Kimura, t986). A common method used in the development of the animal model for sudden deafness of vascular origin is to damage or occlude the major supplying artery to the cochlea. When the anterior inferior cerebellar artery (AICA) is occluded close to the basilar artery, cat cochlea may or may not show pathologic alterations (Bernstein and Silverstein, 1966). When the common cochlear artery is obliterated in guinea pig ears, the most severe morphologi-
T. Ren et al. / Hearing Research 92 (1995) 3 0 - 3 7
cal lesions occur in the sensory cells of the cochlea, saccule, and posterior ampulla, (Kimura and Perlman, 1958a, b). In order to avoid surgical injury to the auditory nerve and increased pressure in the brain, Randolf et al. (1990) used microforceps to clamp AICA and induce acute transient local blood flow impairment in the guinea pig cochlea. They reported that compression of the cerebellar arteries resulted in variable effects on the laser Doppler flowmeter (LDF) signal from the cochlea. The AICA network supplying the rat cochlea has been described by Seidman and Quirk (1992). In our previous study (Ren et al., 1993, 1994), we found that clamping AICA of young adult guinea pigs did not result in a stable or constant decrease but, rather, caused a time-dependent dynamic response in the cochlear blood flow (CBF). A/CA clamping and release caused very regular and repeatable transient biphasic exponential changes in CBF. According to our model-based analysis of CBF responses to AICA clamping, the AICA contributes only about 45% of CBF to the cochlea; 55% of CBF must come from alternative supplying vessels (Ren et al., 1993). AICA clamping is, therefore, not a suitable model for investigation of ischemia effects in the guinea pig cochlea. However, Asai et al. (1993) found that photothrombotic occlusion of the AICA causes inner ear ischemia to various degrees, with or without alterations of the compound action potentials. Photochemical vascular occlusion is based on the production of oxygen radicals by photo-activated rose bengal. The differences between mechanical and photochemical A/CA occlusion were probably contributed by additional microvascular damage caused by oxygen radicals and singlets. These toxic molecules may have traveled with flowing blood to downstream vessels, causing cochlear microvasculature damage. Since the gerbil is one of the most commonly used animals for basic hearing research (Hellstrom and Schmiedt, 1990), a cochlear ischemia and reperfusion model was developed in this animal in the current study. The labyrinthine artery (LA) was exposed ventrally and repeatedly occluded with a custom-made microclamp. Even after multiple LA occlusions, LDF-monitored CBF and continuously recorded cubic distortion product (DP) of otoacoustic emission showed a full recovery. The current study demonstrates that LA clamping in the gerbil is an ideal model for study of ischemia and reperfusion injury in the cochlea.
2. Methods
2.1. Surgical approach The 13 healthy young Mongolian gerbils (45-65 g) used in this study were housed in American Association for Accreditation of Laboratory Animal Care-approved facilities. Experimental protocols were approved by the
31
>) ,..TJ '\
l
lT~
.'
Fig. 1. Diagram of the microclamp used for occlusion of the LA.
University of Michigan Committee on Use and Care of Animals. After tranquilization with pentobarbital sodium (30 m g / k g , i.p.), analgesia was provided by fentanyl (0.32 m g / k g , i.m.). Anesthesia was maintained by a supplementary half-dose of fentanyl every 45 min and half-dose of pentobarbital sodium every 60 min. After the animal's head was firmly fixed in a headholder, a tracheotomy was performed and a ventilation tube inserted into the trachea to insure free breathing. The left carotid artery was cannulated and connected to a pressure transducer (P23KL, Statham Instruments, Oxnard, CA) for continuous blood pressure (BP) measurement with a custom-made electrical circuit. Rectal temperature was maintained at 38 + I°C with a servo-regulated heating blanket. The bulla was exposed and opened through a ventral approach. After the stapedial artery was ligated and cut, the ventral and medial bulla wall was gently removed and the AICA and its branches were observed through the dura.
