An endothelin-1 induced model of optic nerve ischemia in the rabbit.

An Endothelin-1 Induced Model of Optic Nerve Ischemiain the Rabbit

Selim Orgiil* George A. doffi,* David J. Wilson,^ David R. Bacon*and E. Michael Van Buskirk*

Purpose. To evaluate blood flow reduction and topographic optic nerve changes after thelocal administration of endothelin-1 in vivo, delivered to the perineural region of the anterioroptic nerve in the rabbit.

Methods. Endothelin-1 (five rabbits) in a dosage of 0.1 //g/day or balanced salt solution (tworabbits) was delivered to the perineural region of the anterior optic nerve with osmoticallydriven minipumps. Optic nerve blood flow was determined by the colored microspherestechnique after 14 days of local endothelin-1 or balanced salt solution administration to themicrovasculature of the optic nerve. In addition, optic nerve blood flow was determined intwo rabbits that had no minipump implants. The morphologic changes induced by reductionof blood flow were assessed in five additional rabbits implanted with osmotically driven mini-pumps containing endothelin-1 (0.1 fj,g/day). These rabbits were observed for 8 weeks, andthe morphologic optic nerve changes were monitored with a confocal scanning laser ophthal-moscope.

Results. Independent of intraocular pressure, endothelin-1 induced a decrease in blood flowof approximately 38% in the experimental eye, compared to the decrease induced by balancedsalt solution or to the decrease in rabbits without minipumps (analysis of covariance, P =0.0092). Multivariate statistical analysis showed a significant change in topometric parameters(cup area, cup depth, rim volume) obtained with a confocal scanning laser ophthalmoscope,indicating an increase in optic nerve cupping and a decrease of the perineural rim volumein the experimental eyes (P = 0.017).

Conclusions. The current results suggest that morphologic optic nerve alterations can be in-duced experimentally in the rabbit model after ischemia produced by the local administrationof endothelin-1 to the perineural region of the anterior optic nerve. Invest Ophthalmol VisSci. 1996;37:1860-1869.

-M-icrocirculatory compromise of the anterior opticnerve has been invoked as a potential causal factor orcontributor to a variety of optic neuropathies.'"13 Anexperimental model to examine the optic nerve ef-fects of vascular insufficiency has not been available.Recently, however, an in vivo experimental model of-fering a titratable method to examine the effects of

From the *Devers Eye Institute and the R. S. Dow Neurological Science Institute,legacy Portland Hospitals, and the ̂ Department of Ophthalmology, Oregon HealthScience University, Portland, Oregon.Supported in part by the Schweizerische Stiftungfu'r mediunisch-biologischeStipendien and by National Institutes of Health grant EYO5231.Submitted for publication October 27, 1995; revised May I, 1996; accepted May 3,1996.Proprietaiy interest category: N.Reprint requests: George A. doffi, Ocular Microcirculation Unit, Devers EyeInstitute/Good Samaritan Hospital, 1040 N. W. 22nd Avenue, Suite. 200, PortlandOR 97210-3065.

vasoconstriction and vascular insufficiency of the ante-rior optic nerve was developed.1415 In this model,dose-dependent vasoconstriction of the optic nervemicrovasculature of the rabbit has been demonstratedwith local application of endothelin-1 delivered to theperineural region of the anterior optic nerve by osmot-ically driven minipumps.

Endothelins are among the most potent vasopres-sor substances yet discovered.16 Originally isolatedfrom cultured endothelial cells, endothelins exertpowerful and lasting constrictor effects on vascularsmooth muscle.17 These compounds are a family ofvasoactive polypeptides that exist as isoforms, endo-thelin-1 (ET-1), endothelin-2 (ET-2), and endothelin-3 (ET-3). Endothelin-2 is the most potent vasoconstric-tor, followed by ET-1 and ET-3.18 Only ET-1 is synthe-

1860Investigative Ophthalmology & Visual Science, August 1996, Vol. 37, No. 9Copyright © Association for Research in Vision and Ophthalmology

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933194/ on 04/14/2018

Optic Nerve Ischemia 1861

sized by endothelial cells. This group of compoundsis thought to contribute to the regulation of regionalblood flow throughout the body. Previous investiga-tions have shown the vasoconstrictive properties of theendothelins to be dose and concentration dependentin various vascular beds, including the anterior opticnerve microvasculature.15'17 These peptides also havebeen shown to produce, in vivo, localized vasoconstric-tion when injected directly into perivascular cerebraltissues, resulting in regional ischemic damage of thebrain's nervous tissues.19 The purpose of the currentstudy was to evaluate the blood flow reduction andthe alteration in optic nerve morphology after localapplication of ET-1 delivered to the perineural regionof the anterior optic nerve in the rabbit.

MATERIALS AND METHODS

Animals

Fourteen adult New Zealand white rabbits of eithersex, weighing 2.8 kg or more, were used. All experi-ments conformed to the ARVO Statement for the Useof Animals in Ophthalmic and Vision Research. In thefirst experiment, the blood flow effect of local ET-1administration to the microvasculature of the opticnerve was examined using the colored microspherestechnique20 in five experimental (ET-1 administra-tion) and four control rabbits (two rabbits adminis-tered balanced salt solution [BSS] and two rabbitswithout any substance administration). In the secondexperiment, the morphologic optic nerve changeswere monitored for 8 weeks using a confocal scanninglaser ophthalmoscope (Heidelberg Retina Tomo-graph [HRT]; Heidelberg Engineering, Heidelberg,Germany) in five additional rabbits.

