Isolated preparations of ocular vasculature and their applications in ophthalmic research

Progress in Retinal and Eye Research 22 (2003) 135–169

Isolated preparations of ocular vasculature and their applicationsin ophthalmic research

Dao-Yi Yu*, Er-Ning Su, Stephen J. Cringle, Paula K. Yu

Centre for Ophthalmology and Visual Science, The University of Western Australia, Australia

Abstract

The purpose of this review is to outline the techniques and applications for isolated ocular vascular preparations and their

significance to ophthalmic research. Various isolated ocular vascular preparations have been utilized in studies of ocular vascular

biology, physiology and pharmacology, including work in both normal and diseased conditions. However, there is still significant

potential for further studies to improve our understanding of the ocular circulation and its regulation. Experience has shown that

there is no single preparation capable of addressing all of the questions that must be answered if a complete understanding of

mechanisms of vascular regulation in the eye is to be achieved. Rather, it is necessary to select the appropriate preparation and

techniques to address each individual question in the most appropriate manner. In this review, particular emphasis is placed on the

applications for isolated ocular preparations and the relevance of such studies to our understanding of the pathogenesis of eye

diseases involving the vasculature. Examples are given where therapeutic approaches in diabetes and glaucoma are assessed in terms

of their impact on the vasoactive properties of the ocular vasculature.

A significant heterogeneity is present in the different parts of the ocular vasculature, not only in the structural and functional

properties of vessel itself, but also in terms of the tissue environment and innervation. A single vasoactive agent may also have

different effects when applied to the inside or the outside of the same region of a vessel. The vasoactive response of the vascular

system as a whole is what determines the rate of blood flow through the system, but this is regulated by a multitude of factors in

different regions of the vascular network. Isolating individual components of the ocular vasculature is readily achievable for the

extraocular vessels such as the ophthalmic or ophthalmocilliary arteries, which can be studied in myograph type systems measuring

the mechanical vasoactive force generated by the vessel. Retinal vessels from very large animals can also be studied in this way, but

the small diameter of the retinal vessels in most species requires a perfusion rather than myograph based technique. Perfusion based

studies of vessel diameter in response to vasoactive stimuli can be applied to individual retinal arteries and their branches. Perfusion

of more complex elements of the ocular vasculature such as isolated segments of the retina or ciliary body, or whole isolated

perfused eyes may use the perfusate pressure as the determinant of vasoactive state. However, when several components of the

ocular vasculature are being perfused simultaneously it may be difficult to separate out the contribution from the different vascular

elements.

The advantage of isolated preparations is that systemic influences can be eliminated, and vascular components can be studied that

are inaccessible in vivo. The disadvantage is that no matter how well controlled the in vitro environment may be, it will always be a

relatively poor mimic of the in vivo conditions. However, such in vitro work has certainly improved our understanding of the

vasoactive properties of different regions of the ocular vasculature in both health and disease.

r 2003 Elsevier Science Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

1.1. The retina is particularly vulnerable to vascular diseases . . . . . . . . . . . . . . . . . . . . . . 136

1.2. Ocular vasculature and regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137

2. Techniques and procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

2.1. Isolated perfused eye preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

2.1.1. Experimental apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

2.1.2. Tissue preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

*Corresponding author. Lions Eye Institute, 2 Verdun St. Nedlands, WA 6009, Australia. Tel. +61-8-9381-0716; fax: +61-8-9381-0700.

E-mail address: [email protected] (D.-Y. Yu).

1350-9462/03/$ - see front matter r 2003 Elsevier Science Ltd. All rights reserved.

PII: S 1 3 5 0 - 9 4 6 2 ( 0 2 ) 0 0 0 4 4 - 7

1. Introduction

1.1. The retina is particularly vulnerable to vascular

diseases

The retina performs the vital task of transducing andencoding the visual input for later processing by thevisual centres of the brain. The retina has the conflictingconstraints of a high metabolic requirement, yet thetransparency of the retina cannot be compromised by anoverly rich vascular network on the inner retinal side. Itis now established that specific layers within the retinadominate the oxygen requirements of the retina. Theselayers are the inner segments of the photoreceptors(Linsenmeier, 1986), and the outer and innerplexiform layers (Cringle et al., 2002; Yu et al., 1994b;Yu et al., 1999b; Yu and Cringle, 2001). This furthercompounds the problem of adequate provision ofnutrients and the removal of metabolic waste products

to maintain retinal homeostasis. The inner segmentsof the photoreceptors lie in a completely avascularregion of the retina, and the relatively sparse retinalvasculature needs to supply both the inner andouter plexiform layers. This can only be achieved byprecise regulation of vascular elements to matchlocal blood flow with tissue demands. Disruption ofvascular regulation can have severe consequences in theretina. In the western world 75% of blindness and visualloss is attributable to retinal diseases with a vascularcomponent (Cooper, 1990). Such diseases includediabetic and hypertensive retinopathy, glaucoma, age-related macular degeneration, central and brancharterial occlusion, and venous occlusion. Despite manyclinical studies and anatomical descriptions, the under-lying functional basis of these diseases remains poorlyunderstood. Thus, we are not currently in a position toprovide satisfactory procedures for early diagnosisand treatment.

2.1.3. Measurement of vascular resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

2.1.4. Optimal perfusion parameters and determination of extra ocular leakage . . . . . . . . . 140

2.1.5. Experimental regime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

2.2. Isolated arterial ring segment preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142


2.2.2. Tissue preparation and force measurements . . . . . . . . . . . . . . . . . . . . . . . . . 142

2.3. Isolated perfused retinal vessels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143


2.3.2. Vessel cannulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144

2.3.3. Experiment control and diameter and pressure measurement . . . . . . . . . . . . . . . . 144

2.3.4. Intraluminal and extraluminal drug delivery . . . . . . . . . . . . . . . . . . . . . . . . 145

2.4. Cryopreservation of human retinal arterioles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

2.5. Intracellular calcium measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

3. Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146

3.1. Study on the regulation of the normal ocular vasculature . . . . . . . . . . . . . . . . . . . . . . 146

3.1.1. The effects of vasoactive endogenous and pharmacologic substances . . . . . . . . . . . . 146

3.1.2. Other modulators and factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

3.2. Evaluation of vasoactive properties of anti-glaucoma drugs . . . . . . . . . . . . . . . . . . . . . 154

3.2.1. General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

3.2.2. Effect of b-adrenergic antagonists and Ca2+ channel entry blocker on ocular vasculature . 155

3.2.3. Vasoactive effects of a docosanoid and selected prostanoids on retinal arterioles . . . . . . 158

3.3. Applications in the pathogenesis of diabetic retinopathy . . . . . . . . . . . . . . . . . . . . . . . 158

3.3.1. General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

3.3.2. Agonist-induced vasoactivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161

3.3.3. Endothelium dysfunction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161

3.3.4. Tetrahydrobiopterin reverses the impairment of acetylcholine-induced vasodilatation

in diabetic ocular microvasculature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162

4. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162

4.1. Importance and relevance of isolated ocular vascular preparations . . . . . . . . . . . . . . . . . 162

4.2. Important areas to be further addressed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

4.2.1. The role of the endothelium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

4.2.2. Differential effects of intra- and extra-luminal administration of drugs . . . . . . . . . . . 163

4.2.3. Constant perfusion flow or constant pressure for isolated vessel and isolated eye

preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164

4.3. Possibility of prediction of in vivo from in vitro data . . . . . . . . . . . . . . . . . . . . . . . . 164

5. Summary and research directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

D.-Y. Yu et al. / Progress in Retinal and Eye Research 22 (2003) 135–169136

1.2. Ocular vasculature and regulation

The regulatory control of blood flow in the eye is avital component in the maintenance of retinal home-ostasis. Understanding how such control is achieved isfundamental to the development of appropriate ther-apeutic strategies aimed at restoring adequate bloodflow in disease states. However, this is no simple task, asthe mechanisms responsible for vascular regulation arecomplex, and within the different components that makeup the ocular vasculature there is a high degree ofheterogeneity in their vasoactive properties.

In general, the eye is supplied through an ophthalmicartery, which leads into the ciliary arteries feeding thechoroidal/uveal circulation, and a central retinalartery or cilioretinal arteries, which feed the retinalcirculation. These two circulations; choroidal andretinal, possess very different properties and constraints.The outer segments of the photoreceptors are where thevisual image is focused. This region is avascular toensure optimal visual acuity with minimum opticalinterference from any vascular bed. As a consequencethe photoreceptors’ main source of nutrients, thechoroidal circulation, lies totally outside the retina andsupplies the photoreceptor layer with nutrients such asoxygen by passive diffusion (Linsenmeier, 1986; Pour-naras et al., 1989; Yu and Cringle, 2001). It is knownthat the oxygen supply from the choroid is barelyenough to prevent some regions of the outer retina frombecoming hypoxic, which suggests that the high rate ofblood flow through the choroid may be essential tomaintain a high oxygen level in the choriocapillaris(Linsenmeier, 1986).

The choroidal circulation possesses both sympatheticand parasympathetic innervation (Laties and Jacobow-itz, 1966), presumably allowing systemic control ofchoroidal blood flow. However, there has been con-siderable disagreement as to whether the choroidalcirculation is capable of functional regulation (Alm andBill, 1970; Bill, 1962; Friedman, 1970; Hardy et al.,1994; Kiel and Shepherd, 1992; Yu et al., 1988).

In contrast the retinal circulation is differentlyconstrained in its design. Although it is responsible forfeeding a high metabolic rate tissue, it must beanatomically sparse to minimize optical interferencewith the light path to the photoreceptors. A furtherunusual feature of the retinal circulation is that it has noautonomic innervation (Laties, 1967), so total reliancemust be placed on local vascular control mechanisms.These requirements result in a limited flow circulation,with a high arterio-venous oxygen tension difference.This circulation has in general, two capillary beds, onefeeding into the nerve fibre/ganglion cell layer and otherfeeding the middle retinal layers including the innernuclear layer and plexiform layers. There is nocontroversy about the regulatory ability of the retinal

circulation. It has long been accepted that the retinalcirculation has powerful regulatory mechanisms. Hu-man and animal data demonstrate that flow in themajor vessels is regulated (measured by laserDoppler velocimetry) and that the circulation regulatesto blood pressure and hyperoxia (Grunwald et al.,1984a). Moreover we have demonstrated that in the ratretina the oxygen level in the region supported by thesuperficial capillary layer is well regulated, whilst that ofthe deeper capillary layer is not (Yu et al., 1994b). Thisapparent vulnerability of the deep capillary bed area isan important observation as it provides a possibleexplanation for the high incidence of pathologicalinvolvement of the deep capillaries in retinal vasculardisease (Yanoff and Fine, 1989).

The feeder vessels to the eye are also involved in theregulation of ocular blood flow (OBF) and it is knownthat their vasoactive properties can vary significantlyalong their length (Yu et al., 1992c). This adds yetanother dimension to the heterogeneity of vascularcontrol mechanisms in the ocular vasculature.

In any particular vessel there are a number ofcompeting or complementary mechanisms that areresponsible for locally regulating the vessel tone. Theselocal factors combine to ensure that the blood flow tothe tissue is matched to the metabolic requirements. Tounderstand the local control mechanisms, in vitropreparations have been used to study the vascularreactivity of different components of the ocular vascu-lature, determining their response to blood bornefactors, tissue released factors, and factors releasedfrom the autonomic system (Fig. 1).

This integration of total blood flow is known to beachieved by the continuous and dynamic interplaybetween many regulatory factors, including factorsemanating from the blood, the endothelial and smoothmuscle cells of the vessel walls, the surroundingmetabolizing tissue and the input pressure.

Several hypotheses have been tested in an attempt tounderstand whole organ regulation in other organs suchas the brain and kidney (Defily and Chilian, 1995;Holstein-Rathlou and Marsh, 1994), but no suchhypothesis has yet been proposed for the ocularcirculations. To partially remedy this deficit the researchcovered in this review takes the first step in unravellingcontrol mechanisms of the individual components of theocular circulation.

The vascular endothelium is a vital component ofvascular regulation. It consists of a monolayer of thinsquamous cells, which line the inside surface of bloodvessels. One intracellular structure implicated in sensingexternal changes and mediating the output of the hugearray of autocoids known to change smooth muscle cellresponse, is the cytoskeleton. The cytoskeleton gives thecell its shape as well as mediating the transmission ofintracellular signalling. The response of the endothelium

D.-Y. Yu et al. / Progress in Retinal and Eye Research 22 (2003) 135–169 137

to shear stresses associated with local blood flow isanother important mechanism for regulation of bloodflow.

Given the complexity of all the factors involved in theregulation of vascular tone, it is perhaps not surprisingthat systemic diseases such as diabetes disrupt thenormal control mechanisms. Of all the vascular-baseddiseases of the retina, diabetic retinopathy is theprobably the most extensively studied and documented(Cogan et al., 1961; Davis, 1992; Frank, 1995). Thefrustrating feature of human diabetic retinopathy is thatthe disease follows a long, clinically silent course, duringwhich undetectable vascular and neural damage occur,some of which are irreversible by the time the damage isrevealed clinically. It is vital that the cascade of vascularchanges is better understood, and in this respect theavailability of rat models of induced diabetes provides auseful avenue for research (Su et al., 2000; Yu et al.,2001c).

Although a vascular component in glaucoma is morecontroversial, there is an increasing body of evidencethat blood flow changes are involved. There is alsoevidence that some glaucoma medications may helprestore retinal blood flow in addition to their intraocularpressure (IOP) lowering effect (Drance, 1997; Yu et al.,1998). There is clearly scope for investigating newtherapeutic agents that may beneficially influence OBFflow in disease states.

This review outlines the principles and results from arange of studies using isolated preparations ofdifferent elements of the ocular vasculature. Usingsuch an approach we hope to piece together theindividual mechanisms that are the key players in thenormal regulation of OBF. Studies in diseasedanimals begin to identify the mechanisms that areresponsible for vascular dysfunction. It is hoped thatsuch information will ultimately enable a better-targetedapproach to be developed for the treatment of ocularvascular disorders.

2. Techniques and procedures

The techniques we describe are designed to explorethe vasoactive properties of the ocular circulation or itscomponents. We begin with the isolated perfused eyepreparation in which the entire ocular vasculature isstudied as a whole. We then outline the means by whichvasoactive responses of individual vessel can be assessed,moving down the vascular tree from the main extrao-cular arteries to successively smaller vessels, finallyreaching the retinal arteries.

2.1. Isolated perfused eye preparation

The use of the isolated perfused eye preparationallows us to assess the vasoactive properties of the intactocular vasculature. The isolated perfused eye techniquewas pioneered by Gouras and Hoff (1970) and furtherrefined by Niemeyer, who recently reviewed thenumerous applications to which this preparation hasbeen put (Niemeyer, 2001). The most extensivelystudied species for such work is the cat, but we havealso used the same technique with eyes from monkeys,dogs, pigs, guinea pigs and rats. Clearly, with smallereyes, such as those of the rat, the technical demands ofcannulating the ophthalmic artery increase, but thereare significant advantages in terms of the readyavailability of established rat models of retinal disease.A good example is the STZ rat model of diabetes.Although such models are available in larger animals,the rat model has the advantage of being far lessexpensive to maintain, and the relatively fast time scaleof the disease progression in the rat allows largernumbers of animals to be studied over the full lifespanof the animal (Su et al., 2000). We shall describe oursystem for perfusing isolated rat eyes and explain therational behind some of our decisions about optimalperfusion conditions.