2.2. Microclamp and LA clamping For LA occlusion, a microclamp was designed based on our previous development (Ren et al., 1993) (Fig. 1). The blades of a No. 5 microsurgical forceps were separated and the two tips sharpened to approximately 5 /zm wide and 3 ;zm thick. To form the microclamp tip, the ends of the two tips were re-connected with a narrow piece of steel film. The resulting micro-clamp tip was firmly fixed to a plastic handle with a screw. The distal end of a camera shutter release cable was also attached to the handle at the screw. A short metal arm was soldered perpendicular to the steel film spring joining the clamp blades. Pressing the cable button advanced the end of the camera shutter cable to push against the spring arm, thus compressing the clamp tips. Electrical isolation between the clamp tip and the camera shutter release cable was achieved by two plastic washers. The entire microclamp system was held in a micromanipulator and positioned so that the LA was between the clamp tips. Clamping and release of the vessel
T. Ren et a l . / Hearing Research 92 (1995) 3 0 - 3 7
32
were achieved by depressing and releasing the cable button. When the system was in clamp position, the tension from the steel film spring was sufficient to completely occlude the vessel and overcome mechanical resistance from surface tension of fluid around the clamp tip but gentle enough not to damage the LA even after repeated occlusions. Since there is a great variation in vascular
anatomical characteristics within the surgical field, the LA was identified by CBF response to the vessel clamping.
2.3. Laser Doppler flowmetry (LDF) For LDF measurement of CBF, the cochlear mucosa on the first turn was gently removed with a small cotton
1!1
20O
A
BP
0 ~ 10.0 .~ 7.5 ~
5.0t 2.5
~
CBF
0.0 4O
DP go 0 ~ 180
o0t
....... Phase
-180 0
2
4
6
8
10
12
16
14
18
20
Time (min) 2O0
B
~/~BP 10.0
~ C B F
5° i 2.5
0.0
go
~
DP
0 180
gO0 g~ o
:
~
~
-
Phase
-90 -180 20
22
24
26
28
30
32
34
36
38
40
Time (rain)
Fig. 2. Response of BP, CBF and D P O A E in one animal to a series of L A clampings ranging in duration from 1 to 30 sec (A), 45 to 90 sec (B), 180 to 300 sec (C). A clear CBF decrease and increase is initiated by L A clamping and release. Complete CBF recovery occurs after multiple clamps. DP amplitude and phase show a dynamic response to the LA clamping. Formation of a clot in the fine catheter tip at about 14 min reduced the recording of BP oscillations in this animal. The catheter was flushed with heparinized saline at about 39 min.
T. Ren et a l . / Hearing Research 92 (1995) 3 0 - 3 7
pledget. A laser probe (PF 303, Perimed, Stockholm, Sweden) was held securely in a micromanipulator, and the probe tip was positioned against the lateral wall of the basal cochlear turn such that no motion occurred between the probe and cochlea. A light petroleum jelly was used between the laser probe tip and the surface of the cochlea to assure a constant optical coupling of the laser light to the cochlea and to prevent the accumulation of fluid and extraneous blood cells beneath the probe tip. After the probe tip was properly positioned, CBF was monitored with a Periflux PF2B laser Doppler flowmeter (Perimend, Stockholm, Sweden). The time constant was set to 0.2 s and the frequency cutoff to 12 kHz.
derive a reference signal ( 2 f l - f 2 , 5.2 kHz) from two primary tones. DP amplitude and relative phase to reference signal were continuously measured.
2.5. Experimental protocol For producing cochlear ischemia and reperfusion, the microclamp was used to briefly occlude the LA; BP, CBF and DP were simultaneously monitored. Each animal received 1, 2.5, 5, 10, 15, 30, 45, 60, 90, 180, and 300 s LA occlusions. The interval between occlusions was long enough to allow complete recovery of CBF and DP.