Administration of Endothelin-1 to theMicrovasculature of the Optic Nerve

Endothelin-1 was delivered to the perineural regionof the anterior optic nerve by osmotically driven mini-pumps (Alzet minipumps; Alza Corporation, PaloAlto, CA) that deliver a test agent at a controlled andconstant flow rate (0.5 /zl/hour). The delivered dos-age is predetermined by the concentration of the drugwithin the minipump. For implantation of the mini-pumps, the rabbits were anesthetized with intravenoussodium pentobarbital through a marginal ear vein,and the intraocular pressure (IOP) was measured witha tonometer (Tono-Pen 2; Oculab, Glendale, CA) be-fore implantation of the minipump. The tonometerwas calibrated immediately before IOP measurements,and three readings per eye were taken. A mean IOPwas calculated from the three readings per eye.

Minipumps are oval capsules with a polyethylenedelivery tube extending from one end. They were im-

Endothelin-1 Minipump Deliveryto the Rabbit Optic Nerve

Minipump

SubTenonsDelivery 1\ibe

FIGURE l. Minipumps were implanted in a surgically createdsubcutaneous space, superior and nasal to the right eye. Apolyethylene delivery tube was placed, in a subcutaneousmanner, from the minipump through the upper eyelid andits palpebral conjunctiva. A sub-Tenon's capsular channelwas dissected posteriorly in the superiotemporal quadrantto allow direct observation of the optic nerve region. A su-perionasal bulbar conjunctival incision was made, and thedelivery tube was directed under the bulbar conjunctiva andunder the superior rectus muscle, into the superiotemporalsub-Tenon's channel. The distal end of the delivery tubewas fixed in place using a scleral fixation suture, so that theend of the tube was adjacent to the optic nerve and itsvascular supply.

planted in a surgically created subcutaneous space,superior and nasal to the right eye. A polyethylenedelivery tube was directed from the minipumpthrough the upper eyelid into a surgically created su-periotemporal sub-Tenon's channel and was fixed inplace using a scleral fixation suture adjacent to theoptic nerve and its vascular supply (Fig. 1).M Balancedsalt solution (BSS, two rabbits) or a daily dosage ofET-1 (Peptides International, Louisville, KY) 0.1 fig/day (five rabbits) was delivered to the perineural re-gion for 2 to 8 weeks.

Surgical Setup for Blood Flow Determination

Blood flow was determined using the colored micro-spheres technique after 14 days of local drug adminis-tration (five rabbits, ET-1; two rabbits, BSS) to themicrovasculature of the optic nerve. In addition, opticnerve blood flow was determined in two rabbits inwhich no minipumps were implanted. The rabbitswere anesthetized with intravenous sodium pentobar-bital, and IOP was measured. The rabbits were trache-otomized and mechanically ventilated with room airby means of a Harvard small-animal respirator. Bloodsamples for blood gas analysis were obtained throughan arterial catheter in the left forelimb. Mean systemicblood pressure was monitored continually through a


1862 Investigative Ophthalmology & Visual Science, August 1996, Vol. 37, No. 9

saline-filled arterial catheter in the right forelimb thatwas connected to the mercury column of a sphygmo-manometer. Rectal body temperature was monitoredcontinually and kept stable by placing the animal ona thermostatically controlled heating pad.

Injection of Microspheres for Blood FlowDetermination

Black polystyrene spheres 10.2 ± 0.23 /im (mean ±SD) in diameter (E-Z TRAC Ultraspheres; InteractiveMedical Technologies, South Barrington, IL) weresuspended in saline containing 0.05% Tween 80. Thesolution containing the microspheres was exposed tohigh-frequency sound waves (sonicated) for 30 beforeinjection to keep the microspheres from aggregating.

After a sternal thoracotomy, the left atrium wascannulated with polyethylene tubing. After the arterialblood pressure had stabilized, 100 million micro-spheres were injected in a volume of 10 ml within 25seconds through the atrial catheter. This quantity ofmicrospheres has proven to yield a good reproducibil-ity for optic nerve blood flow measurements, andthere is no indication that, with such large numbersof injected microspheres, the estimates of regionalblood flow decrease.2021 From the start of the injec-tion, a 1-minute reference blood sample was collectedfrom the catheter in the left forelimb artery. No sig-nificant changes in blood pressure during the injec-tion or collection period were observed. At the endof the collection period, each animal was killed bycross-clamping the ascending aorta. This producedimmediate and rapid decrease in systemic blood pres-sure, demonstrating a suitable sensitivity of the bloodpressure measurement method. Both eyes were enu-cleated immediately.

Estimation of Blood Flow

The anterior optic nerves (approximately 1.5 mm inlength) were dissected carefully from the eye under adissecting microscope. Tissue included the papilla anda retrolaminar segment of approximately 0.5 mm. Spe-cial care was given to avoid contamination of the opticnerve tissue sample with microspheres from the cho-roidal vascular bed. All solutions used for further pro-cessing of the blood or the tissue samples were ob-tained from the manufacturer of the colored micro-spheres (E-Z Trac Ultraspheres; Interactive MedicalTechnologies). Tissue samples were weighed and dis-solved with Tissue/Blood Digest Reagent I (sodiumhydroxide) in a boiling water bath. Samples were neu-tralized with Tissue/Blood Digest Reagent II (sodiumphosphate) and centrifuged. The supernatant waspoured off, and the samples were resuspended in Mi-crosphere Counting Reagent (Tween 80, deoxycholicacid). The entire suspension was examined in a

Fuchs-Rosen thai hemocytometer, and all the micro-spheres were counted.