Fig. 1. Schematic illustration of local control mechanisms in the vascular system. The vasoactive effects of blood borne factors, myogenic stretch

responses and endothelial cell released factors, and tissue generated factors combine to produce the local regulation of vasoactive tone to regulate

blood flow.


2.1.1. Experimental apparatus

Fig. 2 shows a schematic of the organ bath, the eyeand perfusion system. The organ bath was a speciallyconstructed, indented bath surrounded by a thermostat-controlled jacket through which water flowed at asufficient rate and temperature to maintain bathtemperature at 371C. The bath contained oxygenatedNa+-Krebs solution bubbled with carbogen. The eyewas perfused with carbogen-bubbled Na+-Krebs andcarefully covered to prevent drying of scleral tissue. Apressure transducer connected to a bridge amplifiermonitored the perfusate pressure (P). The setting of flowvalues and data acquisition of pressure values was undercomputer control. The software for this was developedusing the graphical programming language LabView(National Instruments, USA) and run on a conventionalpersonal computer with the appropriate A/D andcommunications cards added. Perfusate flow wasdelivered by a computer controlled syringe pump(Model 22, Harvard Apparatus Inc., MA, USA). Drugsto be delivered intraluminally were administered as a5 ml bolus into the perfusate stream via the sampleinjector valve (7725i, Rheodyne Inc., CA, USA). Theintrinsic design of this injector system permitted primingand inclusion of the drug bolus into the perfusatepathway without any pressure artifact or problems withair bubbles. The volume, and hence the duration of thebolus, was determined to be sufficient for the vasoactivecontraction to stabilize and then recover as the drug waswashed away by the Krebs perfusate. Input pressure andperfusion flow were continuously recorded on a chartrecorder (Yokogawa, LR8100, Japan) and stored to diskwhen required.

2.1.2. Tissue preparation

Each rat was anaesthetized with an intraperitonealinjection of sodium pentobarbitone (Nembutal,50mg kg). The jugular vein was cannulated to allow 30units of heparin to be injected to prevent clotting. After10min both eyes were enucleated ensuring the presence

of a long optic nerve and ophthalmic artery (5–7mm)for subsequent cannulation. The rat was then killed withan anaesthetic overdose. One of the enucleated eyes wasplaced in the organ bath for cannulation and perfusion,and the other was immediately placed in Na+-Krebssolution bubbled with carbogen (95% O2, 5% CO2) andwas maintained at 41C until required. Prior to cannula-tion in the organ bath, the eye was first carefullydissected free of most extraneous tissue, the ophthalmicartery was dissected free, and a cannulation location waschosen sufficiently proximal from the eye to ensure thatall extraocular branches feeding intraocular vessels wereperfused (> 3mm proximal to the disk).

2.1.3. Measurement of vascular resistance

The aim was to perfuse the eye with a constant flow ofoxygenated Na+-Krebs solution, whilst continuouslymonitoring the input pressure just prior to the cannula.The resistance of the cannula is determined prior tocannulation, by noting the perfusate pressure at thegiven flow rate with the open end of the cannula placedunder fluid to remove surface tension effects. Undermicroscopic visualization the ophthalmic artery (outsidediameter B100 mm) was cannulated using a taperedglass pipette and tied with 10/O suture (Ethicon, UK)after which perfusate flow was commenced. The ocularvascular resistance can then be calculated by subtractingthe cannula resistance from the total resistance of thecannula and the eye. Simply dividing the input perfusatepressure by the perfusate flow rate does not give areliable index of total vascular resistance. This is becausethe IOP also opposes the flow of perfusate in theintraocular compartment. The equivalent circuit for thisarrangement is shown in Fig. 3. Since IOP is notroutinely measured in isolated eye preparations wedeveloped a simple procedure for quantifying vascularresistance independently of IOP. The technique involvessystematic increases or decreases in perfusate flow ratewhilst monitoring the input perfusate pressure. We wereable to automate this procedure by taking advantage ofthe computer control of the syringe pump, but manualadjustments of perfusate flow rate could achieve theequivalent data. Fig. 3 shows a data set where the flowrate through an isolated rat eye was ramped up anddown over a small range of flow values. Plotting therelationship between flow rates and resulting perfusatepressure allows the slope of the relationship to bedetermined. This parameter (vascular resistance+can-nula resistance) requires no knowledge of IOP, and alsoeliminates any offset errors in the calibration of thepressure transducers. In the example given the resistanceof the cannula alone was 1.06mmHg/ml/min, whilst thatwhen the eye was added was 3.87mmHg/ml/min. Thevascular resistance of the ocular vasculature in this casewas 2.81mmHg/ml/min. It is important that thedominant resistive element in the circuit is the ocular

Fig. 2. Schematic of isolated perfused rat eye and organ bath, showing

constant flow perfusion by computer controlled syringe pump, input

pressure measured by transducer, the valve through which bolus doses

of the vasoactive agents were added to the perfusion line, leading into

the cannula and the eye maintained in an environmentally controlled

organ bath (Su et al., 1995).


vasculature, as this produces easily recordable pressurechanges when vasoactivity is modulated. If the cannulawere the dominant resistive element then perfusatepressure changes would be muted even in the face ofsignificant vasoactive responses.

2.1.4. Optimal perfusion parameters and determination of

extra ocular leakage

Considerable care and effort was spent on a series ofpreliminary experiments, aimed at achieving a perfusedeye in the optimal physiological condition. We investi-gated choice of perfusate, flow rate, perfusion pressure,and volume of agonist delivered, and methods ofdetermining the extent of extra ocular vascular leakage.

2.1.4.1. Perfusate. In vivo the eye is perfused with highviscosity blood. Using isolated perfused organs forpharmacological experiments has always involved acompromise between using a protein-free isotonic

solution like Krebs, which has the disadvantage of lowoncotic pressure and low viscosity, against a higherviscosity protein-containing solution such as Ficol orHaemaccel, which have the disadvantage of binding anunknown fraction of the pharmacological agents beingtested. Although most previous pharmacological inves-tigations using isolated perfused organs have usedprotein-free solutions, we determined for ourselveswhether it was possible to use higher viscosity solutionsin the eye to mimic the properties of blood flow moreclosely. Therefore, we performed extensive preliminaryexperiments with 16 eyes, varying the perfusing solutionfrom Krebs to Krebs with 0.5–2% albumin as anadditive, to Haemaccel. We found that protein-enrichedsolutions significantly reduced pharmacological re-sponses and also resulted in variable results. This waspresumably a consequence of agonist binding withprotein. Thus, in common with other perfused organstudies, we finally settled on oxygenated Krebs solution.

Fig. 3. Equivalent circuit of perfusion system and isolated perfused eye and typical pressure–flow relationship in an isolated perfused rat eye. The

slope of the pressure–flow relationship indicates the resistance to perfusate flow independently from the influence of the IOP.


Normal Na+-Krebs solution is composed of (in mM)NaCl 119; KCl 4.6; CaCl2 1.5; MgCl2 1.2; NaHCO3 15;NaH2PO4 1.2; Glucose 6.

This choice of perfusate, which has a viscositysignificantly less than blood, makes it impossible toachieve both normal pressure and flow throughout theocular circulation. One can either set the flow to be aclose approximation to the in vivo situation with theresulting pressure being lower than normal, or alter-natively one can set the pressure to the in vivo value witha very much higher flow than normal. Since bothpressure and flow are known to influence the vasoactiv-ity of the vessel, one is faced with the decision of whichparameter to fix. We chose to set the flow rate, a choiceparticularly suitable for a syringe pump delivery systemin which the flow rate is easily set and maintained. Thealternative choice of fixing the perfusate pressure at theinput to the eye to in vivo levels does not guaranteephysiological pressures at locations further down thevascular tree since the low viscosity of the perfusatewould result in lower pressure gradients down thevascular tree and higher pressures at the arteriole andcapillary level.

In the absence of knowledge of the normal ratophthalmic artery flow rate, it was necessary to baseour flow on the size of the vessel using typical valuesfrom the literature for similar sized vessels. The ratophthalmic artery has a diameter of only 100 mm. Otherarteries with similar diameter such as 100 mm retinalarteries from larger eyes have flow rates similar to5 ml min�1 (Riva et al., 1985), as do cerebral vessels ofthis size (Duling and Rivers, 1986). When this flow ratewas used for the perfused rat eye, the basal ocularresistance was stable over several hours and thepharmacological responses repeatable. At this flow rate,input perfusate pressure is lower than seen in vivo due tothe perfusate having lower viscosity than blood. Typicalinput pressure values for other perfused organs wherethe flow is matched to in vivo values are also lower thanin vivo (Bhattacharya et al., 1982; Hendriks et al., 1993).A perfusate flow rate of 5 ml/min therefore seems areasonable estimate of normal flow rate in the rat eyeand this value was used as the baseline for all ourvasoactivity studies in isolated perfused rat eyes.

2.1.4.2. Determination of leakage from extraocular

vasculature. Depending on the specific vascular patternin a particular eye or species, some leakage fromextraocular perfusion pathways may be inevitable.Identifying and tying off such leaking vessels mayreduce the problem but this is a painstaking exercisegiven the transparent nature of the perfusate and thecomplexity of the vasculature involved. We havedetermined a useful procedure for quantifying thepercentage of total perfusate flow that is leaking fromthe extraocular pathways. The principle of the technique

is that manipulation of the IOP affects only theperfusate flow that is passing through the intraocularvasculature. In a constant flow delivery system, asprovided by a syringe pump, the perfusate pressurereflects the total vascular resistance offered by theperfused vasculature, an offset due to the resistance ofthe cannula and associated plumbing, and an offset dueto the IOP. Under these conditions, a change in IOPcreates a change in the measured perfusion pressure,from which the percentage of perfusate flow through theintraocular pathway can be determined. At the two endsof the spectrum, 100% intraocular flow (ie no leakage)results in a 1:1 relationship between changes in IOP andperfusate pressure, whilst 0% intraocular flow (allleakage) results in no association between IOP andperfusate pressure. In this manner the extent ofextraocular leakage can be quantified, and the successof strategies aimed at reducing such leaks can bedetermined. This may be an important issue in situationswhere the vasoactive properties of the ocular vasculatureare to be assessed by monitoring an induced change invascular resistance. A high degree of leakage or shuntingof perfusate through the extraocular vasculature maymask the presence of significant changes in vascular toneof the intraocular vasculature. The data in Fig. 4A ashows the relationship between IOP and perfusatepressure in an isolated pig eye. The slope of the bestlinear fit is 0.75, indicating an intraocular flowpercentage of 75%, which was typical of our experiencewith pig eyes, which have a relatively complex extrao-cular branching pattern (de Schaepdrijver et al., 1992;Simoens et al., 1992) creating the potential for numerousleakage pathways after enucleation.

Fortunately, the eyes of some species have a vasculardistribution that routinely allows extraocular leakage tobe minimized. This is true of the isolated perfused rateye and the isolated perfused guinea pig eye. The rat andthe guinea pig have an ocular vascular system that is fedfrom a single ophthalmic artery that has very fewbranches outside the intraocular compartment. There isconsequently very little chance of significant leakagefrom severed branches subsequent to enucleation andsuccessful cannulation of the ophthalmic artery. Fig. 4Bshows the relationship between IOP and perfusatepressure for an isolated perfused rat eye. In this casethe percentage of intraocular flow was 90% withouthaving to resort to tying off any leaking vessels.Vasoactive responses of the intraocular components ofthe ocular circulation in the rat are therefore relativelyunaffected by extraocular leakage.

2.1.5. Experimental regime

The system was left to stabilize until the basalpressure and hence vascular resistance were stable,usually 30min. A control injection of 5 ml Na+-Krebssolution was given through the valve to confirm an


artifact-free null response. Increasing log M concentra-tions of the selected vasoactive agonist were loadedsequentially into the valve at 10min intervals to producea dose/pressure response curve. At the end of one dose-response trial a repeat control injection was alwaysperformed to ensure that there was no residual drugresponse before a new drug dose-response trial wascommenced. The sequence of agonists was randomlyvaried between experiments. The bath solution wasfrequently replaced to avoid agonist build-up.

2.2. Isolated arterial ring segment preparation

The extraocular vessels in most animals are largeenough to use a myograph-based technique where theforce generated by the vessel can be monitored in order

to assess vasoactive responses. We have used thistechnique in studies of the vasoactive properties ofhuman posterior ciliary arteries (Yu et al., 1992b) and indifferent sections of the cat ophthalmociliary artery (Yuet al., 1992c).


A schematic of our myograph system is shown inFig. 5. It was custom made and consisted of a hollowcylindrical stainless-steel block (B) in which was aconcentric cylindrical depression (volume 10ml) formedthe myograph bath. An isometric force transducer (T)and an XYZ microdrive system (M), driven in the X

direction by a hydraulic microdrive, were mounted on abaseplate on either side of the organ bath. These wereused to measure the ‘‘isometric’’ tension developed bythe muscle and to apply the passive stretch, respectively.The two bathing solutions (composition as required)were held in glass containers with outer water jackets.Bath fluid, and bathing solutions were maintained at371C, by circulating water from a temperature baththrough the stainless-steel base block and outer jacketsof the glass containers. All connective tubing was shortso that temperature gradients were minimized. Bathtemperature, pH and PO2 were continuously monitoredand displayed digitally and on a chart recorder andmaintained close to 371C, 7.4 and 95% O2, respectively.The two organ bathing solutions were delivered to thebath passively under a hydrostatic pressure head byopening a valve (V) in the appropriate line. Thesesolutions, plus that in the bath, were equilibrated withcarbogen (95% O2, 5% CO2), by continuous slowbubbling. When a change of bath fluid was required thechamber was emptied by suction at D, and the feed linefor one solution opened until the bath was full to thepoint where the suction overflow line became opera-tional.

2.2.2. Tissue preparation and force measurements

Two 75 mm tungsten wires (W), coated with gold paintto render the surface smooth, were threaded throughopposite ends of the vessel lumen. All handling of thevessel was by means of forceps and connective tissueuntil the vessel was completely mounted on the tungstenwires, after which the connective tissue was carefullydissected away. The tungsten wires were then clipped tothe two steel yokes (Y), which were rigidly attached tothe transducer and microdrive system, respectively, sothat the wires lay parallel to each other. The bath wasfilled with the Na+-Krebs bathing solution and thevessel left to equilibrate for no less than 30min underzero tension. The two tungsten wires were stiff enoughto remain parallel throughout all active and passivetension changes. As the passive or active tension varied,the resulting change in distance b was directly measuredby means of a calibrated beam splitter. The isometric

Fig. 4. Relationship between input perfusate pressure and manipu-

lated intraocular pressure in isolated perfused pig (A) and rat (B) eyes.

The 451 slope seen in the rat eye indicates minimal leakage from the

extraocular vasculature.


tension output from the transducer was amplified,filtered (DC to 3Hz) and continuously recorded on achart recorder.

Using such a system, our group have studied thevasoactivity of isolated arterial ring segments ofophthalmic arteries or long posterior ciliary arteries(LPCA) from various species such as cats, dogs, pigs orhuman donors (eye bank eye or surgical specimens).