2.6. Data collection and process
2.4. Distortion product otoacoustic emission (DP)
BP, CBF and DP amplitude and phase were continuously collected using a computer-based chart recorder throughout the experiment. Sampling frequency of the data collection system was 2 or 4 Hz. CBF baseline (BL) was determined over 1 min immediately before each occlusion. The level of CBF measured with the LDF is expressed in arbitrary units based on voltage output from the instrument. BP and DP amplitude and phase are presented in absolute units: mm Hg, dB SPL, and degree.
Primary tones (fl and f2) were generated by two oscillators (Hewlett-Packard 3594A sweeping local oscillator and Fordham audiogenerator AG-260). Two primary tones were independently attenuated and applied to the test ear through separate earphone (Sony MDR-A22L). The frequency of f] was 6.6 kHz and of f2 8 kHz. The intensities of fl and f2 were 55 dB SPL. The output of each earphone was coupled to a 20-ga hypodermic needle and transmitted to a low-noise microphone (Etymotic Research ER-10B, Elk Grove Village, IL). The sound system was connected to the external ear canal by a plastic microphone tip and was calibrated in situ. In order to measure DP, the microphone output was introduced to a lock-in amplifier (Stanford Research Systems, Palo Alto, CA). A signal generator was used to
.I-
E E
2OO 150 100 5O
33
3. Results All animals tolerated the surgical procedure well. BP was 113.4 _+ 10.7 mm Hg (n = 13) immediately before the
C BP
0 10.0
¢U
~
5.0 2.5
C
B
F
0.0
413 ~
DP
20
go 0
g6o ~
-16O 2 ~
,
,
,
40
42
44
46
48
j
~
4
~
~
i
i
i
50
52
54
T i m e (rain)
Fig. 2. (continued).
~
P
h
56
a
s
e
i
i
58
6O
T. Ren et a l . / Hearing Research 92 (1995) 3 0 - 3 7
34
first occlusion and was maintained above 90 mm Hg throughout all experiments. A series of BP, CBF, and DP magnitude and phase responses to LA clamping of different durations in one animal is shown in Fig. 2A-C. Although irregular BP fluctuations are present, none are obviously clamp-related. The LA clamping initiates an abrupt CBF decrease, which is identical for different clamping durations. CBF reaches the minimum in a few seconds and maintains this level throughout the clamping duration. Upon clamp release, CBF increases immediately to a level above the baseline. The mean of original CBF decrease amplitude across all animals is 83.2 _+ 3.7% BL (n = 13). Amplitude and time delay of CBF recovery following the LA reperfusion appears clamping-duration related. The CBF pattern of response to 1 s LA clamp (Fig. 2A) is different from the response pattern for longer duration clamping, since the time required for full CBF decrease is longer than 1 s. A non-flow phenomenon was not observed during the LA reperfusion. Complete CBF recovery occurred after each clamping. The time required for CBF recovery seems to have a positive relationship to the clamping duration. A sharp negative and positive BP change in Fig. 2B at about 39 min was caused by flushing the catheter. Mean BP values before the flush show no significant difference from the value after the flush, although the baseline is more noisy, suggesting that even a flat BP trace probably indicates a correct or nearly correct mean BP. The DP amplitude shows a series of dynamic responses to the various clamp durations, except for the 2 s clamp which causes no obvious change in DP. A clamp of 5 s provokes only a brief DP decrease. DP responses to l0 and 15 s ischemia define the transition from 5 to 30 s or longer clamping. During the first 30 s clamp time, DP amplitude gradually decreases, reaches a minimum, and maintains this level throughout the clamp time. Even for 5 min LA clamping (Fig. 1C), no DP amplitude increase is observed
120"
\\\\ ~
6O 40 20
""e (s)
""
"
~o~
Fig. 3. The first 5 s of CBF responses to LA clamping of different durations from 1 animal, showing a typical exponential decrease.