Reference blood samples were diluted with BloodHemolysis Reagent (ethanol) and centrifuged. Thesupernatant was poured off and the sediment was dis-solved with Tissue/Blood Digest Reagent I in a boilingwater bath. Digests were neutralized with Tissue/Blood Digest Reagent II and centrifuged. The super-natant was poured off, and the sediments were resus-pended in Microsphere Counting Reagent. Micro-spheres in an aliquot of this suspension were countedin a Fuchs-Rosenthal hemocytometer, and the totalnumber of microspheres in the reference blood sam-ple was calculated. Subsequently, the optic nerveblood flow could be calculated.20 Changes of bloodflow were analyzed by means of an analysis of covari-ance in a 3 (between groups: ET-1, BSS, no pump) X2 (within subject: optic nerve blood flow minipumpeye, optic nerve blood flow control eye) design withIOP readings included as changing covariates.

Morphologic Monitoring During Optic NerveIschemia

To assess the morphologic effects of ischemia, mini-pumps filled with ET-1 (0.1 /ig/day) were implantedin five rabbits. These rabbits were observed for 8weeks. Because the capacity of the minipumps in thisstudy allowed for only 2-week use, a new minipumpwas implanted every 2 weeks. Only the reservoir wasreplaced, whereas the delivery tube was left undis-turbed. Intraocular pressure was measured, as de-scribed above, before minipump implantation, after 4weeks of local ET-1 administration, and at the endof the 8-week observation period. Morphologic opticnerve changes were monitored with the HRT. Topo-graphic images were obtained before minipump im-plantation, after 4 weeks of local ET-1 administration,and at the end of the 8-week observation period.

The HRT is a confocal laser scanning ophthalmo-scope designed for three-dimensional measurementsof the retinal surface topography. Sequential topo-graphic images can be analyzed with respect to thesame region of interest, and changes in the topogra-phy can be quantified. The Heidelberg Software, Ver-sion 1.11 (Heidelberg Engineering, Heidelberg, Ger-many) was used for the current analysis. In prelimi-nary experiments, previous studies with photographsor ophthalmoscopy have been unable to demonstratemorphologic optic nerve changes convincingly.Therefore, a quantitative method to analyze the opticnerve topography was needed. The HRT is a quantita-tive method to analyze clinically the topography of theoptic nerve. It seemed reasonable to use this methodin the current study because previous experiments us-ing the HRT have shown comparable reproducibilityof the topometric data obtained in rabbits compared



to humans.22 This is important because good repro-ducibility is essential for a useful statistical power indetermining small changes. For scaling the topo-graphic images, the HRT software uses Gulstrand'smodel eye and the value of the corneal curvature todetermine the focal length of the eye. Gulstrand'smodel cannot be applied to the rabbit eye. Its averageaxial length is 17 mm, and its corneal curvature is 7.5mm.23 By using a corneal curvature of 7.5 mm, thecalculated focal length would not be correct. There-fore, an artificial corneal radius of 4.85 mm, based onan assumed focal length of 16 mm, was used.22'2'1

Rabbits were anesthetized with intravenous so-dium pentobarbital through a marginal ear vein dur-ing optic nerve imaging, and the pupils were dilatedwith tropicamide 0.5%. To enhance imaging quality,the cornea was moistened with artificial tears betweenimage acquisition. At each session, a series of 10 topo-graphic images was obtained from each eye by alter-nating both eyes during acquisition of the 20 images.

The software of the HRT can be used to generatea mean topographic image from a set of individualtopographic images by determining the averageheight measurement at each image location. A meantopographic image was generated with the 10 imagesper eye, and the region of interest was defined by acontour line on this mean topographic image. Thiscontour line was used for each topographic image ofthe sample to compute the topographic parameters("export-import contour line" function). This con-tour line also was "imported" on the subsequent(after 4 and 8 weeks) samples of mean topographicimages and consequently could be used to computethe topographic parameters in later samples with re-spect to the same region of interest. The HRT uses a"curved surface"23 as an upper delimitation of theexamined structure. The effective area (EA) estimatesthe optic nerve cup area, the volume below surface(VBS) estimates the volume of the optic nerve cup,the volume above surface (VAS) estimates the opticnerve rim volume, the effective mean depth (EMD)estimates the mean depth of the optic nerve cup area,the maximum depth (MaxD) estimates the meandepth of the 5% picture elements with the highestdepth values within the optic nerve cup, the thirdmoment (TM) estimates the overall shape of the opticnerve head, and the quotient EA/Area (Area estimatesthe total optic nerve area) estimates the cup:disk ratio.The change over time of each parameter was analyzedin a 5 (between groups: five rabbits) X 2 (betweengroups: minipump eye, control eye) X 3 (betweengroups: baseline, 4 weeks, 8 weeks) analysis of covari-ance design, with the mean height of contour (MHC;mean height value along the contour line) includedas a covariate to correct for imperfect alignment ofthe contour line.24 Because seven parameters were

compared, the probability that at least one of this com-parisons would be significant by chance was 30%. Tokeep this probability at 5%, only a P value lower than0.0073 was considered significant for the evaluationof the univariate comparisons. In addition to theseunivariate comparisons, a multivariate approach wasused. To preserve a reasonable statistical power to thestudy, three arbitrarily chosen topographic parame-ters—EA, VAS, and EMD—were analyzed. These vari-ables provided little redundant information. Thechange over time of VAS, EA, and EMD in these fiverabbits was analyzed in a 5 (between groups: five rab-bits) X 2 (between groups: minipump eye, controleye) X 3 (between groups: baseline, 4 weeks, 8 weeks)multivariate (VAS, EA, EMD) analysis of covariancedesign, with the mean height of contour (MHC; meanheight value along the contour line) included as acovariate to correct for imperfect alignment of thecontour line.24 The interocular difference in variabil-ity of IOP over time was analyzed in a multivariateapproach of a 2 (within subject: minipump eyes, con-trol eyes) X 3 (within subject: baseline IOP, IOP after4 weeks, IOP after 8 weeks) analysis of variance design.