2.3. Isolated perfused retinal vessels

For vessels less than about 200 mm in diameter themyograph technique becomes unsuitable. It is notfeasible to reliably insert the two stretching wires intothe lumen. The alternative is a microperfusion system inwhich the monitored parameter to assess vasoactivity isthe diameter of the vessel. We adapted the perfusionsystem described by Duling and Rivers (1986), andcreated a system for perfusing retinal vessels in theregion of 100mm in diameter (Alder et al., 1996; Su et al.,1996; Yu et al., 1994a).


A schematic of our system is shown in Fig. 6. Anisolated retinal artery is cannulated at both ends andperfused intraluminally in the orthograde directionwhilst being maintained in an environment controlledbath. Vessel diameters are measured on-line by digitalanalysis of the microscope image. Krebs solution isnormally used for perfusion, but could be switched toother solution such as a 124mM K+ -Krebs solutionwhen a sustained potassium contracture was required.Alternatively, retinal vessels could be pre incubated withchosen agents extraluminally. For example, precontrac-

tion could be induced by an extraluminal dose of ET-1(10�9M). The effect of either intraluminally or extra-luminally delivered test agents on the outer diameter ofthe precontracted vessel can then be assessed.

2.3.1.1. Incubator chamber. The incubator chamber wasmounted on the stage of an inverted microscope. Thechamber contained 5ml Krebs solution maintained at371C using a bipolar temperature controller. Theincubating solution was equilibrated with 95% O2, 5%CO2 gas flowing over the surface of the chamber so as tomaintain PO2, PCO2 and pH of the incubating solution.This was verified by occasional aspiration of samples forblood gas analysis.

2.3.1.2. Cannula and perfusion system.General. The principle used in these experiments was

that vessels with a side branch were cannulated at bothends, and perfused through one end (proximal) with aflow of 5 ml/min in the orthograde direction with the sidebranch acting as the exit route, as shown in Fig. 6. Theother end (distal) was generally left with a low flow(0.3 ml/min). This residual flow helped ensure that drugsdelivered from the other pipette did not collect in thedistal end of the vessel. The vessel diameter wasmeasured on-line by a computer analysing the user-selected window in a frame-grabbed image of the vessel.Diameter changes in response to intraluminal orextraluminal delivery of test agents were compared.

Cannulas. Each cannula system consisted of con-centric pipettes (Fig. 6). The outer or holding pipettewas shaped on its innermost surface to present twoconstrictions. The constriction nearer the open end ofthe pipette served as a surface against which the inner

Fig. 5. Schematic myograph system and associated plumbing, inset shows yolk type mounting system for ring segments (Yu et al., 1992c).


pipette, the pipette that entered the vessel lumen,squeezed and sealed the vessel wall. The innermostconstriction served to centralize the tip of pipette and toseparate it from the inner wall of holding pipette.Another pipette, delivered the perfusate and intralum-inal drugs directly at the shoulder of the perfusingpipette. The fluid resistance of the smallest pipetteoffered less resistance to perfusate flow than thatpresented by a typical second order branch. Thus anyvasoactive effects on the vessel or the branch result in amodulation of perfusate pressure at the monitoringpoint just distal to the pipettes.

Perfusion and mounting system. Perfusate flow wasdelivered by a computer controlled syringe pumpcontaining a 5ml gas tight syringe connected to theinnermost pipettes. A specially manufactured pipetteholding system was developed which allowed relativemovement between holding and inner pipettes whilstmaintaining a pressure seal and minimal compliance.The whole assembly was mounted on a joystickcontrolled XYZ microdrive and angled at 351 to thehorizontal. An equivalent system of pipettes, manipula-tors, and pumps was used for either end of the vessel.

2.3.2. Vessel cannulation

The vessel was positioned horizontally in the incuba-tion bath close to the bottom of the dish. By conventionthe proximal end of the vessel was mounted on the lefthand pipette system. The XYZ drive was manipulated toplace the tip of outer pipette adjacent to the proximalend of the vessel. Relative movement between outer andinner pipettes was achieved using a hydraulic microd-rive. The inner pipette was retracted into the cannulaand gentle suction was applied to the outer pipette todraw the end of the vessel into the cavity between thetwo constrictions. The inner pipette was then advanced

into the vessel lumen such that it squeezed the vessel wallagainst the constriction. This procedure was performedwhilst a continuous perfusion of 5 ml/min was flowing. Itwas found that this helped keep the vessel lumen open.Once this end of the vessel was sealed the perfusate flowflushed out the red blood cells from the lumen of thevessel and a similar procedure was repeated on the otherend. Once both ends were cannulated, perfusion deliveryat the distal end was reduced to 0.3 ml/min. The vesselwas left to stabilize for 30min.

2.3.3. Experiment control and diameter and pressure

measurement

All of the data recording and much of the instrumentcontrol was under computer control. The software wasdeveloped using the graphical programming languageLabView, and run on a conventional personal computerwith the appropriate serial and IEEE communicationscards added. The inverted microscope image of thevessel was captured with a CCD camera, displayed on acolour video monitor, and frames were grabbed at twosecond intervals. The external diameter of the vessel wasmeasured on-line at the chosen location, and theperfusate pressures were monitored at key points byconventional transducers. The monitoring of pressure attwo points in each pipette system was found to be avaluable aid in detecting small leaks or partial blockagesin the system. These pressure signals, together withvessel diameter, the flow rate of the two pumps, andinjection signals from the sample injector valves were allrecorded on a chart recorder. Using IEEE-GPIBcommunication between the host computer and thechart recorder all 8 channels were sampled every twoseconds and the data reproduced on the computerscreen and streamed directly to a spreadsheet file whenrequired.

Fig. 6. Schematic representation of the experimental apparatus for isolated perfused retinal vessels. Selective delivery of intraluminal or extraluminal

test agents can be investigated whilst maintaining physiological flow rates through the vessel. Vessel diameter is monitored by on-line image analysis

of digitised microscope images at the chosen vessel location (Yu et al., 1997).


2.3.4. Intraluminal and extraluminal drug delivery

Drugs to be delivered intraluminally (IL) wereadministered as a 5 ml bolus into the perfusate streamvia the sample injector valves. The intrinsic design ofthis injector system permitted inclusion of the drugbolus without any pressure artefact or problems with airbubbles. A built in switch allowed a signal to begenerated to indicate loading and injection phases of theprocedure and the chart recorder and the computerrecorded this signal. At a typical perfusion rate of 5 ml/min the drug arrived at the vessel about 90 s afterinjection. Spreading of the original bolus in transitresults in a slight dilution of the drug but this is smallcompared to the log unit increments in concentration.The size, and hence the duration of the bolus, wassufficient for the vasoactive response to stabilize.Extraluminal drug delivery was achieved by directpipetting into the incubating solution to achieve therequired concentration. The bathing solution wasflushed frequently.

We have used this system to study vasoactiveresponses of retinal arteries from pigs and human donoreyes. Using a dissecting microscope the eyes weresectioned at pars plana ciliaris, separating the anteriorsegment and adherent vitreous body from the posteriorpole. The retina, choroid and sclera were divided intoquadrants using a razor blade, taking care not to sectionany major retinal arterioles. The retina was carefullyseparated from the underlying choroid and sclera usingan iris speculum. A quadrant of retina was placed on ahollowed glass slide containing Krebs solution kept atless than 41C. Using a combination of transmitted lightwith a stereomicroscope, individual first order retinalarterioles (outside diameter 60–120 mm) were dissectedfree of retinal tissue with a fire polished micropipette. Asegment of arteriole about 800–1500 mm long wasselected, in most experiments care was taken to includea side branch of outside diameter B50 mm whereas in afew cases a segment with no side branch was chosen forstudies of pressurized but not perfused vessels. Thisallowed the study of myogenic vasoactive responses inthe absence of intraluminal flow.

2.4. Cryopreservation of human retinal arterioles

The restricted availability of freshly available humaneyes meeting the criteria for suitability of vascularstudies meant that considerable effort was warranted indeveloping cryopreservation techniques to preservevascular function in human tissue. In this way it becamefeasible to perform many experiments on individualsegments of retinal arteries dissected from each donoreye. The generally agreed on requirements for successfulcryopreservation are to place cells in a high osmolaritysolution, use a slow rate of freezing, and ensure fastthawing with gradual dilution (Bank and Brockbank,

1987). Vessel segments from both human and pig eyeswere used in preliminary testing of various cryopreser-vation regimes. The optimum regime proved to be thesame for both species, resulting in responses in the pigvessels only slightly smaller than measured in freshtissue. The vessels were carefully dissected into segmentsfrom a region about two to four disk diameters from thedisk margin under microscope control in a dishcontaining Krebs solution at 41C. 8 to 10 vesselsegments from the same eye were placed in separate2ml cryotubes containing the cryopreservation mediumand allowed to equilibrate in a refrigerator at 41C for60–90min, before their transfer to a controlled ratefreezing system (Su et al., 1998). The cryopreservationmedium contained 1.8M DMSO, 50% fetal calf serum,0.1M sucrose, combined with Hepes buffered Krebs-Henseleit solution (mmol/L) (NaCl 118, KCl 4.7,KH2PO4 1.2, MgSO4 1.2, CaCl2 1.2, Glucose 11,NaHCO3 25, EDTA 0.03, HEPES 10) with a pH of 7.4.

The cryotubes containing the retinal arteriolar seg-ments were placed in the control rate freezer using thefollowing protocol: from 41C to �701C at a rate of�11C/min, followed by �701C to �1961C at a rate of�51C/min and finally placed into liquid N2 for storage.For thawing the cryotubes were placed into 371C Krebssolution in the water bath, and monitored continuouslyto ensure the vessel segments were removed before theice had melted completely, the vessel segments were thenplaced into Krebs solution at 371C. This whole thawingprocess takes about 2min.

Each vessel segment was subsequently tested forvascular pharmacological function. Segments wereplaced in the organ bath cannulated and perfused, anddiameter responses monitored and compared to those offreshly enucleated tissues.

2.5. Intracellular calcium measurement

Each of the isolated vessel preparations that we havedescribed is also amenable to studies of intracellularcalcium levels during induced vasoactivity. One of ourinverted microscopes was also equipped with a dualwavelength fluorescence excitation system (PTI deltascan) and high-speed photometers. Incorporation ofFura-2, a calcium sensitive fluorophore, into the smoothmuscle cells, allows a ratio measurement (340 nm/380 nm) to reflect the level of intracellular calcium atsample rates of up to 650 measurements per second. Inthe case of ring segment preparations the forcegenerated by the ring segment could be simultaneouslymonitored in real time. In this review we will summarizeour [Ca2+]i investigations in the isolated ring segmentsof LPCA preparations. Schematic representation of thesystem used to measure intracellular calcium and thecontractile force generated by a ring segment is shown inFig. 7.


3. Applications

The isolated ocular vascular preparations that wehave described have been used to study the physiologicaland pharmacological regulation of OBF in the normaland diseased conditions. Over the last 10 years we havebeen particularly interested in three aspects of vascularregulation:

* Control mechanisms of the healthy ocular micro-circulation, particularly the effects of vasoactiveendogenous and pharmacologic substances on thevascular smooth muscle and the mechanism by whichthese substances alter vascular tone.

* Pathogenesis and therapeutic strategies of oculardiseases suspected to involve vascular abnormality.

* Characterization of the vasoactive properties ofselected ophthalmic drugs currently in clinical useas IOP lowering agents.

3.1. Study on the regulation of the normal ocular

vasculature

In this review we will introduce the effects ofvasoactive endogenous and pharmacologic substanceson the vascular smooth muscle and the mechanism bywhich these substances alter vascular tone. Althoughmost ocular vessels consist of only one cell layer of

endothelium and a few layers of smooth muscle cells,clear-cut definitions of the mechanisms and effects ofmany vasoactive substances have not yet been obtained.

3.1.1. The effects of vasoactive endogenous and

pharmacologic substances

Although our review cannot be encyclopaedic eitherin providing definitive answers to each of the questionsor in covering the entire pharmacological studies in theocular vasculature, it is hoped that the presentedinformation from isolated ocular vascular preparationswill serve as a summary of our current knowledge and asa basis for considering fruitful avenues for furtherinvestigation.

3.1.1.1. Adrenoreceptor agonists. The microvasculatureof most species is endowed with receptors on the smoothmuscle cells for catecholamines. Adrenergic receptorshave been classified into a- and b-adrenergic receptorsand subclassified into a1 and a2 and b1;b2 and b3: a-adrenergic receptors may be presynaptic or/and post-synaptic. b-adrenergic receptors are located primarily onsmooth muscle cells and pre synaptically.

Unlike most vasculature throughout the body, theretinal vasculature has no (or very weak) autonomicinnervation (Laties, 1967). A controversy exists concern-ing the responses of the retinal vasculature to catechola-mines. Sympathetic stimulation and intra-arterial

Fig. 7. Schematic representation of the system used to measure intracellular calcium levels [Ca2+]i and the contractile force generated by a ring

segment of LPCA. Dual wavelength excitation of a Fura-2 loaded vessel allows the ratio of the emitted intensities to reflect the [Ca2+]i level during

K+ contraction and subsequent administration of betaxolol (Yu et al., 1997).


injection of noradrenaline (Alm, 1972) appear to haveno consistent effect on the retinal circulation. Theexistence of adrenergic a1 and a2 binding sites in theretinal vessels has been demonstrated (Forster et al.,1987). However, direct vascular responsiveness must bedemonstrated before concluding that any inducedvasoactivity may be important. Using ring segmentpreparations, it has been shown that noradrenaline,adrenaline and phenylephrine induce a contractiveresponse in bovine retinal arteries (Hoste et al., 1989;Nielsen and Nyborg, 1989a) the cat ophthalmociliaryartery (Yu et al., 1992c), and in human LPCA (Yu et al.,1992b). We have also shown that noradrenaline,adrenaline and phenylephrine induce vascular contrac-tion in isolated perfused eyes (Su et al., 1995), in isolatedring segments of ophthalmic artery, and in retinalarteriole and vein preparations (Alder et al., 1993; Yuet al., 1999a, 1992c, 1994b). Studies in isolated perfusedretinal arteries also demonstrated an asymmetry in theresponses to adrenergic agonists with contractionssignificantly larger when the drug was applied to theintraluminal surface rather than extraluminal surface(Yu et al., 1994a). These results imply that a-adrenergicreceptors are present in the ocular vasculature. a1-adrenergic receptors mediating contraction are presentwhereas a2-adrenergic receptors appear to be absent asdemonstrated in the cat ophthalmociliary artery (Yuet al., 1992c).

The smooth muscle cells of the majority of themicrovasculature appear to be endowed with b-adre-nergic receptors, which mediate vasodilatation, suppres-sion of the responses to adrenergic nerve stimulation,and inhibition of the response to noradrenaline and 5-HT. However, there is clear evidence that much of theocular vasculature is relatively devoid of b-adrenergicreceptors, and, as demonstrated in vitro in severalspecies of animals and human, that any b-adrenergicreceptors present are at best only weakly functional(Hoste et al., 1989, 1990; Nielsen and Nyborg, 1989a, b;Su et al., 1995; Yu et al., 1999a, 1992c, b). Thus, onewould anticipate a minimal b-adrenergic effect in theretinal vasculature.