180 8 o ~
120
~
~
:
"-
100
6O
~oo
40 ~ 2O
~o0 Fig. 4. Typical CBF response to the first 200 s of reperfusion plotted against clamping duration. After a short clamping, such as 5, 10, 15 or 30 s, CBF increases immediately and shows a brief overshoot. For longer duration clamping ( > 60 s), CBF overshoot reaches the maximum in approximately 150 s after reperfusion and then gradually recovers.
during the cochlear ischemia. The DP amplitude decrease across all animals is 76.3 -I- 5.2 BL (n = 13). On clamp release, a brief DP increase occurs following CBF increase. Amplitude of this DP increase is negatively correlated with clamping duration. This peak is not detectable from DP curves in response to > 60 s clamps. The second peak of DP amplitude can be identified from these responses to > 30 s clamps. Amplitude of the second DP peak seems to have a negative relationship to clamping duration. The third increase of DP amplitude is observed from responses to 30 s clamps. Amplitude of the third DP increase peak appears to be clamping-duration related. Although obvious overshoot of DP amplitude occurs after the LA reperfusion, no overshoot is observed before clamp release. The DP pattern in response to cochlear reperfusion is different from the CBF response pattern. A full recovery of DP is observed for all clamps. Full recovery time shows a positive relationship to clamping duration. DP phase shows changes that are consistent with DP magnitude. The first 5 s of CBF responses to LA clamping of different durations from one animal are plotted in Fig. 3. Despite different total clamping times, CBF shows a very similar decline. CBF changes show a typical exponential decrease with a time constant of approximately 0.83 s, which provides a measurement of cochlear vascular mechanics. The similarity of CBF decreases in the figure demonstrate excellent reliability for LA clamping-induced ischemia in the gerbil cochlea. CBF response to the first 200 s of reperfusion is plotted against different clamping durations (Fig. 4). After a short clamping, such as 5, 10, 15 or 30 s, CBF increases immediately and reaches an overshoot maximum in less than 2.5 s during LA reperfusion. The CBF overshoot completely recovers in less than 1 min. With clamping
T. Ren et a l . / Hearing Research 92 (1995) 3 0 - 3 7
140 " > 120 100 o_,,
80
~"
60
~
x
g
EJ
(
Fig. 5. DP amplitude during the first 200 s of reperfusion plotmd against clamping duration. Time delay of recovered DP exhibits a positive relationship to clamping duration. DP response patterns also appears to be clamping duration-related. Rough surface at long clamping duration range ( > 30 s) suggests fluctuations of DP amplitude.
duration increase, reperfusion-caused CBF increase rate becomes slower. For longer duration clamping ( > 60 s), CBF overshoot reaches the maximum at approximately 150 s during reperfusion. Full recovery time for these CBF overshoots is longer than 4 min. A long-lasting CBF overshoot occurs at approximately 150 s reperfusion time following a long-duration clamping ( > 60 s). In summary, Fig. 4 shows that CBF changes during cochlear reperfusion are reperfusion time- and ischemia duration-related. In order to show ischemia and reperfusion effects on DP, DP amplitude response during the first 200 s of reperfusion time from one animal is presented against the clamping duration in Fig. 5. DP response patterns appear clamping duration-related. For short duration clamping ( < 30 s), DP fully recovers in less than 60 s reperfusion time with no oscillation. The rough surface shown in the figure for long clamping duration range ( > 30 s) suggests fluctuations of DP amplitude. Overshoots and decreased amplitudes after overshoots seem to be clamping-duration related. DP full recovery time is positively related to clamping duration. These data show that following different durations of ischemia, cochlear reperfusion results in complicated multiphase DP responses. Compared to data in Fig. 4, DP amplitude responses are quite different from CBF response during the reperfusion.