After 8 weeks, the rabbits were killed by anestheti-zation with intravenous sodium pentobarbital througha marginal ear vein and mechanical ventilation. Theabdominal aorta was exposed and catheterized. Thesuperior circulation of the rabbits was flushed initiallywith a heparin solution; subsequently, buffered 4%paraformaldehyde was perfused into the ascendingbranches of the aorta. Coincident with the start ofperfusion with the heparin solution, the right atriumwas incised to permit efflux of venous blood and ex-sanguination. After perfusion with the fixative, theeyes were enucleated, opened coronally, and im-mersed in buffered 4% paraformaldehyde at 4°C tocomplete the fixation. Anterior optic nerves were care-fully dissected and preserved in buffered 4% para-formaldehyde.

HistologySubsequent to the detection of statistically significantmorphologic changes in the experimental eyes, asmeasured with the HRT, the optic nerves were as-sessed for potential histologic correlates for thesefindings. Anterior optic nerves were embedded inplastic (LR White resin), sectioned, and stained withtoluidine blue. In the central region of the optic nervecup, within the medullary rays, 1- to 2-//m thick sec-tions were screened for abnormalities by light micros-copy in a masked fashion.

RESULTS

Optic Nerve Blood Flow

The average (mean ± SD) weight of the examinedtissue among the 18 eyes of the 9 rabbits used to assess


1864 Investigative Ophthalmology 8c Visual Science, August 1996, Vol. 37, No. 9

TABLE l. Optic Nerve Blood Flow in Rabbits Implanted WithEndothelin-1 or BSS-Filled Osmotic Minipumps for 2 Weeks

Rabbit Group(number of rabbits)

ET-1-filled minipumps (5)BSS-filled minipumps (2)

Values are mean ± SD.BSS = balanced salt solution.

Minipump Eye(ftl/mg per minute)

0.12 ± 0.010.18 ± 0.04

Contralateral Eye(nl/mg per minute)

0.19 ± 0.030.17 ± 0.05

P Value

0.000820.63

optic nerve blood flow was 2.9 ± 0.6 mg. The average(mean ± SD) number of microspheres counted in thetissue was 128 ± 46. Blood flow data among the rabbitsimplanted with minipumps are outlined in Table 1.Optic nerve blood flow (mean ± SD) was 0.17 ± 0.01and 0.17 ± 0.04 in the right and the left eyes, respec-tively, among the rabbits with no minipump implants.The IOP readings at baseline (before minipump im-plants) and the IOP readings immediately beforeblood flow determination among the rabbits subjectedto ET-1 and the rabbits implanted with BSS-filled mini-pumps are shown in Table 2. Changes in optic nerveblood flow in the right eyes (minipump eyes for therabbits implanted with minipumps) compared to theleft eyes was significantly different among the threegroups (ET-1, BSS, no minipumps), independent ofthe changes in IOP (Fig. 2; P = 0.0092). A statisticalcontrast analysis, which allows testing of the statisticalsignificance of specific differences in particular partsof a complex design, disclosed a significantly loweroptic nerve blood flow in the ET-1 minipump eyescompared to the contralateral eye (mean ± SD: 37.8± 7.2%; P = 0.00082), whereas the rabbits implantedwith BSS-filled minipumps and those without mini-pumps had a comparable blood flow in both opticnerves (P= 0.63 and P = 0.97, respectively). Statisticalpower for a difference similar to that observed amongthe rabbits subjected to ET-1 to be statistically signifi-cant at the level of 0.05 was more than 99% and 88%among the rabbits implanted with BSS-filled mini-pumps and those without minipumps, respectively.

Optic Nerve Morphology

From a total of 300 HRT topographic images obtainedfor the five rabbits, 269 images were included in this

analysis. The prerequisite for a topographic image tobe included was the ability of the HRT software algo-rithm to align the original contour line on it. Amongthe univariate comparisons of the seven topometricparameters, only EMD showed a statistically significantdifference in the interocular difference in change overtime (P = 0.0037). The interocular difference inchange of EMD from baseline to the fourth week wasstatistically significant (contrast analysis, P< 0.00001),whereas no difference was seen between the fourthand the eight weeks. For the multivariate approach,the change of the three topometric variables—EA,EMD, and VAS—was statistically significantly differentbetween the eyes subjected to ET-1 and the contralat-eral eyes (multivariate analysis of variance, P = 0.017).Adjusted average values (corrected for the variationin MHC) of VAS, EA, and EMD are summarized inTable 3. Statistical contrast analysis disclosed a signifi-cant change of these three topographic parametersfor the ET-1-administered eyes ( P < 0.00001), whereasno significant change occurred in the contralateraleyes (P = 0.13). A canonical analysis extracted tworoots, with the first root showing an explanatory powerof 92.74% of the variation of the data. A canonicalanalysis correlates the weighted sums of two sets ofvariables in such a way that this correlation is max-imized. More than one weighted sum may be neces-sary to describe the complex structure of a set of vari-ables. Such weighted sums can be thought of asdescribing some underlying "latent" variables (ca-nonical roots). The standardized coefficients (canoni-cal weights) of the present discriminant function (firstcanonical root) were 0.21, -0 .33, and -0.72 for VAS,EA, and EMD respectively. The more negative the re-