3.1.1.2. 5-Hydroxytryptamine. 5-hydroxytryptamine (5-HT) released from the vascular wall or the platelets mayact as a local modulator of microvascular tone. 5-HTcan directly affect the smooth muscle cells, or indirectlyact through adrenergic nerves to alter neurogenicallymediated tone. The contractile responses of smoothmuscle cells to 5-HT are believed to be receptormediated. 5-HT receptors are located on the membraneof smooth muscle cells and cross-reactivity with a-adrenergic receptors may exist. Using isolated arterialring segment preparations, contractile responses to 5-HT have been demonstrated in retinal arteries (Hosteet al., 1990), and long posterior ciliary and ophthalmic

arteries (Yu et al., 1992c, b). 5-HT induced contractionshave also been observed in isolated perfused eyepreparations (Su et al., 1995). These results imply that5-HT receptors are present in the ocular vasculature.

3.1.1.3. Histamine. Histamine is a potent mediator ofboth vascular tone and peripheral edema. The concept isgenerally accepted that microvascular responses tohistamine are mediated by the classical histaminereceptors (H1 and H2).

Histamine is known to induce endothelium-dependentrelaxation in bovine retinal arteries (Benedito et al.,1991a). We have extensively studied the effects ofhistamine on the ocular vasculature, and demonstratedthat histamine induces potent contractile responses inthe cat ophthalmociliary artery (Yu et al., 1992c), andthat the ophthalmociliary artery is heterogenous in itsresponse to histamine, with the most proximal segmentproducing the largest responses compared with the mostdistal segment, nearest the eye (Yu et al., 1992c).However, in the human posterior ciliary artery, hista-mine produced biphasic responses with a mild relaxationfor low concentrations and small contraction for highconcentrations (Yu et al., 1992b). More recently, wehave compared the differences between histamine-induced responses in retinal arterioles and in theposterior ciliary artery in the pig (unpublished data).We demonstrated that histamine induces opposingvasoactive effects at different levels of the porcineocular vasculature. In retinal arterioles histamineproduced a dilation, whilst in the LPCA, histamineproduced a contraction. In examining the mechanism ofaction in the retinal arterioles, the involvement of H2

receptors was implied by the attenuation of thehistamine response in the presence of cimetidine, anH2 antagonist. Further attenuation of the histamineresponse by the addition of pyrilamine, an H1 antago-nist, suggests the involvement of H1 receptors. In retinalarterioles denuded of endothelial cell function, thehistamine response was only slightly reduced. Pyrila-mine then had no effect, but cimetidine furtherdiminished the histamine response. In intact retinalarterioles the attenuating influence of cimetidine waspreserved in the presence of methylene blue, an inhibitorof endothelial cell function. This combination ofevidence suggests that in porcine retinal arterioles thehistamine-induced vasodilatation may be mediated byendothelial cell H1 receptors, and by H2 receptors on thesmooth muscle cells.

3.1.1.4. Acetylcholine. Acetylcholine has not only beenimplicated as a mediator of vascular tone but alsorecognized as an important pharmacological tool toassay the endothelium function.

The endothelium modulates smooth muscle cellactivity by releasing vasoactive substances such as


endothelium derived relaxing factor/nitric oxide, andpotent vasoconstrictor endothelin-1. Damaged or dys-functional endothelium therefore has an important rolein the pathology of vascular diseases.

Acetylcholine is routinely used to assess endotheliumfunction in blood vessels and is widely accepted as anendothelium dependent vasodilator. It is known to causeendothelium dependent vasodilatation by stimulatingthe release of nitric oxide from the endothelium. Thisendothelium dependent dilatation function has beenshown to be impaired in the vasculature of many organsin diseases such as hypertension and diabetes. Hyperten-sion and diabetes have associated ocular pathologies,such as hypertensive or diabetic retinopathy that hasspecific relevance to the vascular component of the eye.It is therefore suspected that endothelial dysfunctionalso occurs in the eye vasculature. Several studies(Bakken et al., 1995; Stowe et al., 1997; Su et al.,1994) on normal ocular circulations have been per-formed using acetylcholine. The responses varied withspecies studied but most studies demonstrated endothe-lium dependent vasodilation effects of acetylcholineadministration on isolated retinal vessels, isolated eyesand ring segments of ophthalmic arteries (Benedito et al.,1991b; Hoste and Andries, 1991). In vitro studies fromour laboratory on human retinal vessels (Yu et al., 1998)have also demonstrated vasodilatation to acetylcholineadministration. Furthermore, we have specifically ad-dressed the question of tone dependency in acetylcholineinduced relaxation response using isolated eye prepara-tion (Yu et al., 2000b). The issue of tone dependency isimportant in view of vascular disease situations as insystemic hypertension or glaucoma where perfusionpressure is altered. If tone dependency exists inacetylcholine-induced dilatation responses, then thishas implications for the ability of vessels to dilate toacetylcholine in disease states and also for the inter-pretation of acetylcholine studies. Our results show thatacetylcholine-induced relaxation responses are highlydependent on vascular tone in the rat ocular vascula-ture. Particularly, there is a strong linear relationshipbetween acethylcholine-induced vascular relaxation andprecontracted perfusion pressure at higher dosages ofacetylcholine. Therefore, when evaluating endothelialfunction using acetylcholine-induced relaxation, vascu-lar tone has to be counted as an important factor.

3.1.1.5. Dopamine. The dopaminergic system, like thoseof all other neurotransmitters, can effect alterations inOBF by a number of distinct mechanisms. One of themis to directly interact with specific receptors on thevasculature. Dopamine has potent contractile effects onmajor cerebral arteries in vitro. The maximal contrac-tion provoked by dopamine can exceed that producedby noradrenaline under the same conditions. It has beendemonstrated that the contractile effects of dopamine

are not the results of activation of specific dopaminereceptors, but results from the interaction of dopaminewith other neurotransmitter receptors, particular the a-adrenergic receptor and the 5-HT receptors. Dopaminehas much less potent contractile effects in the ocularvasculature than that in cerebral arteries. For example,contractile response of dopamine in cat ophthalmocili-ary arteries has only B50% of that of noradrenaline,and 30% of that of K+ induced contraction under thesame conditions (Yu et al., 1992c). Dopamine inducedcontraction in the human LPSA but these were lesspotent than seen in the cat ophthalmociliary artery (Yuet al., 1992b).

3.1.1.6. Prostanoids and thromboxanes. The prosta-noids and other arachidonate-mediated metabolitesrepresent a vast family of compounds. However, thesevarious eicosanoids do not share one common biologicaleffect, and in many systems, different eicosanoidspossess opposite actions. The various siblings of thearachidonic acid pathway rival each other in terms ofthe complexity and diversity of their effects.

There are at least two issues of direct interest to theocular vasculature. A large number of diverse eicosa-noids such as prostaglandin F2a (PGF2a) and prostacy-clin are known to be potent modulators of the ocularcirculation. Secondly, a number of eicosanoids havebeen used as therapeutic agents for glaucoma.

Prostaglandins are known to have vasoactive effectsin many organs. In ocular tissues, the vasoactive effectsof closely related members of the prostaglandin familycan vary significantly depending on the specific vesselsinvolved. PGF2a is known to have contractile effects onthe feeder vessels to the eye (Hoste, 1997; Ohkubo andChiba, 1987; Stjernschantz et al., 2000; Su et al., 1994,1995) and also in bovine retinal arteries (Hoste andAndries, 1991; Nielsen and Nyborg, 1989b, 1990).However, comparatively little is known about thevasoactive effect of PGF2a and other prostanoids onretinal arterioles. We have recently studied vasoactiveeffects of selected prostanoids on retinal arterioles (Yuet al., 2001b) and demonstrated that in normal tonearterioles without endothelin-1 contraction, PGF2a andthe thromboxane A2 analogue U46619 both produced apotent dose dependent contraction. In endothelin-1contracted retinal arterioles, U46619 produced furthercontraction, whereas PGF2a produced a slight vasodi-latation.

3.1.1.7. Endothelins. Endothelins are endogenous vaso-contracting peptide agents. Endothelin-1 (ET-1), ET-2,and ET-3 are produced in a variety of tissues, where theyact as modulators of important cell processes. Thesepeptides act through binding to two classes of trans-membrane receptors, ETA and ETB, where the stimula-tion of several signalling pathways leads to their


mitogenic, vasoconstriction, and developmental actions.ET-1 is very clearly an important vascular hormone.

ET-1 has been shown to be one of the most potentvasocontractors in the ocular vasculature (Meyer et al.,1993; Nyborg et al., 1991). Significant reduction in theblood flow in the retina, choroid and optic nerve headby exogenously administrated ET-1 has been reported inhumans and in a number of animal species (Dallingeret al., 2000; Granstam et al., 1992; Kiel, 2000; Polaket al., 2001; Sugiyama et al., 1996). Altered plasmaconcentration of ET-1 has been demonstrated in avariety of ocular diseases, such as retinal vein occlusion(Masaki and Yanagisawa, 1992), glaucoma (Kaiseret al., 1995; Liu et al., 1990), diabetic microangiopathy(Ak et al., 2001) and ocular microangiopathy syndromein patients with AIDS (Geier et al., 1995). Normalplasma level of ET-1 are about 1–8 pg/ml (0.4–3.2� 10�12M) (Iannaccone et al., 1998; Masaki andYanagisawa, 1992), and the values in these reporteddiseases are roughly twice that of normal subjects.However, 10�12–10�7M ET-1 induced potent contrac-tion of porcine ciliary artery with EC of 8.3 (Meyer et al.,1995b). ET-1 induced dose dependent vasocontractionin retinal arterioles with a similar range of dosage.Extraluminal application of ET-1 10�9M, BEC50 of thevasocontraction, produced a potent and stable vasocon-traction in the pig and human retinal arterioles (Yuet al., 1998, 2001b). More recently, we have found thatretinal arterioles exhibit asymmetry in their responses toET-1 with contractions significantly larger when thedrug was applied to the extraluminal surface rather thanthe intraluminal surface as shown in Fig. 8.

The paracrine mechanism of endothelin in vascularwall is shown in Fig. 9 (Masaki and Yanagisawa, 1992).Vasoactive factors stimulate endothelial cells to releaseET-1. Release of ET-1 stimulates ETA receptors onsmooth muscle and ETB on endothelial cell. Twofactors, prostacylcin and EDRF, are released, and acton smooth muscle cells as vasorelaxation factors. It islikely that ET-1 levels in the vicinity of the smoothmuscle cells could be more than 1 ng/ml, which is

significantly higher than seen in the plasma. As ET-1 islocally secreted, only that portion, a several-thousand-times smaller amount of ET-1, crossing back across theendothelium will enter the plasma, making plasmaconcentration of 1–2 pg/ml. Therefore, there is adiffusion barrier for ET-1, and an efficient regulatorysystem in the vascular wall, particularly in microvessels.

In general, data on levels of circulating ET-1 in bloodor plasma do not provide any valid information on localactivity of this system, nor do they allow association ofenhanced formation rates to particular cells or tissues.At best, levels of circulating ET-1 give some informationregarding the ‘‘overall’’ activity of the system.

3.1.1.8. Adenosine. Adenosine, along with the excita-tory amino acids glutamate and aspartate, is known tobe released from ischemic and hypoxic neural tissue(Rudolphi et al., 1992; Sciotti et al., 1997). Indeed,

Fig. 8. Comparison of the vasoactive effects of endothelin-1 when

applied intraluminally or extraluminally to an isolated perfused pig

retinal artery. Extraluminal application produces a greater vasocon-

traction than intraluminal delivery.

Fig. 9. Schematic of the paracrine mechanism in the vascular wall. Vasoactive factors stimulate endothelial cells to release ET-1 which stimulates

ETA receptors on the smooth muscle cells and ETB receptors on the endothelial cell. Modified from Masaki and Yanagisawa (1992).


elevation of aspartate and glutamate are both accom-panied by an increase in extracellular adenosine(Sciotti et al., 1997). The excitatory amino acids aredamaging to neural cells, whereas the concomitantrelease of adenosine has been shown to ameliorate thedamage. Indeed, extracellular concentrations of adeno-sine increase in ischemia in a graded fashion by factorsof greater than 10, with extracellular adenosineconcentration being related to the level of tissueischemia, a necessary criterion if it is to function as avasoactive signaller. In the retina, adenosine has beenshown to cause retinal artery dilatation after vitrealmicrosuffusion onto the vessels, acting through A2

receptors (Gidday and Park, 1993a, b). Experimentalincrease of endogenous adenosine was also accompaniedby dilatation.

We have demonstrated an asymmetrical response toexogenous adenosine in retinal arterioles in that extra-luminal administration of adenosine produces a dosedependent dilatation, whereas intraluminal adenosinefails to produce a significant dilatation response (Alderet al., 1996). The extraluminal adenosine caused a dosedependent dilatation which commenced at 10�6M, andreached a percentage dilatation of 23% at 10�3M. Forconcentrations of 10�4M and above spontaneousoscillations in diameter were observed for extralumin-ally applied adenosine. In contrast, intraluminal adeno-sine failed to cause dilatation or spontaneousoscillations at any concentration tested yet the dilatoryability of these vessels was confirmed by intraluminaland extraluminal application of the Ca2+ channelblocker verapamil. The two sides of the retinal arterywall are differentially sensitive to adenosine, with theintraluminal route being ineffective. In vivo, in hypoxicor ischemic situations, adenosine is released by extra-luminal neural tissue and minimizes tissue damage,partially by acting as a signaller of metabolic status tothe vasculature leading to vasodilatation and increasedlocal blood flow. Thus, adenosine is capable of bothsignalling metabolic needs and causing vasodilatation inthe retina.

Normally, extracellular levels of adenosine are about10�7M, which is close to the threshold values found inour and other studies for extraluminal application(Alder et al., 1996; Park and Gidday, 1990; Rudolphiet al., 1992). Adenosine probably acts as a neuromodu-lator as well as a vascular mediator in the retina. A1 andA2 receptors are also present in neural tissue. A1

receptors are found in ganglion cells, and their axonalprocesses in the optic nerve, as well as in the innerplexiform layer (Blazynski, 1990; Blazynski and Perez,1991). These receptors have a high affinity with aKmE10�9 M (Blazynski and Perez, 1991), so they aresensitive to nM concentrations of adenosine, well belowthe threshold we demonstrated for vascular responses.There are also A2 receptors in the retinal pigment

epithelium and photoreceptors (Blazynski, 1990) thatare sensitive to mM concentrations of adenosine.

3.1.1.9. Angiotensins. Angiotensin II is a potent con-tractile substance in larger arteries and arterioles. Theorder of sensitivity of most arterial smooth muscle toangiotensins is angiotensin IIbangiotensin IIIbangio-tensin I. Contractile responses to angiotensin I may bemediated by its conversion to angiotensin II by theaction of angiotensin converting enzyme (ACE). In-hibitors of ACE not only inhibit the formation ofAngiotensin II, but also increase the levels of bradykininwhich actives the nitric oxide pathway and reduces theformation of ET-1.