4. Discussion
4.1. Reversibility and repeatability of the current animal model Reversibility is a basic requirement for a cochlear ischemia animal model for study of ischemia/reperfusion effects and other cochlear physiological and pathophysio-
35
logical mechanisms. Before LDF was introduced for the study of cochlear microcirculation (Miller et al., 1983), it was difficult, if not impossible, to continuously monitor CBF. Without the real time dynamic CBF measurement provided by LDF, it was impossible to directly study reversibility of the cochlear ischemia model. Although LDF has been used in a few studies of cochlear ischemia, no data are available which clearly indicate complete reversibility. Variation in the anatomy of the feeding vessels of the cochlea plus methods used to manipulate cochlear supplying vessels probably were responsible for incomplete reversibility in these models. To demonstrate LDF application in the cochlea, Miller et al. (1983) and Short et al. (1985) found that local pressure on the contents of the internal auditory meatus reduced the cochlear LDF signal to 15-30% BL in a reversible pattern. This preparation causes local pressure on the auditory nerve, with potential damaging and as well as stimulating effects. Therefore, it is inadequate as an ischemia model. Using a microclamp to mechanically occlude the AICA, Randolf et al. (1990) found a variable CBF decrease during the clamping and partial recovery after the release of the clamping. Clamp-caused mechanical damage may have contributed to their experimental results, since Ren et al. (1993, 1994) showed a complete CBF recovery following a large overshoot ( > 160% BL) after AICA clamping. Although Seidman et al. (1991) had investigated superoxide dismutase and allopurinol effects on ischemia and reperfusion in the rat cochlea, there was a little information on the CBF dynamic response to the clamping of the AICA. Since photochemically induced thrombosis was introduced to develop a cochlear ischemic model (Umenura et al., 1990; Asai et al., 1993), a great effort has been made to improve reversibility in this model. Because of difficulties in controlling thrombosis formation and resolution, full reversibility has not yet been achieved. Moreover, direct cochlear sensory cell damage caused by photochemically induced free radicals make interpretation of experimental results difficult. Data presented in Figs. 2 and 4 clearly show a full recovery following dynamic responses to different duration LA clamping, indicating complete reversibility of the current model. Good repeatability is illustrated also in Fig. 3. However, proper tension on the microclamp tip is essential: great enough to completely clamp the LA but soft enough to avoid any vascular damage. Quantitative calibration of the microclamp was not feasible because of the varying effects of vascular tension, fluid surface tension, viscosity-caused friction, and the non-linearity of the clamp spring and vascular wall mechanics. Therefore a practical approach was utilized to achieve repeatable non-invasive clamping in each animal. Tension on the clamp was increased until the LA blood flow completely stopped, as observed under a stereomicroscope. Then, the distance advanced was marked and fixed by a screw on the distal end of the camera release cable. This approach improved the efficiency of the ischemia/re-
36
T. Ren et al. / Hearing Research 92 (1995) 30-37
perfusion model, achieving full reversibility and good repeatability. 4.2. Controllable and quantitative ischemia in the cochlea A basic requirement for an ideal quantitative cochlear ischemia model is the ability to control and monitor the level and duration of ischemia. To the best of our knowledge, no cochlear ischemia model which fully meets these criteria has been reported in the literature. The dynamic and relatively quantitative properties of L D F have made it the most commonly used method to measure CBF. Instant change on L A clamping or release (Figs. 2 and 3) demonstrates the good time resolution in the LDF-measured C B F data obtained from gerbil. The L D F time resolution and the speed o f the L A clamping and releasing made a clear o n / o f f marker on the C B F signal. Thus, a controllable and quantitative duration o f cochlear ischemia was achieved in the gerbil cochlea. However, the autoregulatory response of cochlear vessels made it difficult to control the level of ischemia in the cochlea (Ren et al., 1993, 1994). Development of a microclamp with the capability of partial L A clamping may be a way to control ischemia level in future experiments. 