TABLE2. Intraocular Pressure in Rabbits Implanted With Endothelin-1 (ET-1) or BSS-FilledOsmotic Minipumps for 2 Weeks

Rabbit Group(number of rabbits)

ET-1-filled minipumps (5)BSS-filled minipumps (2)

MinipumpEye (mm Hg)

10.1 ± 0.611.8 ± 0.7

Baseline

ContralateralEye (mm Hg)

9.9 ± 1.010.7 ± 0.5

Before Blood Flow Assessment

MinipumpEye (mm Hg)

13.2 ± 1.012.0 ± 0.5


12.9 ± 0.911.7 ± 0.5

Values are mean ± SD.



n?4

0.22

0.20

0.18

0.16

0.14

0 12

0.10

0.08

0.06

0.04

0.02

n nn

ET-1

|

BSS NCI MINIPUMP

CONTROL MINIPUMP CONTROL MINIPUMP RIGHT LEFT

FIGURE 2. Optic nerve blood flow (mean ± SD) among the rabbits (five) implanted withminipumps containing endothelin-1 (ET-1), those (two) implanted with minipumps con-taining balanced salt solution (BSS), and those (two) without minipump implants. Bloodflow was determined 2 weeks after the minipump implant among the rabbits implanted withminipumps. The interocular optic nerve blood flow difference varied significantly betweenthe three groups, independently of the IOP (P = 0.0092). Blood flow in optic nervessubjected to ET-1 was significandy lower compared to the contralateral side (P = 0.00082).The interocular difference was not significant among the rabbits implanted with BSS-filledminipumps and those without minipumps (P = 0.63 and P = 0.97 respectively).

suit of this discriminant function becomes, the moreEA and EMD increase and the more VAS decreases(Fig. 3), which is consistent with an increase in cupsize and a decrease in optic nerve rim volume. There-fore, approximately 93% of the variation seen in thecurrent data seems to be accounted for by a morpho-logic optic nerve alteration in the eyes subjected to

TABLE 3. Mean Values of TopometricParameters After Adjustment for theVariation in the Alignmentof the Contour Line*

Eye

MinipumpMinipumpMinipumpControlControlControl

Time

Baseline4 weeks8 weeksBaseline4 weeks8 weeks

VAS(mm3)

0.0720.0580.0630.0670.0740.076

EA(mm2)

3.563.633.613.823.763.75

EMD(mm)

0.340.380.390.360.370.37

VAS = volume above surface (optic nerve dm volume); EA =effective area (optic nerve cup area); EMD = effective meandepth (optic nerve cup mean depth).* The change of the three topometric variables was statisticallysignificantly different between the eyes subjected to endothelin-1and the contralateral eyes (multivariate analysis of variance; P =0.017).

ET-1. However, contrast analysis showed that the mor-phologic optic nerve changes in the minipump eyesbetween the fourth and the eighth weeks were statisti-cally not significant (P = 0.10).

The IOP values at baseline, after 4 and 8 weeks,are shown in Table 4. The change over time was com-parable between the eyes subjected to ET-1 and thecontralateral eyes (P = 0.88). Therefore, the interocu-lar differences in IOP cannot account for the morpho-logic changes of the optic nerve in the eyes subjectedto ET-1.

HISTOLOGY

Definite histologic abnormalities were seen by lightmicroscopy in the optic nerve of the eyes subjected toET-1 among two of the rabbits after 8 weeks of localET-1 administration. These abnormalities included aloss of myelin and a gliosis of the prelaminar portionof the optic nerve (Fig. 4). The authors disagreed onthe classification (minipump eye versus control eye)of the optic nerves in a third rabbit, and no definitehistologic differences between the eyes subjected toET-1 and the contralateral eyes were seen, whenscreened in a masked fashion, among the remainingtwo rabbits. These differences between the rabbitsmost probably reflect the variable degree of optic



-10.5

O -11.5

i= -15.5

-12.5

-13.5

-14.5

-16.5

-17.5

-18.5

\

\\

BASELINE FOUR WEEKS EIGHT WEEKS

MINIPUMP EYESBASELINE FOUR WEEKS EIGHT WEEKS

CONTRALATERAL EYES

FIGURE 3. The change in optic nerve morphology was significantly different between theminipump eyes and the contralateral eyes, indicating an increase in optic nerve cuppingand a decrease in optic nerve rim volume (P = 0.017). Discriminant function (mean ± SD)pertaining to optic nerve cup area, optic nerve cup depth, and optic nerve rim volumemeasured with the Heidelberg Retina Tomograph before minipump implants (baseline), 4weeks after the minipump implant, and 8 weeks after the minipump implant. Statisticalcontrast analysis disclosed a significant decrease for the right eyes (P < 0.00001), whereasno significant change occurred in the left eyes (P = 0.13).

nerve injury after a relatively brief period of ischemiaamong the rabbits, as already suggested by the smallmorphologic changes observed with the HRT.