The presence of angiotensin II binding sites in bothbovine and human retinal vessels has been demonstratedby special radioligand binding studies (Ferrari-Dileoet al., 1991). High levels of ACE has been detected inbovine and human retinal and choroidal vessels(Ferrari-Dileo et al., 1988). There are diverse resultsfrom various preparations and different species. Forexample, bovine retinal arteries have been shown to beinsensitive to angiotensin II using ring segment prepara-tions (Nyborg et al., 1990), but dose dependentcontractile responses to angiotensin II have beendemonstrated in posterior ciliary arteries from bovine,porcine and human eyes (Meyer et al., 1995a; Nielsenand Nyborg, 1990; Nyborg and Nielsen, 1990) and inthe retinal artery in cats (Rockwood et al., 1987).Regional heterogeneity, contractile and dilatory re-sponses with angiotensin II administration have beenobserved in bovine retinal arteries and veins and likelymodulator roles of the vascular endothelium have beenproposed (Kulkarni et al., 1999). However, studies usingour isolated perfused rat eye preparation failed to find asignificant dose dependent contractile response toangiotensin II (Su et al., 1995). These results have tobe treated with caution. From our recent experimentsusing isolated perfused retinal vessels, not only the siteof action may play an important role but also theangiotensin II delivered intraluminally may be pre-vented from reaching the smooth muscle cells by theendothelium. In addition, the endothelium may play anactive modulating role on the effects of angiotensin II onvascular tone. It is expected that the renin–angiotensinsystem may play an important role in retinal physiologyand pathology in terms of the presence of all thecomponents required for the intra ocular generation ofangiotensin II (Murata et al., 1997; Wagner et al., 1996)and there is evidence that the concentration ofangiotensin II in the retina is 10 times higher than thatin plasma (Kohler et al., 1997). The effects ofangiotensin II in the ocular vasculature, particularly inthe retinal microvasculature clearly warrant furtherinvestigation.


3.1.1.10. Insulin. Dynamic alterations in blood insulinand glucose levels are hallmark features of both insulindependent and non-insulin dependent diabetes. Withinthe retina, the vasculature is the prime site of diabetesinduced changes (Cogan et al., 1961) with the earlystages of both clinical and experimental diabetesmellitus characterized by blood flow changes (Cringleet al., 1992; Grunwald et al., 1984b; Small et al., 1987;Tilton et al., 1989). In diabetes the initiating insult is toinsulin and glucose levels. Glucose-induced changes inretinal blood flow have already been reported (Ernestand Goldstick, 1983) but there have been very fewstudies describing the direct effect of insulin on theretinal vasculature except for those in which both insulinand glucose levels are changing (Grunwald et al., 1987),i.e. with no euglycaemic clamp. We raised the questionto what extent could some of these changes in bloodflow, oxygen tension and pharmacological response bemediated by insulin concentration.

Insulin is known to exert very diverse biologicaleffects. Recently evidence has accumulated that insulinhas a physiological role to play in controlling vascularactivity in some vessels (Baron, 1994). This is in additionto its more established roles as a glucoregulatoryhormone, a promoter of cell proliferation and metabo-lism, and a mediator of ion transport.

We tested the hypothesis that insulin dilates retinalarterioles by a direct mechanism. Because of thedifficulty of separating out direct and indirect effectsof insulin in the in vivo animal, we utilize the isolatedperfused retinal arteriole. Our results showed thatextraluminal delivery of insulin alone had no significanteffect on vessel diameter. Intraluminal delivery ofinsulin produces a dose dependent dilatation of 5.6%of the K+ contracted diameter at 200 mU/ml and up to12% by 2000 mU/ml, whereas combined intraluminaland extraluminal application of insulin causes dilatationat all concentrations rising to 15% at 200 mU/ml and19.7% at 2000 mU/ml. Intraluminal application ofindomethacin (5� 10�5M) had no significant effect onthe insulin induced dilatation whereas intraluminalapplication of L-NAME (10�4M) inhibited insulindilatation completely. Addition of extraluminal verapa-mil (10�6M) during insulin-induced dilatation resultedin further dilatation to 38%. However, addition ofinsulin to verapamil dilated vessels caused no furtherdilatation. Exposure to extraluminal application ofinsulin whilst measuring the intraluminal applicationof K+ contraction dose response curve had no effect.Results in main arteries and their branches did notdiffer. Our results indicate that intraluminal applicationof insulin dilates potassium contracted pig retinalarteries. This effect is enhanced by extraluminalapplication of insulin that does not result in a dilatationwhen administered alone. The dilatation response ismediated by nitric oxide but not by prostaglandins.

There is some evidence for the involvement of Ca2+

channels in the insulin-induced dilatation. These resultsimply that insulin is a vascular regulator in normalconditions and may have relevance to the vascularchanges occurring in diabetes and hypertension in theretina.

3.1.1.11. Calcium channel entry blockers. Both plasma-lemma and sarcoplasmic reticulum membranes establisha Ca2+ concentration gradient of about 10,000-fold.[Ca2+]i in the resting smooth muscle cell lies between120–270 nM and rises to 500–700 nM in the activatedsmooth muscle cell. It means that [Ca2+]i of smoothmuscle cells in the activated condition is only 3–4 timeshigher than that in the rest condition. The voltage-activated Ca2+ channels are often loosely referred to asCa2+ channels although three major Ca2+ controlpathways, voltage-activated Ca2+ channels, receptor-operated Ca2+ channels, and Na+-Ca2+ exchangers,are present. The biological role of l-type Ca2+ channelsis well established and most current calcium channelentry blockers act on l-type Ca2+ channels.

The three major subclasses of Ca2+ entry blockershave demonstrated important differences in theirinhibitory potency, when their action on cerebral vesselsis compared. The dihydropyridines are the most potentwith IC50 values usually around 10�8–10�9M, reachingeven 10�10M in some cases, whereas the phenylkyla-mines and benzothiazepines have significantly higherIC50 values. We have extensively studied on the effectsof calcium channel entry blockers. Our data fromhuman and pig retinal arteries are comparable withthese studies on the cerebral vessels (Yu et al., 1998).

The relative importance of the intracellular andextracellular Ca2+ sources in contributing to vasodila-tation has been shown to vary in different regions of thevasculature, as well as with different vasoactive agents.For example, we have demonstrated that the phasiccomponent of an a1-adrenergic contraction is mainlydependent on intracellular Ca2+ stores, whereas thetonic component relies almost exclusively on extracel-lular Ca2+ (Yu et al., 1992a). It is well recognized thatthe selectivity of Ca2+ entry blockers depends on thetissue (with greater selectivity in cerebral arteries whencompared with other peripheral arteries), the species, thevasoconstrictor agent used and the chemical type of theCa2+ entry blocker, which probably reflects differencesnot only in the quantity, but also in the quality, of theCa2+ channels activated in the different cases.

3.1.2. Other modulators and factors

It is important to understand the effects of theexperimental set up and the tissue bath environmentson isolated ocular vascular preparations. The para-meters used in isolated ocular vascular preparations canonly be a best guess mimic of the in vivo physiological


conditions. It is important to examine the influence ofeach parameter on the vasoactive responses in order toappreciate its potential importance. Critical parametersshould be closely monitored or controlled throughoutthe experiments. Deliberate manipulation of suchparameters outside the physiological range can be usefulfor creating pathological conditions of relevance to thepathogenesis of particular diseases.

3.1.2.1. Oxygen tension. Oxygen tension in vivo isknown to play an important role in regulating retinalblood flow (Grunwald et al., 1984b). Accuratelycontrolling the oxygen tension in isolated preparationsrequires a close attention to uptake or losses of oxygenby the perfusate media. Our approach was to maintainthe preparations at a fixed oxygen tension governed bythe oxygen content of the gas phase above theincubation chambers (95% O2, 5% CO2). In an earlierstudy in ring segments of the ophthalmociliary artery weinvestigated the modulating role of oxygen tension onnoradrenaline and KCl induced responses in theophthalmociliary artery (Alder et al., 1993). Our resultsindicated that endothelial cells modify the intrinsicsmooth muscle response to a gradual reduction in PO2

by releasing relaxing and contracting factors causing theobserved dichotomous response in NA activated vessels.However the KCl induced response is only modulatedby low oxygen tensions.

An unknown retinal-derived relaxing factor whichplays a role in regulation of retinal arterial tone has beendemonstrated using isolated arterial ring segmentpreparations with and without adhering retinal tissue(Delaey and van de Voorde, 1998). Adherent retinaltissue was shown to mediate the hypoxia-inducedvasodilatation of bovine retinal arteries in vitro. NeitherNO, prostanoids, adenosine, excitatory amino acids,lactate or pH changes seem to be involved in thehypoxia-induced vasodilatation (Delaey et al., 2000).

3.1.2.2. Extracellular pH. We have studied the effectsof changes in extracellular pH (pHe) on passive tone andagonist responses in the cat ophthalmociliary artery andmediator roles of the endothelial cells in any pH inducedeffect in order to explore the ability of the ophthalmo-ciliary artery to influence retinal and choroidal bloodflow in response to metabolic stimuli (Su et al., 1994).

Our results show that PGF2a produces a concentra-tion dependent contraction that is insensitive to analkaline shift from control pHe (7.4) in the bathingmedium. For acidic shifts to pHe 7.0, there is nosignificant change in the magnitude of the PGF2a

contraction, whereas at pHe 6.0 the PGF2a contractionis reduced to 23% of its value at pHe 7.4. Thresholdresponse concentration remains unaffected. Deliberatedamage to the endothelial cells does not significantlyaffect the magnitude of the 10�5M PGF2a response at

pHe 7.4 nor the effect of acidic pH on this response. The10�5M NA response is reduced in a graded fashion toacidic shifts to pH 7.0 and 6.0 (40%) and also to alkalineshift to pH 8.0 (22%) when compared to the inducedtension at pHe 7.4. For the acidic shift only, endothelialcell damage causes a further significant reduction in theNA response to 20%. For vessels contracted with K+

Krebs solution there is a small but significant reductionin response at pHe 6.0 to 84%, whereas for pHe 8.0there is a much larger reduction to 45%. All pHeinduced relaxations of K+ are endothelium indepen-dent. Passive tension is unaffected by all pHe manipula-tions.

Conclusions: vessel responses to vasoactive agents areselectively mediated by pHe changes. Major acidic shiftscause reduced responses (relaxation) to NA, PGF2a, orK+, whereas only vessels preactivated with NA and K+

relax to alkaline shifts. This implies that NA or K+

induced vascular responses are maximal close to neutralpHe with major shifts from neutrality in either the acidicor alkaline direction causing a reduced response. Theseresults imply that the ophthalmociliary artery probablydoes not play a major role in controlling OBF inresponse to pHe changes within the normal metabolicrange, but may become important in ischemic condi-tions.

3.1.2.3. Tension and perfusion pressure. It is critical toset up a suitable pretension for myograph ring segmentof ophthalmic artery or LPCA preparation, andperfusion pressure for isolated perfused vessels andisolated perfused eye preparations.

To determine the effects of pretension for themyograph ring segment of ophthalmic artery or LPCApreparation on vascular responses to a given vasoactivesubstance, length tension curve has been studied (Yuet al., 1990, 1992c). The form and magnitude of K+-Krebs induced contraction has been measured for thedog ophthalmociliary artery. Ring segments of the vesselwere mounted in a myograph. For low passive tensionsthe K+ induced contractions show both a phasic andtonic component with the tonic tension componentbeing the smaller. As passive tension is increased thephasic component grows and plateaus whilst the toniccomponent continues to increase. For high passivetensions the phasic component disappears and the toniccomponent reduces in magnitude and the contraction isnot fully reversible. Comparison of the length tensioncurves measured during K+-Krebs induced contractionsand Ca2+ free Krebs induced relaxation shows that theactive tension generated by this vessel increases withincreasing passive tension until a peak active tension of3.1mN/mm is reached at an effective radius of 357 mm.For greater values of passive tension the active tensiondecreases. The calculated transmural pressure requiredto maintain the vessel at this optimal radius is


51.9mmHg, which is very similar to the only availablemeasurements of ophthalmic artery pressure, viz51.8mmHg (Riva et al., 1981).

Tone or pressure dependency in vasoactive substanceinduced vascular responses is also present in the intactocular vascular preparation. As mentioned we havedemonstrated perfusion pressure dependency in acet-ylcholine induced relaxation response using isolated eyepreparation (Yu et al., 2000b).

3.1.2.4. Pharmacological and mechanical heterogeneity

of the ocular vasculature. The potential for the sameagent to have different vasoactive effects on differentregions of the ocular vasculature has been welldocumented. Our own studies demonstrating thisphenomenon include work in different segments of thecat LPCA, different branching types of the retinalcirculation in pig, and differential intraluminal andextraluminal responses in both the pig and humanretinal arteries.

Ophthalmociliary artery. The mechanical, histologicaland pharmacological properties of the isolated catophthalmociliary artery are reported and compared forring segments from three locations on the artery withinthe orbit. The vasoactive properties of different sectionsof the cat ophthalmociliary artery were assessed. Theophthalmocillary artery was divided into proximal(POA), middle (MOA), and distal (DOA) segments(Fig. 10).

The most distal segment (DOA), closest to the eye,developed the smallest maximal tension whereas themost proximal segment (POA) had the greatest lumendiameter. All segments had a phasic and tonic compo-nent in response to passive stretching and duringactivation with K+-Krebs and agonists. Cumulativedose response curves were measured for nine agonists:histamine (HIS), 5-hydroxytryptamine (5-HT), dopa-mine (DOPA), adrenaline (A), noradrenaline (NA),tyramine (TYR), phenylephrine (PHE), isoproteronol(ISOP) and xylazine (XYL) and the maximum activetension developed at a concentration of 10�3M com-pared with that for 0.124mM K+-Krebs. In the threesegments the agonist response order was:

POA HISbKmaxbNA > A > DOPA > PH

¼ TYR > 5-HT ¼ ISOP > XYL;

MOA KmaxbNAbA ¼ HISbDOPA > ISOP

¼ PHE > TYR ¼ 5-HT > XYL;

DOA KmaxbNAbA > HIS ¼ DOPA > ISOP > PHE

¼ 5-HT > XYL ¼ TYR:

Although no antagonists were used it is concluded thatall segments possess a rich concentration of functionalHIS and a1 receptors, a relatively poor concentration ofb and 5-HT receptors, and no a2 receptors. A significantdifference in the response to HIS between the POA and

the MOA suggests different sections of the retrobulbarophthalmociliary artery are heterogeneous in receptorbalance, with the most proximal and larger segmenthaving a greater concentration of histamine receptors.The responses to A and NA were found to be a functionof passive tension, whereas the response to histaminewas independent of passive tension.

Although the diameter of these sections of theophthalmociliary artery is similar, they displayed a verydifferent response to mechanical stretch in the myo-graph system. The DOA showed only about one third ofthe active vasoactive tension of the POA and MOAwhen each were subjected to K+-Krebs at increasinglevels of passive stretch. Pharmacological heterogeneitywas also evident in the differing response to histamine.Compared to the magnitude of a K+-Krebs inducedcontraction, 10�6M histamine induced contractions of137% in the POA, 55% in the MOA and 41% in theDOA. Thus indicating a decreased vasoactive effect as afunction of distance down the ophthalmociliary artery.

Different types of retinal branch. The branchingpattern of retinal arteries has two distinct patterns,which we can readily distinguish by the angle at whichthe branch emanates from the main vessel (901- and Y-branches). Studying isolated retinal vessels with both

Fig. 10. Schematic illustration of the ophthalmociliary artery in the

cat and the three different regions studied for vasoactive properties.

Different sections of the same vessel can have markedly different

vasoactive responses to the same agent (Yu et al., 1992c).


types of branches allows the vasoactive properties of thedifferent branch types to be compared. This may beparticularly important, since we also have evidence thatthe 901 branches tend to feed the capillary layers, whilstthe ‘‘Y’’ branches remained in the superficial retina. Thismay well be an important component in the regulatorycontrol of retinal blood flow that we have previouslydescribed (Yu et al., 1994b).