4.3. Continuous cochlear function measurement during ischemia / reperfusion There are several methods, such as the compound action potential, cochlear microphonic potential, endocochlear potential, auditory evoked brain-stem response, and the different otoacoustic emissions, for revealing ischemia-caused cochlear function changes. Since i s c h e m i a / r e p e r f u s i o n damage is a time-dependent phenomenon, cochlear function should be monitored continuously a n d dynamically. W h e n two continuous tones with appropriate parameters are introduced in the ear, cubic combination tone at frequency of 2 f l - f 2 ( f l < f 2 ) can be generated in the cochlea and measured in the external ear canal (Brown and Kemp, 1984). This combination tone is c o m m o n l y known as the cubic DP and has been widely used as an indication of the non-linearity of the cochlear active process and normal cochlear function. Because of its high sensitivity and adequate time resolution, cubic DP was used to assess cochlear function in this experiment. Regarding the ischemia effect on otoacoustic emissions, Schmiedt and A d a m s (1981) reported that DP in the gerbil can persist up to 2 h after death from anoxia. They found that the cubic difference tone ( 2 f i e f 2) decreased initially, then recovered to approximately the normal magnitude. Schmiedt (1986) later noted that the effect of anoxia on DP ( 2 f l - f 2) is dependent on the level of the stimuli. K e m p and Brown (1984) found gerbil cubic DP declines 5 - 1 0 min after death. Lonsbury-Martin et al. (1987) found that a stimulus level of 75 dB SPL produced DP that lasted longer after death than the 65 dB SPL evoked emissions.
Whitehead et al. (1992) also noted a similar characteristic in the rabbit ear. Rebillard and Lavigne-Rebillard (1992) observed that after hypoxia, cubic DP in the guinea pig often initially increased (4 dB) then decreased. Data presented in Figs. 2 and 5 clearly show constant DP decrease during ischemia and dynamic change upon reperfusion. Transient DP decrease during cochlear reperfusion may indicate reperfusion-induced changes in cochlear function, since this DP change cannot be interpreted simply as ischemia-caused changes. The difference between previously reported data and the current experimental results may be due to the different animal model and protocol variances, since anoxia must have caused complicated systematic changes and uncontrollable cochlear metabolic disturbances. Since the current controllable quantitative ischemia model has complete reversibility and good repeatability, it will be useful in studies o f ischemia/reperfusion-induced cochlear damage and other basic mechanisms of cochlear physiology or pathophysiology.
Acknowledgements Authors gratefully acknowledge thoughtful suggestions given by Dr. A l f Axelsson and editorial help by Carole Tice. This work was supported by NIH Grants DC 00105 and A G 08885.
References Asai, Y., Umemura, K., Kohno, Y., Uematsu, T. and Nakashima, M. (1993) An animal model for hearing disturbance due to inner ear ischemia: photochemically induced thrombotic occlusion of the rat anterior inferior cerebellar artery. Eur. Arch. Oto-rhino-laryngol.250, 292-296. Bernstein, J.M. and Silverstein, H. (1966) Anterior cerebellar and labyrinthine arteries: a study in the cat. Arch. Otolaryngol. 83, 422-435. Brown, A.M. and Kemp, D.T. (1984) Suppressibility of the 2 f l - f 2 stimulated acoustic emissions in gerbil and man. Hear. Res. 13, 29-37. Cole, R.R. and Jahrdoerfer, R.A. (1988) Sudden hearing loss: an update. Am. J. Otol. 9, 211-215. Hellstrom, L.I. and Schmiedt, R.A. (1990) Compound action potential input/output functions in young and quiet-aged gerbils. Hear. Res. 50, 163-174. House, W. (1975) Mrni~re's disease. Management and theory. Otolaryngol. Clin. N. Am. 8, 515-535. Hultcrantz, E. (1988) Clinical treatment of vascular inner ear diseases. Am. J. Otolaryngol. 8, 317-322. Kemp, D.T. and Brown, A.M. (1984) Ear canal acoustic and round window electrical correlates of 2 f l - f 2 distortion generated in the cochlea. Hear. Res. 13, 39-46. Kimura, R. (1986) Animal models of inner ear vascular disturbances. Am. J. Otolaryngol. 7, 130-139. Kimura, R. and Perlman, H.B. (1958a) Arterial obstruction of the labyrinth. 1. Cochlear changes. Ann. Otol. Rhinol. Laryngol. 67, 5-25.