DISCUSSION

Optic nerve blood flow was assessed in an in vivo ani-mal model after local administration of endothelin-1.In addition, the effect of optic nerve ischemia wasmonitored by means of a confocal scanning laser oph-thalmoscope. Administration of endothelin-1 to theanterior optic nerve region induced a significant de-crease in local blood flow of approximately 38%, com-pared to the contralateral eye. Multivariate analysisdisclosed a small, but statistically significant, change in

TABLE 4. Intraocular Pressure in RabbitsImplanted With Endothelin-1 FilledOsmotic Minipumps for 8 Weeks*

MinipumpEye (mm Hg)


BaselineAfter 4 weeksAfter 8 weeks

9.7 ± 0.511.2 ± 1.711.2 ± 1.4

9.8 ± 1.010.9 ± 0.811.3 ± 1.7

Values are mean ± SD.* The change over time was comparable between the eyessubjected to endothelin-1 and the contralateral eyes (P = 0.88).

optic nerve morphology, as measured with a confocalscanning laser ophthalmoscope, after 8 weeks of localadministration of endothelin-1, compared to the con-trol eyes. These changes were consistent with an opticnerve cupping and a decrease in optic nerve rim vol-ume. Histologic analysis showed loss of myelin andgliosis of the prelaminar portion of the optic nervein at least two optic nerves subjected to endothelin-1during 8 weeks. Blood flow and morphologic changeswere independent of changes in intraocular pressure.

Endothelin-1 was chosen in the current study toinduce an animal model of optic nerve ischemia, notonly because it is the most potent vasoconstrictorknown but because it may play a role in some regionalischemic processes throughout the body. In healthyyoung humans, the circulating levels of endothelinare low.26 In pathologic conditions such as an ischemiccerebrovascular insult, the plasma level of endothelin-1 has been reported to be elevated.27 Experimentally,endothelin-1 has been shown to produce, in vivo, lo-calized cerebral vasoconstriction, resulting in regionalischemic damage of the brain nervous tissues.19 Endo-thelin-1 is a strong vasoconstrictive agent of the ocularcirculation as well, and the vasomodulation potencyof endothelial mediators increases with decreasing di-ameters of the blood vessel.28"31 This compound hasbeen demonstrated to cause, in vivo, a dose-depen-dent vasoconstriction of the anterior optic nerve mi-



B

FIGURE 4. Light-microscopic view of two pairs of optic nerves (toluidine blue stain). Opticnerves subjected to endothelin-1 during 8 weeks (A,C) showed a loss of myelin (dark areas)and a gliosis of the prelaminar portion of the optic nerve compared to the contralaterateyes (B,D). Arrows point to myelinated nerve fibers in the control eyes (B,D).

crovasculature in the rabbit.H ir> Recently, Moriya etal32 found a statistically significant increase in the lev-els of plasma endothelin-1 in patients with low tensionglaucoma compared to normal controls. Therefore,the role of endothelin-1 in optic nerve pathology de-serves further investigation.

Earlier investigations have demonstrated that re-peated exposure of the ophthalmic artery to endo-thelin-1 in vitro shows a marked tachyphylaxis,2"' aneffect which most probably caused by a downregula-tion of endothelin receptors/3 When cultured smoothmuscle cells are incubated with endothelin-1, amarked decrease in the endothelin-1 binding sites canbe observed, probably because endothelin-receptorsare internalized in the cells/™ In contrast, the endo-thelin-induced pressor response seems not to undergotachyphylaxis in normotensive animals.3334 Possibly,the turnover rate of the endothelin-receptors respon-sible for the pressor response is relatively high in vivo,and these receptors are synthesized prompdy after in-ternalization. This might be the reason no tachyphy-laxis of the vasoconstrictor response was observed after2 weeks of local administration of ET-1 in the currentstudy.

The intravascular microsphere injection techniquewas used for blood flow measurements in the currentstudy. Microspheres allow blood flow measurements insmall, inaccessible tissues. Application of the intravascu-lar microsphere injection technique, however, may notseem practical for studies on the eye. Because of therelatively small size of the optic nerve, one could expectthat too few spheres would localize to the optic nervevascular bed to permit flow determinations. However,with larger injected quantities of 8 to 10 fjbm spheres,optic nerve blood flow can be determined with reason-able precision.20 With increasing numbers of injectedmicrospheres, the number of those near the vessel wallincreases. These latter spheres may have a greater ten-dency to become entrapped in branching vessels, whichmight explain the low variability observed in previousstudies with this technique. Furthermore, colored micro-spheres have a density of 1.05 g/ml, which is closer tothat of red cells (1.098 g/ml) than that of radioactivemicrospheres (1.4 g/ml). This may facilitate the mixingof the spheres in the blood stream, thereby enhancingthe accuracy of the estimates of regional blood flow withcolored microspheres compared to radioactive micro-spheres.