We found that the proximal region of the 901-branches of isolated porcine retinal arterioles demon-strated a localized vasoconstriction to intraluminalapplication of 124mM potassium when compared tothe more distal region (Fig. 11). In contrast, the Y-branches displayed a more even constriction along thelength of the bifurcation (Fig. 12). The localized activitynear the branch point of 901 type branches may explainprevious observations of so called sphincter-like activityin these vessels which has been difficult to demonstratemorphologically.

3.2. Evaluation of vasoactive properties of anti-glaucoma

drugs

3.2.1. General

Although lowering of IOP has been the mainstay ofglaucoma treatment for many years, more scientificevidence emerged recently that vascular factors areprobably also involved in the pathogenesis of glauco-matous optic neuropathy (GON) (Flammer, 1994).Normal-tension glaucoma (NTG) or high-tension glau-

coma patients whose disease progresses despite normalor normalized IOP have been shown, for example, tohave slower blood flow velocity in the retina (Wolf et al.,1993), in the choroid (Duijm et al., 1997; Prunte andFlammer, 1989) and in the optic nerve head (Michelsonet al., 1996). Increased plasma levels of endothelin 1(ET-1), the most potent physiologic vasoconstrictorpresently known (Flammer, 1996; Liu et al., 1990), havealso been reported in NTG patients (Kaiser et al., 1995;Sugiyama et al., 1995).

Since the causative factors for the impaired OBF inglaucomatous eyes are not well defined, evidence of thehaemodynamic effects of existing IOP-lowering medica-tions is still contradictory. While some investigatorshave suggested that certain topical IOP-lowering med-ications might have a beneficial effect on the OBF withan unclear mechanism and clinical relevance (Harriset al., 2000; Lachkar et al., 1998), others have focused onthe importance of the lack of a detrimental effects oftopical IOP-lowering drugs on the ocular circulation(McKibbin and Menage, 1999; Pillunat et al., 1999;Seong et al., 1999). In glaucoma therapy, any vasoactiveeffects on the retinal vasculature may be particularlyrelevant, given the increasing acceptance of an ischemiccomponent to the pathophysiology of glaucoma(Drance, 1997; Flammer, 1996). Thus, it is clear thatidentifying and abolishing the etiological factors forimpaired OBF in glaucomatous eyes may lead to

Fig. 11. Normalized vessel diameter changes of the proximal and the

distal regions of 901-branches as a function of time in vessels from 11

different eyes when 124mM potassium was applied intraluminally. In

the proximal region, the mean baseline diameter was 49.171.1 mm and

the average percentage constriction in the proximal region was

15.571.9%. In the distal region, the mean baseline diameter was

51.171.1 mm and the average percentage contraction was 1.371.6%.

This was significantly less than that in the proximal region (po0:0001).

Fig. 12. Normalized vessel diameter changes of the proximal regions

and the distal regions of Y-type branches as a function of time in

vessels from fourteen different eyes when 124mM potassium was

applied intraluminally. In the proximal region, the mean baseline

diameter was 79.671.3 mm, and the mean contraction was

18.371.2%. In the distal region, the mean baseline diameter was

84.771.4mm, and the mean contraction was 16.371.7%, which was

not statistically different from that in the proximal region of these Y

type branches (p > 0:4).


slowing down of the disease progress and ultimately topreservation of the visual field.

There are several lines of new evidence to explain whythe retina and the optic nerve are particularly vulnerableto ischemic insult. The energy demands of the visualprocess are high, and much of the energy is derived fromoxidative metabolism coupled to ATP synthesis. Theoxygen consumption of the retina on a per gram basishas been described as higher than that of the brain(Ames, 1992; Anderson and Saltzman, 1964). Giventhat the brain consumes a highly disproportionateshare of the total body oxygen uptake (Coyle andPuttfarcken, 1993), this places the retina as one of thehighest oxygen consuming tissues in the body (Ander-son, 1968). Since oxygen cannot be ‘‘stored’’ in tissue, aconstant and adequate supply must be guaranteed inorder to preserve function (Vanderkooi et al., 1991).Oxygen supply to the retina is arguably more vulnerableto vascular deficiencies than any other organ. Therequirement for a relatively unobstructed light path tothe photoreceptors presumably places a constraint onthe degree to which the retina can be vascularized. Thisresults in a very delicate balance between the availableoxygen supply and the consumption of oxygen withinthe retina.

In contrast to the brain, the retina has a highly layeredstructure in which the different cell types and thesupporting vascular components are spatially separated.The likelihood of differing oxygen requirements ofdifferent cell types and their relationship to the nearestoxygen source makes it difficult to predict the oxygenenvironment in any particular retinal layer. Our currentunderstanding of the oxygen requirements of the innerand outer retina is that there are three dominant oxygenconsuming layers in the rat retina. These are the innersegments of the photoreceptors, the outer plexiformlayer, and the deeper region of the inner plexiform layer(Yu et al., 1994b, 1999b; 2000a; Yu and Cringle, 2001).Oxygen supply to the plexiform layers is heavilydependent on retinal circulation, particularly the deepcapillary layer. Even the inner segments of the photo-receptors are partially supplied by the deep capillaryduring dark adaptation (Cringle and Yu, 2002; Yu andCringle, 2002).

We have demonstrated that there is high tissuepressure gradient across the lamina cribrosa (Morganet al., 1995, 1998), that requires considerable energy tosupport axonal transport indicated by localized dis-tribution of abundant mitochondria.

These findings illustrate that particular regions of theretina and optic nerve, those with high metabolic ratesand relatively spare vascular supply are more vulnerableto ischemic insults than may have been anticipated. Asischemic insult is likely to be a critical pathogenic factorin glaucoma, ideally, any drugs used for glaucomatreatment should lower IOP but have either no direct

vascular effect, or an effect which improves ocularperfusion.

3.2.2. Effect of b-adrenergic antagonists and Ca2+

channel entry blocker on ocular vasculature

The b-adrenergic antagonists betaxolol and timololare commonly used antiglaucoma agents, which arethought to lower IOP, the greatest risk factor forglaucoma, by reducing the rate of aqueous humorformation. Aqueous formation is influenced by b-adrenergic mechanisms in the ciliary epithelium (Bru-baker, 1991; Collignon-Brach, 1992; Frishman et al.,1994; Neufeld and Bartels, 1982). It has long beendebated whether these agents also act directly on theretinal circulation to change retinal blood flow.Although the influence of b-adrenergic antagonists onretinal artery diameter has been studied in animals, in-vivo and in a clinical setting, the results remaincontroversial with some reports suggesting that someb- and the b1-selective adrenergic antagonists decreasedOBF (Chiou and Chen, 1993; Yan and Chiou, 1987),whereas other studies have indicated increased bloodflow (Grunwald, 1986). As mentioned before, much ofthe ocular vasculature is devoid of b-adrenergic recep-tors, or that they are only weakly functional (Hosteet al., 1989, 1990; Nielsen and Nyborg, 1989b, a; Nyborgand Nielsen, 1995; Su et al., 1995; Yu et al., 1999a,1992c, b), which would suggest that only a minimal b-adrenergic vascular effect would be expected.

Topical application of both timolol and betaxolol hasbeen shown to produce a statistically significant decreasein IOP in humans, although the IOP decrease wasgreater for timolol (Collignon-Brach, 1992). However,there is evidence that betaxolol also improves visual fieldperformance through as yet unknown mechanisms(Collignon-Brach, 1992; Messmer et al., 1991). Onepossible factor in this improved visual outcome whichhas been suggested, is that betaxolol may improveretinal blood flow as well as reduce IOP. Betaxolol hasbeen shown to have vasodilatory properties in pigLPCA and bovine retinal arteries using ring segmentpreparations (Hester et al., 1994; Hoste, 1999; 3Hosteand Sys, 1998). We were interested in the underlyingmechanism of vasodilatation induced by b-adrenergicantagonists and in determining whether such vasodila-tation also occurs in human retinal arterioles, which aremuch smaller (B100 mm) than bovine retinal arteries.

3.2.2.1. Effect of b-adrenergic antagonists on pig ciliary

artery and possible mechanisms. To define the under-lying mechanisms of the vasodilatory properties of b-adrenergic antagonists, we used the ring segmentpreparation system and dual wavelength fluorescenceexcitation system to follow intracellular calcium changes(Yu et al., 1997). The force generated by the ringsegment was also monitored in real time. Potassium was


administered into the bath to produce a concentrationof 60mM. Once equilibrium was attained, increasingdoses of betaxolol or timolol were applied to the bathand the force and calcium ratio monitored throughout.

The effect of betaxolol administration on the Fura-2ratio and the contractile force in K+ contracted ringsegments of LPCA is shown in Fig. 13A. The data isnormalized to the ratio and force measured during theK+ application. Note the close correlation between thecontractile force and the [Ca2+]i level at increasing dosesof betaxolol. A similar response to timolol was alsofound (Fig. 13B), although the concentration of timololrequired to produce the same level of relaxation wassignificantly higher. A direct comparison between thenormalized contractile force at increasing doses ofbetaxolol and timolol is shown in Fig. 13C.

In this study we demonstrated that b-adrenergicantagonists, such as betaxolol and timolol, can alsoact as Ca2+ channel entry blockers in the pig LPCAwhere [Ca2+]i was shown to decrease in a dosedependent way inversely with the magnitude of therelaxation generated by these b-adrenergic antagonists(Yu et al., 1997). This conclusion is supported by studiesin other vascular beds where betaxolol has beenconfirmed to produce vasodilatation though a Ca2+

channel blocking mechanism independently of its b1-selective adrenergic antagonist action (Bessho et al.,1990).

3.2.2.2. Effect of b-adrenergic antagonists and Ca2+

channel entry blocker on human and pig retinal arter-

ioles. Isolated perfused retinal arteries of pig andhuman were used to compare the potency of vasoactiv-ity of betaxolol and with l-type Ca2+ entry blockersand timolol (Yu et al., 1997, 1998).

We have demonstrated that betaxolol produced adose dependent dilatation of K+ contracted retinalarterioles, with a threshold of 10�12M. The effect wasmore pronounced with intraluminal rather than extra-luminal delivery (Fig. 14). Betaxolol also dilated ET-1contracted retinal arterioles in a dose dependent mannerabove 10�10M. Similar effects were observed withverapamil and diltiazem (Fig. 15). Nimodipine wassignificantly more potent, whilst timolol was less potentthan betaxolol.

Nimodipine, a dihydropyridine, was chosen after arange of Ca2+ entry blockers (nimodipine, nifedipine,verapamil, diltiazem) were screened for vasodilatoryactivity in the pig retinal arteriole, which demonstratedthat nimodipine was the most potent (Yu et al., 1997).ET-1 was chosen as the precontracting agent as it hasbeen shown to exert its contraction effect in part byincreasing influx of extracellular Ca2+ through dihy-dropyridine sensitive (l-type) voltage operated Ca2+

channels (Randall, 1991). Certainly ET-1 is a verypotent vasoconstrictor in the retinal circulation and

increases in plasma ET-1 which would cause ocularvascular constriction, have been linked to NTG. Thismight implicate potent l-type Ca2+ channel entryblockers as potential therapeutic agents for NTG(Sugiyama and Goldman, 1995).

Because of the infrequent and unpredictable nature ofaccess to human donor tissue for this study, wedeveloped a specific protocol for the optimal cryopre-servation of human retinal arterioles, comparing freshand cryopreserved pig tissue using several different

Fig. 13. Force and fluorescence ratio in ring segment preparations

administered increasing doses of betaxolol and timolol. Betaxolol was

more potent than timolol at relaxing the ring segments (Yu et al.,

1998).


cryopreservation procedures. Examples of fresh andcryopreserved human retinal arteries are shown inFig. 16 (Su et al., 1998; Yu et al., 1997).

We tested the hypothesis that the b-adrenergicantagonists betaxolol and timolol, cause retinal arter-iolar vasodilatation in addition to their ability to reduceIOP, and compared their vasodilatory ability with thatof a known Ca2+ channel entry blocker nimodipine

in donor human and pig isolated perfused retinalarterioles.

Precontraction with endothelin-1 (ET-1) reduced thediameters to 74% for human retinal arterioles and 72%for pig retinal arterioles respectively. Nimodipine andbetaxolol caused a significant dose dependent dilatation,both with the same thresholds of 10�12M in humanvessels and 10�10M in pig vessels. Timolol caused asmall but significant dose dependent dilatation with athreshold of 10�12M in pig vessels, but was ineffective inhuman vessels. The nimodipine and betaxolol doseresponse curves were not significantly different inhuman vessels, but for pig vessels, nimodipine producedsignificantly greater dilatation than betaxolol. Bothnimodipine and betaxolol were significantly moreeffective vasodilators than timolol in both human andpig vessels (Fig. 17).

Fig. 15. Normalised vessel diameter during administration of either,

betaxolol (n ¼ 13), timolol (n ¼ 13), diltiazem (n ¼ 12), nimodipine

(n ¼ 15), or verapamil (n ¼ 13) in ET-1 contracted retinal arteries.

Each agent produced a dose dependent dilatation, with nimodipine

being significantly more potent (po0:05) than betaxolol, diltiazem,

and verapamil, which in turn were more potent than timolol (po0:05)(Yu et al., 1998).

Fig. 16. Frame grabbed image of two isolated perfused human retinal

arterioles: (A) fresh tissue, (B) cryopreserved tissue. The tips of the

inner (cannulation) and outer (suction) pipettes of the left-hand

cannulation system can be seen (B). Images were grabbed at slightly

different focus levels along the vessels so that in some areas the focal

plane is located in the lumen of the vessels (left-hand side (A) and right

hand side (B)). In these locations the vessel wall structure was clearly

defined, and showed that the endothelial cells (EC) lined up as an

irregular dark line on the inside of the smooth muscle cells. At other

locations the focal plane is located at the bottom of the vessel, so that

the endothelial cell nuclei (ECN) can be seen longitudinally arranged

along the vessel wall (right-hand side (A) and left-hand side (B)) (Yu

et al., 1998).

Fig. 14. Dose response curve for intraluminally (filled circles, n ¼ 16)

and extraluminally (empty circles, n ¼ 15) delivered betaxolol in retinal

vessels preconstricted with 124mM K+. Continuous perfusion with

K+ Krebs produces a sustained contraction which is relaxed in a dose

dependent manner by intraluminal application of betaxolol. The data

is normalized to the diameter prior to K+ contraction and expressed as

a percentage. The data points marked (*) are significantly dilated

(po0:05) when compared to the K+ contracted diameter in the

absence of betaxolol (Yu et al., 1998).


3.2.3. Vasoactive effects of a docosanoid and selected

prostanoids on retinal arterioles

Prostaglandin (PG) analogues, such as latanoprost,and docosanoids, such as unoprostone isopropyl (Res-culas) have been developed as an IOP-lowering agentfor the clinical management of glaucoma.

Unoprostone isopropyl has been reported to increaseOBF both in animals (Ogo, 1996; Sugiyama and Azuma,1995) and in humans (Nishimura and Okamoto, 1998).We wanted to elucidate the mechanisms by which it

exerts its beneficial effects on OBF. We compared thevasoactive properties of the docosanoid unoprostone, itsfree acid and different members of the prostanoid familyon isolated perfused pig retinal arterioles to assess theirpotential to modulate retinal blood flow (Yu et al.,2001b).