T. Ren et a l . / Hearing Research 92 (1995) 3 0 - 3 7
Kimura, R. and Perlman, H.B. (1958b) Arterial obstruction of the labyrinth. 2. Vestibular changes. Ann. Otol. Rhinol. Laryngol. 67, 26-41. Lonsbury-Martin, B.L., Martin, G.K., Probst, R. and Coats, A.C. (1987) Acoustic distortion products in rabbit ear canal. I. Basic features and physiological vulnerability. Hear. Res. 28, 173-189. Miller, J.M., Hultcrantz, E., Short, S.O. and Nuttall, A.L. (1986) Pharmacological effects on cochlear blood flow measured with the laser Doppler technique. Scand. Audiol. Suppl. 26, 11-20. Miller, J.M., Marks, N.J. and Goodwin, P.C. (1983) Laser Doppler measurements of cochlear blood flow. Hear. Res. 11,385-394. Randolf, J.B., Hanpt, H. and Scheibe, F. (1990) Cochlear blood flow following temporary occlusion of the cerebellar arteries. Eur. Arch. Otorhinolaryngol. 247, 226-228. Rebillard, G. and Lavigne-Rebillard, M. (1992) Effect of reversible hypoxia on the compared time courses of endocochlear potential and 2 f l - f 2 distortion products. Hear. Res. 62, 142-148. Ren, T., Nuttall, A.L. and Miller, J.M. (1993) Contribution of the anterior inferior cerebellar artery to cochlear blood flow in guinea pig: a model-based analysis. Hear. Res. 71, 91-97. Ren, T., Avinash, G.B., Nuttall, A.L., Miller, J.M., Laurikainen, E.A. and Quirk, W.S. (1994) Dynamic response of cochlear blood flow to occlusion of anterior inferior cerebellar artery in guinea pigs. J. Appl. Physiol. 76, 212-217. Schmiedt, R.A. (1986) Effects of asphyxia on levels of ear canal emissions in gerbil. Abstr. Assoc. Res. Otolaryngol. 9, 112-113.
37
Schmiedt, R.A. and Adams, J.C. (1981) Stimulated acoustic emissions in the ear canal of the gerbil. Hear. Res. 5,295-305. Seidman, M.D., and Quirk, W.S. (1992) The anterior inferior cerebellar arterial network supplying the rat cochlea and its role in autoregulation of cochlear blood flow. Eur. Arch. Otorhinolaryngol. 249, 332335. Seidman, M.D., Quirk, W.S., Nuttall, A.L. and Schweitzer, V. (1991) The effects of superoxide dismutase and allopurinol on ischemia and reperfusion in the cochlea. Otolaryngol. Head. Neck. Surg. 105, 457-461. Short, S.O., Goodwin, P.C., Kaplam, J.N. and Miller, J.M. (1985) Measuring cochlea blood flow with laser Doppler spectroscopy. Otolaryngol. Head Neck Surg. 43, 786-793. Umenura, K., Kohno, Y., Uematsu, T. and Nakashima, M. (1990) A new model for photochemically induced thrombosis in the inner ear microcirculation and the use of hearing loss as measure for microcirculatory disorders. Eur. Arch. Otorhinolaryngol. 248, 105-108. Whitehead, M.L., Lonsbury-Martin, B.L. and Martin, G.K. (1992) Evidence for two discrete sources of 2 f l - f 2 distortion-product otoacoustic emission in rabbit. II. Differential physiological vulnerability. J. Acoust. Soc. Am. 92, 2662-2682. Yamasoba, T., Kikuchi, S., Higo, R., Ofchi, T. and Tokumaru, A. (1993) Sudden sensorineural hearing loss associated with slow blood flow of the verterbrobasilar system. Ann. Otol. Rhinol. Laryngol. 102, 873877.