Morphologic studies on peripheral nerves haveshown that, after local ischemia at an extent compara-ble to that of the current study, nerve fiber injurypredominantly took the form of demyelination.35'36

With more severe ischemia, however, Wallerian degen-eration also could be seen.35'36 Definite histopatho-logic changes were observed in the current study, butonly in 2 of 5 rabbits. These differences reflect thevariable degree of optic nerve injury after a relativelybrief period of ischemia among rabbits. This is consis-tent with the overall small morphologic changes ob-served with HRT. It seems reasonable to assume thatthe changes seen in light microscopy were not in-duced by fixation because of the observed asymmetry.Furthermore, the observed demyelination is inagreement with results obtained in peripheralnerves.35'36 Because of the fixation with paraformalde-hyde, axon counts by electron microscopy were notpossible in this study. Our results demonstrate thatdefinite morphologic changes can be observed in ananimal model of optic nerve ischemia. Because of theanatomic differences between the rabbit and the hu-man optic nerves, the implications of the observedhistopathologic alterations are uncertain. In contrastto humans, most of the anterior optic nerve in therabbit is myelinated. It appears, therefore, that futurehistologic investigation of the potential role of isch-emia in optic neuropathy should be performed in pri-mates and should include observations for longer timeintervals. The use of any other nonprimate than therabbit would not have improved the significance ofthe current results for humans. It seemed reasonableto use the rabbit for this study because its optic nervearterial supply is similar to that in humans. Further-more, definite vascular effects of endothelin havebeen demonstrated in the anterior optic nerve vascula-ture of rabbits, and investigational methods, such asthe colored microspheres technique and HRT im-aging, already have been validated in this animal.15'20'22

Morphologic changes observed with the HRT inthe current study suggest optic nerve cupping re-sulting from optic nerve ischemia. Such morphologicchanges are expected to occur in glaucomatous opticneuropathy. However, anatomic differences betweenthe optic nerve in humans and rabbits preclude adefinite comparison of the current results with humanglaucomatous optic neuropathy. Anterior ischemicoptic neuropathy probably would cause similar mor-phologic changes in a rabbit optic nerve. The poten-tial role of vascular insufficiency in glaucoma has beendebated for more than a century and remains contro-versial.7 It has been difficult to settle these controver-sies with current experimental techniques. Indeed,previously there was no experimental model thatcould be used to examine the optic nerve effects ofvascular insufficiency. Experimental optic nerve atro-

phy, similar to that seen in patients with glaucoma,can be induced by increased intraocular pressure.37'38

However, increased intraocular pressure cannot ac-count for the changes observed in the current study.Because these experiments have been performed inrabbits, their significance for human optic neuropa-thies must be considered cautiously, and the presentmodel of optic nerve ischemia must be carried into aprimate model to allow more reliable comparisonswith conditions observed in the human eye.

Key Words

blood flow, demyelination, glaucoma posterior segment, op-tic nerve, scanning laser

References

1. Hayreh SS. Anterior ischemic optic neuropathy. ArchNeurol. 1981;38:675-678.

2. Hayreh SS. Inter-individual variation in blood supplyof the optic nerve head: Its importance in variousischemic disorders of the nerve head, and glaucoma,low-tension glaucoma and allied disorders. Doc Oph-thalmol. 1985; 59:217-246.

3. Phelps CD, Corbett JJ. Migraine and low-tension glau-coma: A case-control study. Invest Ophthalmol Vis Sri.1985;26:1105-1108.

4. Drance SM, Douglas GR, Wijsman K, Schulzer M, Brit-ton RJ. Response of blood flow to warm and coldin normal and low-tension glaucoma patients. Am JOphthalmol. 1988; 105:35-39.

5. Olver JM, Spalton DJ, McCartney ACE. Microvascularstudy of the retrolaminar optic nerve in man: Thepossible significance in anterior ischaemic optic neu-ropathy. Eye. 1990; 4:7-24.

6. Gasser P, Flammer J. Blood-cell velocity in the nailfoldcapillaries of patients with normal-tension and high-tension glaucoma. Am J Ophthalmol. 1991; 111:585-588.

7. Van Buskirk EM, Cioffi GA. Glaucomatous optic neu-ropathy. Am J Ophthalmol. 1992; 113:447-452.

8. Tielsch JM, Katz J, Sommer A, Quigley HA, Javitt JC.Hypertension, perfusion pressure, and primary open-angle glaucoma. Arch Ophthalmol. 1995; 113:216-221.

9. Hayreh SS, Zimmerman MB, Podhajsky P, AlwardWLM. Nocturnal arterial hypotension and its role inoptic nerve head and ocular ischemic disorders. Am]Ophthalmol. 1994; 117:603-624.

10. Dielemans I, Vingerling JR, Algra D, Hofman A,Grobbee DE, dejong PTVM. Primary open-angle glau-coma, intraocular pressure, and systemic blood pres-sure in the general elderly population: The RotterdamStudy. Ophthalmology. 1995; 102:54-60.

11. Graham SL, Drance SM, Wijsman K, Douglas GR, Mi-kelberg FS. Ambulatory blood pressure monitoring inglaucoma: The nocturnal dip. Ophthalmology. 1995;102:61-69.

12. Orgul S, Flammer J. Perilimbal aneurysms of conjunc-tival vessels in glaucoma patients. Ger J Ophthalmol.1995; 4:94-96.



13. Stroman GA, Stewart WC, Golnik KC, Cure JK,Olinger RE. Magnetic resonance imaging in patientswith low-tension glaucoma. Arch Ophthalmol 1995;113:168-172.

14. Cioffi GA, Van Buskirk EM. Microvasculature ofthe anterior optic nerve. Surv Ophthalmol. 1994;38(suppl):S107-Sll7.

15. Cioffi GA, Orgiil S, Onda E, Bacon DR, Van BuskirkEM. An in vivo model of chronic optic nerve ischemia:The dose-dependent effects of endothelin-1 on theoptic nerve microvasculature. Curr Eye Res. 1995;14:1147-1153.