In normal tone arterioles without endothelin-1 con-traction, PGF2a and U46619 both produced a potentdose dependent contraction (Fig. 18), but neitherunoprostone isopropyl, nor unoprostone free acid hada significant vasoactive effect (Fig. 19). In endothelin-1contracted arterioles, U46619 produced further contrac-tion, PGF2a produced a slight vasodilatation (Fig. 18),whereas unoprostone isopropyl and its free acidproduced a pronounced dilatation (Fig. 19).

Of the agents tested, unoprostone isopropyl and itsfree acid were the most potent vasodilators of endothe-lin-1 contracted pig retinal arterioles. Members of theprostanoid family demonstrated a different effect on thediameter of isolated retinal arterioles as compared to thedocosanoids. The potential therefore exists for thedocosanoid unoprostone to have a beneficial effect onretinal blood flow in addition to any reduction in IOP.

3.3. Applications in the pathogenesis of diabetic

retinopathy

3.3.1. General

Diabetic retinopathy is the leading cause of newblindness in the working age population. Althoughtherapies such as retinal photocoagulation and vitrect-omy can be remarkably effective when administered atan appropriate stage in the disease process, there is aneed for further investigation of the pathogenesis ofdiabetic retinopathy, particularly in the earlier stagesbefore the appearance of severe pathological damageand visual impairment. If improved treatment regimensare to be developed it is crucial that the underlyingpathophysiological mechanisms responsible for diabeticretinopathy are better understood. The multifactorialnature of the many pathways implicated in diabeticretinopathy requires a very detailed approach toelucidate the key mechanisms involved and theirinteractions in order to develop logical strategies aimedat therapeutic intervention. There are many candidatemetabolic and chemical pathways implicated in diabeticretinopathy. These include advanced glycation end-products (AGE), aldose reductase, oxidative stress,tissue hypoxia, carbonyl stress, substrate stress, alteredlipoprotein metabolism, increased PKC activity, andaltered growth factor or cytokine activity. It is alsolikely that there could be significant interaction betweenthese pathways. A schematic diagram summarizing thissituation is shown in Fig. 20 (Yu et al., 2001a).Hemodynamic and vascular cell changes play a criticalrole in the pathogenesis of diabetic complications.

Fig. 17. Direct comparison between the vasodilatory effects of

nimodipine (open circles), betaxolol (filled circles) and timolol (filled

triangles) on ET-1 (10�9M) contracted pig (A) and Human (B) retinal

arterioles. Diameters are expressed as percentage of the uncontracted

baseline (B), and the abscissa is the Log ½ðMÞ� concentration of the test

vasodilator. Nimodipine, betaxolol and timolol all induced significant

dose dependent dilatation but that due to nimodipine and betaxolol

were significantly greater than with timolol (Yu et al., 1998).


Fig. 18. Vasoactive response of isolated perfused pig retinal arteries to

intraluminal or extraluminal application of PGF2a or U46619 in

uncontracted vessels or vessels contacted with 10�9M ET-1. Both

PGF2a and U46619 contacted the previously uncontracted vessels.

PGF2a dilated ET-1 contracted arteries but U46619 produced a further

contraction (Yu et al., 2001b).

Fig. 19. Vasoactive response of isolated perfused pig retinal arteries to

intraluminal or extraluminal application of unoprostpone isopropyl or

it’s free acid in uncontracted vessels or vessels contracted with 10�9M

ET-1. Neither unoprostone isopropyl or it is free acid affected

uncontracted vessels. Both unoprostone isopropyl and it is free acid

dilated ET-1 contracted arteries (Yu et al., 2001b).


Diabetic retinopathy is characterized by graduallyprogressive alterations in the retinal microvasculature,leading to three fundamental morbidities: (1) vascularhyper permeability, (2) vascular occlusion, and (3)neovascularization. Isolated ocular vascular prepara-tions have significant advantages in diabetic retinopathyresearch because they provide the possibility to look atthe individual pathways and their effects on the ocularmicrovasculature. The areas that we have focused on arethe direct effects of insulin on the retinal arterioles,alterations of agonist induced vascular responses indiabetic ocular vasculature, and the roles of theendothelium in diabetic ocular vasculature.

Unravelling the most critical mechanisms is unlikelyto be achieved in studies that are limited to the clinicallyobservable changes to the retina that are seen in humandiabetic retinopathy. Far more detailed and invasivestudies are required, preferably in a readily availableanimal model. In our study of long-term diabetes in therat (Su et al., 2000) the development of capillary fallout,pericyte loss, ghost vessel formation and microaneur-ysms paralleled many aspects of the human pathology.In this large group of single dose STZ treated rats thepathology progressed despite the spontaneous recoveryof normoglycaemia after about 40 weeks. Fig. 21A andB shows examples of retinal microaneurysms in the deepretinal capillary layer in both a diabetic rat after 111weeks of STZ induced diabetes, and in a human donoreye from a 51 year old diabetic patient (Yu et al., 2001a).The final and most debilitating stage of diabeticretinopathy is the proliferative stage. Retinal vascularproliferation is a rare occurrence in conventional animalmodels of diabetes. However, we have observed severalexamples of vascular proliferation in our longer termdiabetic rats. Fig. 21C shows the development of newvessels in the disk region of an STZ rat after 111 weeksof diabetes.

We review some of our recent experimental work inthe STZ rat compare our findings to the human

pathology, and outline potential new avenues fortherapeutic intervention. In particular, our improvedunderstanding of which layers of the inner retina havethe most stringent metabolic demands has helpedidentify which retinal layers are most susceptible tometabolic or hypoxic/ischemic insult. We conclude thatimproved treatment outcomes may ensue if the therapy

Fig. 20. Schematic representation of the possible pathways by which

hyperglycaemia leads to diabetic retinopathy (Yu et al., 2001a).

Fig. 21. (A) Retinal microvasculature (trypsin digested specimen)

showing a typical microaneurysm and ghost vessel formation in a

diabetic rat 111 weeks after STZ induction. (B) Retinal microvascu-

lature (trypsin digested specimen) showing a typical microaneurysm in

the retina from a 51 year old diabetic patient. (C) Retinal

microvasculature (trypsin digested specimen) showing proliferative

preretinal neovascularization (curved arrows) in the retina and disk

region in a diabetic rat 111 weeks after STZ induction. (Yu et al.,

2001a).


is targeted at the appropriate tissue at specific stages ofthe disease.

3.3.2. Agonist-induced vasoactivity

There are numerous reports showing that onefunctional consequence of diabetic retinopathy isalteration to retinal blood flow and also to theautoregulatory power of the retinal circulation (Cringleet al., 1993; Konno et al., 1996; Lanigan et al., 1990;Kohner, 1976) To date, despite some speculation, thereis no evidence for accompanying changes to thechoroidal circulation. However, there is some contro-versy regarding the magnitude and direction of thechanges.

Recent studies focussed on the initial stages ofdiabetic retinopathy have demonstrated early increasesin blood flow values in the retinal circulation as well aschanges to ocular oxygen distribution within B5 weeksafter the onset of diabetes in rats as well as an increase inheterogeneity of retinal tissue blood flow (Cringle et al.,1992; Tilton et al., 1989). A greater heterogeneity intissue blood flow values in the retina implies aredistribution of blood flow. Tissue blood flowdistribution normally provides a tight match betweenblood supply, tissue demands and the satisfactoryremoval of the waste products of metabolism. It is adynamic, complex and hierarchical task. The observedredistribution of blood flow so early in diabetes, wellbefore histological changes to the vasculature areobserved, may merely be a consequence of diabetes-induced heterogeneity of tissue demand. However,an alternative explanation is that the normal mechan-isms of blood flow control are disturbed, possiblyleading to a mismatch between tissue demands andblood supply.

The aim was to compare the pharmacologicalresponses of ocular vessels in control and age-matcheddiabetic animals early in the disease. In vivo pharma-cological experiments on the ocular vasculature areconfounded by the systemic effects of the injectedvasoactive agents as well as by the uncontrollable andunknown binding of the vasoactive agents to bloodproteins, so that effective concentrations available to thevasculature are variable. Moreover, there may beproblems with recirculation of the vasoactive agents.All these disadvantages are overcome if an isolatedperfused eye preparation is used.

In the STZ diabetic rat, blood glucose levels areelevated and insulin levels are reduced so that theendothelial and smooth muscle cells of the vessel lumenare bathed in an abnormal milieu, which is thought toaffect endothelial cells which play a vital role inregulating and mediating many vasoactive responses.As we are interested in a direct comparison of thefunctioning of the endothelial cells and smooth musclecells of the ocular vasculature in the same milieu, both

control and diabetic isolated perfused eyes wereperfused with a normoglycaemic medium during theexperiment. This answers the question whether or notocular vascular responses in the same environment areaffected at this stage of diabetes.

The effect of streptozotocin-induced diabetes onocular vascular resistance responses to noradrenalin(NA), adrenalin (A), phenylephrine (PHE), isoproter-enol (ISOP), prostaglandin F2a (PGF2a), 5-hydroxy-tryptamine (5-HT) and angiotensin II (ANG II), wasdetermined using isolated arterially perfused rat eyesfrom control and diabetic (4 week) rats (Su et al., 1995).In control eyes, NA, A, PHE, PGF2a, and 5-HT allproduced dose-dependent increases in total vascularresistance, with the following order of potency:NA=A > 5-HT > PHE=PGF2a at 10�4M. Theocular circulation was not sensitive to isoproterenol andangiotensin II. In diabetic eyes, responses to NA, A,PGF2a and 5-HT were altered. Diabetic responses toNA and A had lower thresholds with larger resistanceincreases at low concentrations. However, the rate ofincrease in resistance with concentration was moregradual in diabetic eyes so that at 10�4M controlresponses were larger. Diabetic resistance responses toPGF2a had the same threshold as in control eyes, butwere greater in magnitude with an earlier peak at10�4M. In contrast diabetic resistance responses to 5-HT were reduced, peaked at a lower resistance at10�4M, but had the same threshold as those in thecontrol eye. Basal vascular resistances in control anddiabetic eyes were not significantly different.

Vasoactivity in the early diabetic eye is disturbed withthe effective balance between different agonists alteredin favour of catecholamines at physiological concentra-tions. This may be related to the early changes in bloodflow and oxygen distribution already reported in the rateye, as well as changes to autonomic function. Theisolated perfused rat eye is a valuable technique forinvestigating such vascular reactivity in animal modelsof retinal disease.

3.3.3. Endothelium dysfunction

Alteration in endothelium-dependent vasodilatatoryfunction of the vascular endothelium has been impli-cated in diabetic vascular diseases including diabeticretinopathy. We assessed the vasoactive properties ofthe intact vascular system of the streptozotocin induceddiabetic rat model. Rats were sacrificed at 5–8 weeks,22–28 weeks, or 85–116 weeks post-induction, and thevascular resistance of the eye was assessed in an isolatedperfused eye preparation. The vasoactive response toacetylcholine (Ach, endothelium-dependent vasodilator)and sodium nitroprusside (SNP, endothelium-indepen-dent vasodilator) was compared in precontractedvasculature from eyes in each group. The function ofthe endothelium and smooth muscle cells were therefore


assessed at different ages of this diabetic model.Concurrently blood glucose and body weight weremonitored.

In the streptozotocin rats, hyperglycaemia wasestablished and maintained in the week followinginduction to 28 weeks, then spontaneous recoverygradually to 78 weeks, becoming normoglycaemic.

Acetylcholine induced a dose dependent dilatationresponse, which was significantly increased at 5–8weeks, but impaired at 22–28, and 85–116 weeks postinduction in the STZ group when compared to agematched controls. Vasodilatation responses inducedby sodium nitroprusside in the ocular vasculature ofthe STZ groups were not affected at any stageexamined. Aging had no significant effect on theAch-induced endothelium-dependent or SNP inducedendothelium-independent vasodilatory function in therat ocular circulation. These results implicate thatendothelial dysfunction appeared during thehyperglycaemia and remained after the rats becamenormoglycaemic.

3.3.4. Tetrahydrobiopterin reverses the impairment of

acetylcholine-induced vasodilatation in diabetic ocular

microvasculature

We tested whether tetrahydrobiopterin, an essentialcofactor in nitric oxide synthesis, can reverse endothe-lium dysfunction in diabetic ocular circulation (Yu et al.,2001c). Using the streptozotocin-induced diabetic ratmodel and the isolated perfusion eye technique, theresponse to the acetylcholine-induced vasodilatation ofthe diabetic ocular vasculature before and after tetra-hydrobiopterin administration was compared. Agematched normal rats were used for reference responses.Six streptozotocin-induced diabetic rats and elevencontrol rats at 21 weeks post-induction were used. Thedose response curves from the diabetic eyes werefound to be significantly different from that of thecontrol eyes with significantly reduced responses to10�4M acetylcholine. After 30min of acute administra-tion of tetrahydrobiopterin to the diabetic eyes,however, the acetylcholine-induced vasodilatationresponse was significantly increased compared withthe response prior to tetrahydrobiopterin administra-tion. The vasodilatory response in the diabetic eyesafter tetrahydrobiopterin administration was at a levelthat was comparable with the control response.We have shown that acute administration of tetrahy-drobiopterin is effective in reversing to control levelthe impaired acetylcholine-induced vasodilatory re-sponse at 21 weeks post-induction. Our result suggeststhat a decreased level of tetrahydrobiopterin in the eyesof the streptozotocin-induced diabetic rats may beresponsible for the ocular vascular endothelium dys-function.

4. Discussion

Vascular involvement is clearly evident, or is im-plicated in a wide range of retinal diseases whichtogether account for the majority of new blindness inour community. There is potential for new drugtherapies to prevent or reduce the extent of vascularimpairment, and therefore ameliorate the extent ofvisual loss. To achieve this we need to improve ourunderstanding of the vasoactive properties of differentelements of the ocular vasculature in both health anddisease. Isolated preparations of the ocular vasculatureor its sub components have proven to be a valuablesource of such information. We have reviewed much ofour own experimental work in this area, and we nowseek to draw together the threads that bring together theoverall picture of vasoactive properties of the ocularvasculature.

4.1. Importance and relevance of isolated ocular vascular

preparations

Since in vivo experimentation can be complicated byfactors such as anaesthesia, systemic condition, andsurgical limitations to access the required tissue, manyisolated vascular preparations have been developed toinvestigate the vasoactive properties of the vasculature.For studies of the vasoactive responses of the ocularvasculature, isolated vascular preparations have anumber of the advantages: (1) removal of systemicinfluences, (2) precise control of environmental condi-tion to the vascular tissues, (3) ease of measurement andcontrol of all the relevant input parameters, (4) ability toperform studies at the cellular level (endothelium orsmooth muscle cells), (5) Applicable to many speciesincluding human tissue from eye banks or surgicalspecimens, (6) no anaesthesia required, (7) use offortuitously available eyes from other experiments, (8)use of expanding range of disease models, particularly inrats. However, in vitro techniques also have unavoidablelimitations, such as: (1) impossible to exactly recreatethe in vivo environment, (2) choice of perfusionparameters can influence vasoactivity, (3) loss of neuralcontrol mechanisms.