16. Vane JR, Anggard EE, Botting RM. Regulatory func-tions of the vascular endothelium. N Engl J Med.1990; 323:27-36.

17. Yanagisawa M, Kurihara H, Kimura S, et al. A novelpotent vasoconstrictor peptide produced by vascularendothelial cells. Nature. 1988; 332:411-415.

18. Inoue A, Yanagisawa M, Kimura S, et al. The humanendothelin family: Three structurally and pharmaco-logically distinct isopeptides predicted by three sepa-rate genes. ProcNatl Acad Sci USA. 1989; 86:2863-2867.

19. Sharkey J, Ritchie IM, Kelly PAT. Perivascular mi-croapplication of endothelin-1: A new model of focalcerebral ischemia in the rat. / Cardiovasc Pharmacol.1993; 13:865-871.

20. Orgiil S, Cioffi GA, Bacon DR, Van Buskirk EM. Mea-surement of optic nerve blood flow with nonradioac-tive colored microspheres in rabbits. Microvasc Res.1996;51:in press.

21. Geijer C, Bill A. Effects of raised intraocular pressureon retinal, prelaminar, and retrolaminar optic nerveblood flow in monkeys. Invest Ophthalmol Vis Sci.1979; 18:1030-1042.

22. Orgiil S, Cioffi GA, Bacon DR, Van Buskirk EM. Re-producibility of topometric data with a scanning laserophthalmoscope in rabbits. Jpn J Ophthalmol. 1995;39:438-442.

23. Prince JH. Cornea, trabecular region, and sclera. In:Prince JH, ed. The Rabbit in Eye Research. Springfield:Charles C Thomas; 1964; 86-139.

24. Orgiil S, Cioffi GA, Bacon DR, Van Buskirk EM.Sources of variability of topometric data with a scan-ning laser ophthalmoscope. Arch Ophthalmol. 1996;114:161-164.

25. Rohrschneider K, Burk ROW, Volker HE. Reproduc-ibility of topometric data acquisition in normal andglaucomatous optic nerve heads with the laser tomo-graphic scanner. Graefe's Arch Clin Exp Ophthalmol.1993; 231:457-464.

26. Hartter E, Woloszxzuk W. Radioimmunoassay of endo-thelin. Lancet. 1989; 1:909. Letter.

27. Ziv I, Fleminger G, Djaldetti R, Achiron A, Meland E,Sokolovsky M. Increased plasma endothelin-1 in acuteischemic stroke. Stroke. 1992;23:1014-1016.

28. Ke Y, Tschudi M, Flammer J, Luscher TF. Endothe-lium-dependent regulation of vascular tone of the por-cine ophthalmic artery. Invest Ophthalmol Vis Sci.1991;32:1792-1798.

29. Haefliger IO, Flammer J, Luscher TF. Nitric oxideand Endothelin-1 are important regulators of humanophthalmic artery. Invest Ophthalmol Vis Sci. 1992;33:2340-2343.

30. Haefliger IO, Flammer J, Luscher TF. Heterogeneityof endothelium-dependent regulation in ophthalmicand ciliary arteries. Invest Ophthalmol Vis Sci. 1993;34:1722-1730.

31. Meyer P, Flammer J, Luscher TF. Endothelium-depen-dent regulation of the ophthalmic microcirculationin the perfused porcine eye: Role of nitric oxide andendothelins. Invest Ophthalmol Vis Sci. 1993; 34:3614-3621.

32. Moriya S, Sugiyama Y, Shimizu K, Hamada J, TokuokaS, Azuma I. Low tension glaucoma and endothelin(ET-1). Folia Ophthalmol Jpn. 1992; 43:554-559.

33. Hirata Y, Yoshimi H, Takaichi S, Yanagisawa M, MasakiT. Binding and receptor down-regulation of a novelvasoconstrictor endothelin in cultured rat vascularsmooth muscle cells. FEBS Lett. 1988;239:13-17.

34. Le Monnier de Gouville A, Lippton H, Cohen G, Cav-ero I, Hyman A. Vasodilator activity of endothelin-1and andothelin-3: Rapid development of cross-tachy-phylaxis and dependence on the rate of endothelinadministration. J Pharmacol Exp Ther. 1990; 254:1024-1028.

35. Myers RR, Heckman HM, Galbraith JA, Powell HC.Subperineurial demyelination associated with re-duced nerve blood flow and oxygen tension after epi-neural vascular stripping. Lab Invest. 1991;65:41-50.

36. Myers RR, Yamamoto T, Yaksh TL, Powell HC. Therole of focal nerve ischemia and Wallerian degenera-tion in peripheral nerve injury producing hyperesthe-sia. Anesthesiology. 1993; 78:308-316.

37. Gaasterland DE, Tanishima T, Kuwabara T. Axoplas-mic flow during chronic experimental glaucoma: I:Light and electron microscopic studies of the monkeyoptic nerve head during development of glaucoma-tous cupping. Invest Ophthalmol Vis Sci. 1978; 17:838-846.

38. Quigley HA, Hohman RM, Addicks EM. Chronic ex-perimental glaucoma in primates: II: Effect of ex-tended intraocular pressure elevation on optic nervehead and axonal transport. Invest Ophthalmol Vis Sci.1980; 19:137-152.


An endothelin-1 induced model of optic nerve ischemia in the rabbit.

Documents

Transcript of An endothelin-1 induced model of optic nerve ischemia in the rabbit.