The real dilemma in using isolated vascular prepara-tions is that whilst it becomes possible to measure thevasoactive properties of the vessels, there is always thequestion of how this reflects their behaviour in the morecomplex in vivo environment. The ultimate parameter ofinterest is after all the local blood flow in vivo. Thishowever is often difficult or impossible to measure withexisting techniques. The degree of dissociation from thein vivo environment varies in each of the preparationsthat we have discussed. The isolated eye preparation hasthe advantage that the vascular system is essentiallyintact and the relationship between the vascular tissues


and the tissues that they supply is maintained. However,any perfusate solution is a poor mimic of healthy blood,and any neural control of the vasculature is clearly lost.With isolated vessel preparations the same limitationsapply, along with the additional disruption of dissectionfrom surrounding tissues. In ring segment preparationsthe intraluminal flow of blood is lost. Now that weunderstand the importance of the role of shear stressassociated with intraluminal flow (Yu et al., 1994a) thisis one aspect of vascular control that cannot be wellreproduced in ring segment preparations. In isolatedperfused vessel preparations the normal shear stressenvironment can be partially restored. However, differ-ences in viscosity between perfusate solutions andhealthy blood mean that this is only a partial solution.Intravascular pressure is another factor that is difficultto accurately mimic in vitro under normal flow condi-tions. The open ends of the branch vessels in our isolatedvessel preparations means that intravascular pressure islower than present in the in vivo situation under thesame flow conditions. It is not feasible to match bothflow and pressure to the in vivo environment.

However, as long as the limitations of in vitrotechniques are taken into account, studies on isolatedvessel preparations continue to yield a great deal ofuseful information about the vasoactive properties ofthe ocular vasculature. Although the isolated ocularvascular preparations and their importance and rele-vance have been demonstrated and accepted as a well-established technique, the assessment of vasoactiveresponse of ocular vessels, whether in vitro or in vivo,

can never provide a solution to all the problems ofocular vascular research. In vitro and in vivo techniqueswill continue to contribute significantly to ocularcirculation research. The basic approaches in theisolated vascular study must be seen as complementaryto in vivo studies on tissue blood flow and metabolism.

4.2. Important areas to be further addressed

The majority of studies using isolated ocular vascularpreparations have focussed on the basic pharmacologyof ocular vascular smooth muscle. Typically thisincludes quantitative assays of agonists or antagonists,comparison of the activities of different agonists,competitive antagonists, dose ratios and affinity con-stants. This basic pharmacology still remains a crucialelement of ocular circulatory research. New classes ofcompounds with possible effects on ocular vessels (andwith therapeutic possibilities) are worth identifying. Therecent assessment of the direct effects of anti-glaucomaagents on ocular circulation is a typical example.However, more optimistically, investigations in vasoac-tive properties may provide insight into the etiology andtreatment of pathological conditions such as retinalischemic diseases and diabetic retinopathy. The results

from our limited studies on altered agonists inducedeffects in the diabetic animal model and functionalchanges in the vascular endothelial cells suggest thatisolated ocular vascular preparations may help us toinvestigate the multifactorial pathways and underlyingmechanisms in the pathogenesis of diabetic retinopathy.

It is essential to further improve or develop newtechnologies for isolated vascular preparations toexpand our knowledge in this field. A number ofimportant areas should be further addressed in the nearfuture.

4.2.1. The role of the endothelium

How the endothelium plays a modulator role onsmooth muscle function has not been sufficientlyexplored in either normal or diseased conditions. Thevascular endothelium, the largest endocrine, paracrine,and autocrine regulatory tissue in our body, participatesin numerous processes related to the functioning of thevessel wall. In response to various stimuli the endothelialcells synthesize and release numerous biologically activefactors such as nitric oxide, endothelins, prostanoids,endothelium-derived hyperpolarizing factor, oxygen freeradicals, growth factors, tissue plasminogen activatorand inhibitor, and vascular cell adhesion molecules. Thestimuli for such release include hemodynamic forces(shear stress, cyclical strain) and circulating and locallyderived vasoactive substances (angiotensin II, catecho-lamines, vasopressin, arachidonic acid, thrombin, etc.),as well as interactions with leukocytes or platelets.Short-term interactions between these factors contributeto the regulation of vascular tone, while longer-terminterplay of such factors modulates smooth-muscle cellproliferation, extracellular matrix production, hyper-trophic growth, and apoptosis to regulate vascularstructure and remodelling. Endothelium-derived factorsalso regulate platelet aggregation, coagulation, andthrombosis, as well as leukocyte adhesion and inflam-mation via the expression of chemotactic and adhesionmolecules. Most information and knowledge that wehave today has been obtained from cardiovascularresearch from other organs. There are relatively limitedstudies available on the ocular vasculature. To addressthe role of the endothelium in the regulation of vasculartone, it is crucial to maintain and assess normalendothelium function, and also critical to combine otheradvanced techniques from vascular biology, such asfrom biochemical and molecular levels to determinealteration in the intracellular and intercellular signallingduring the physiological and pathological changes incellular environments.

4.2.2. Differential effects of intra- and extra-luminal

administration of drugs

Isolated perfused retinal arteriole preparations pro-vide an improved method for comparing the differences


between the intra- and extra-luminal administrations ofdrugs compared to that offered by the myographapproach. We have demonstrated the differential effectsof intra- and extra-luminal administrations of a numberof vasoactive substances such as noradrenaline, adrena-line, adenosine, and endothelin-1 (Alder et al., 1996; Yuet al., 1994a). The nature of the asymmetry depends onthe drug tested. ET-1 and adenosine were more effectivewhen the drug entry site was extraluminal, whereasnoradrenaline, adrenaline and K+ were more effectivewhen administered to the intraluminal surface. It is clearthat the endothelium, at least in the porcine and humanretinal arterioles, presents a diffusion barrier for thosevasoactive substances with differential effects of intra-and extra-luminal administration. The exact mechan-isms of such differential effects are unknown.

Several hypotheses have been raised to explainvascular wall asymmetry to the same agonist. Theseinclude (1) the lipid or water solubility of the drug (Lewand Duling, 1990, 1992; Matsuki et al., 1993), (2)possible differences in the types, affinity and numbers ofthe receptors on the intra- and extra-luminal side of thevascular smooth muscle and endothelial cells (Headricket al., 1992; Yu et al., 1994a), (3) interaction with theendothelial cell either via released factors, or by uptakemechanisms (Headrick et al., 1992; Ngai and Winn,1993; Ogawa et al., 1993) (4) the blood-flow-inducedstress signal to the endothelial cell (Bevan et al., 1988),(5) interaction with factors released from the metabolis-ing tissue adjacent to the vessel, and (6) the role playedby the autonomic innervation of the vessels (Mazmanianet al., 1993; Tesfamariam and Halpern, 1988).

Differential effects of intra- and extra-luminal admin-istration of drugs could be a valuable model formimicking the effects of vasoactive agents from theblood steam or tissue environments, and also forexploring the interactions between the vascular en-dothelium and smooth muscle cells.

4.2.3. Constant perfusion flow or constant pressure for

isolated vessel and isolated eye preparation

The technique of continuous perfusion adopted in ourisolated retinal vessel preparation confers severalimportant advantages over the more commonly usedring segment preparation which we and others have usedto investigate the vasoactivity of the larger oculararteries. A perfused artery corresponds more closely tothe in vivo situation. It allows differentiation of intra-and extra-luminal responses, and may cause lessendothelial cell damage, and is readily applied to smallervessels such as the human or porcine retinal arteries.Moreover the use of a controlled perfusion techniquemeans that the flow rate is known, whilst the intralum-inal pressure and vessel diameter can be measured. Druginduced changes in perfusion pressure provides informa-tion about the combined vascular resistance of the main

vessel and its single side branch through which theperfusate exits.

Interestingly, our comparison of pressurized vesselswith no flow, and pressurized vessels with flow,uncovered a strong myogenic contraction with increas-ing pressure for vessels in which there was no flow (Yuet al., 1994a). However, with flow, at the same pressure,the vessel dilates. Bevan (Bevan and Joyce, 1988) alsoobserved flow induced dilatation and concluded that itwas independent of endothelial cell removal. It seemslikely that vessel diameter reflects the balance ofopposing forces generated by a combination of mechan-isms which are not currently well understood. Underconditions where flow was present the arterial diameterwas relatively unaffected by incremental increases inflow.

In attempting to correlate results between isolatedperfused eyes and isolated perfused vessels a constantflow technique at least ensures a similar flow environ-ment. A constant pressure approach is not feasible inisolated perfused eyes where there is inevitably apressure drop along the vascular tree.

From a practical point of view, it is hard to matchboth pressure and flow to in vivo values when using invitro techniques. Although continuous perfusion createsmore technical demands than constant pressure and noflow conditions, it yields a preparation closer to thatseen in vivo.

4.3. Possibility of prediction of in vivo from in vitro data

To address this issue of differential control and othercontrolling mechanisms in the retinal vascular bed, wehave developed several in vitro techniques, which ruleout systemic influences and overcome the problem ofblood flow measurement.

What do we already know about retinal blood flow?We believe there must be a match between the flow ofnutrients to the tissue and the tissue demands at alllocations. What are the mechanisms that may controlretinal blood flow? The retinal circulation appears to beunique in having no autonomic innervation, implyingthat distal to the optic nerve head all blood flow controlmust be by local mechanisms. The most likely con-tenders are blood-borne factors such as the catechola-mines, tissue released factors signalling hypoxia orischemia such as adenosine, endothelial cell mediatingfactors such as nitric oxide and the endothelins, as wellas flow induced effects and pressure induced effects (alsoknown as the myogenic response). At any one locationdown the vascular bed, these five factors may interact toproduce the final response. We propose a segmentalhypothesis of blood vessel function in the normal retina.We hypothesize that the balance of the above fivemechanisms varies down the vascular bed, and betweendifferent regions of the vasculature. For example, it


could be imagined that, close to the capillaries, vascularcontrol may be dominated by feedback tissue releasedfactors, whereas in first order arterioles pressure andflow effects may be dominant.

These issues of local control mechanisms in differentorders of retinal microvasculature remain to be deter-mined and can only be sorted out by experimentation.With a range of in vitro techniques, it becomes possibleto unravel these changes in balance between thesefactors. A marriage of the knowledge obtained withthe in vitro techniques on the myogenic responses, theflow induced dilatation, the major blood-borne, en-dothelial-cell-derived and tissue released factors will becombined and integrated for testing the hypothesesraised in the in vivo animal. Ultimately this integratedapproach linking in vivo and in vitro work will lead to adeeper level of under standing of the physiology,pathology and pharmacology of ocular circulation.

5. Summary and research directions

This review has described several isolated vascularpreparations currently used in our group. Importantinsights into understanding the local control mechan-isms of the ocular circulation in normal conditions andfollowing pathogenesis in ocular diseases involvingblood flow changes have been gained. The possibilityof predicting in vivo responses from in vitro data needsto be addressed further. Vascular endothelial cells play akey role in the regulation of blood flow in normal anddiseased conditions. Up to now we have only verylimited knowledge of the function of the endotheliumand the signalling between the endothelium and smoothmuscle cells. The communication between vascular cellsand retinal tissue has also not been studied sufficiently.Integration of the different orders of ocular vasculatureis another important topic to be considered. Newisolated vessel preparations and techniques for studyingthe physiology and pharmacology in the different ordersof the choroidal vasculature need to be developed. Webelieve that such isolated ocular vascular preparationswill allow further investigation of normal vascularbiology, exploration of the pathogenesis of ocularvascular diseases and ultimately assist in the develop-ment of new therapeutic strategies.

References

Ak, G., Buyukberber, S., Sevinc, A., Turk, H.M., Ates, M., Sari, R.,

Savli, H., Cigli, A., 2001. The relation between plasma endothelin-1

levels and metabolic control, risk factors, treatment modalities, and

diabetic microangiopathy in patients with Type 2 diabetes mellitus.

J. Diabetes. Complicat. 15, 150–157.

Alder, V.A., Su, E.N., Yu, D.Y., Cringle, S.J., 1993. Oxygen reactivity

of the feline isolated ophthalmociliary artery. Invest. Ophthalmol.

Vis. Sci. 34, 1–9.

Alder, V.A., Su, E.N., Yu, D.Y., Cringle, S.J., Yu, P.K., 1996.

Asymmetrical response of the intraluminal and extraluminal

surfaces of the porcine retinal artery to exogenous adenosine.

Exp. Eye Res. 63, 557–564.

Alm, A., 1972. Effects of norepinephrine, angiotensin, dihydroergota-

mine, papaverine, isoproterenol, histamine, nicotinic acid, and

dxdanthinol nicotinate on retinal oxygen tension in cats. Acta

Ophthalmol. 50, 707–719.

Alm, A., Bill, A., 1970. Blood flow and oxygen extraction in the cat

uvea at normal and high intraocular pressure. Acta Physiol. Scand.

80, 19–28.

Ames, A., 1992. Energy requirements of CNS cells as related to their

function and to their vulnerability to ischemia: a commentary

based on studies on retina. Can. J. Pharmacol. 70, S158–S164.

Anderson, B., 1968. Ocular effects of changes in oxygen and carbon

dioxide tension. Trans. Am. Ophthalmol. Soc. 66, 423–474.

Anderson, B., Saltzman, H.A., 1964. Retinal oxygen utilization

measured by hyperbaric blackout. Arch. Ophthalmol. 72, 792–795.

Bakken, I.J., Vincent, M.B., Sjaavaag, I., White, L.R., 1995.

Vasodilation in porcine ophthalmic artery: peptide interaction

with acetylcholine and endothelial dependence. Neuropeptides 29,

69–75.

Bank, H.L., Brockbank, K.G.M., 1987. Basic principles of cryobiol-

ogy. J Cardiovascular Surg. 1, 137–143.

Baron, A.D., 1994. Hemodynamic actions of insulin. Am. J. Physiol.

30, E187–E202.

Benedito, S., Prieto, D., Nielsen, P.J., Nyborg, N.C.B., 1991a.

Histamine induces endothelium-dependent relaxation of bovine

retinal arteries. Invest. Ophthalmol. Vis. Sci. 32, 32–38.

Benedito, S., Prieto, D., Nielsen, P.J., Nyborg, N.C.B., 1991b. Role of

the endothelium in acetylcholine-induced relaxation and sponta-

neous tone of bovine isolated retinal small arteries. Exp. Eye Res.

52, 575–579.

Bessho, H., Suzuki, J., Narimatsu, A., Tobe, A., 1990. Antihyperten-

sive effect of betaxolol a cardioselective beta-adrenoceptor

antagonist in experimental hypertensive rats. Nippon Yakurigaku

Zasshi. 95, 347–354.

Bevan, J.A., Joyce, E.H., 1988. Flow-dependent contraction observed

in a myograph-mounted resistance artery. Blood Vessels 25, 261–

264.

Bevan, J.A., Joyce, E.H., Wellman, G.C., 1988. Flow-dependent

dilation in a resistance artery still occurs after endothelium

removal. Circ. Res. 63, 980–985.

Bhattacharya, J., Nanjo, S., Staub, N.C., 1982. Micropuncture

measurement of lung microvascular pressure during 5-HT infusion.

J. Appl. Physiol. 52, 634–637.

Bill, A., 1962. Intraocular pressure and blood flow through the uvea.

Arch. Ophthalmol. 67, 336–348.

Blazynski, C., 1990. Discrete distributions of adenosine receptors in

mammalian retina. J. Neurochem. 54, 648–655.

Blazynski, C., Perez, M.T., 1991. Adenosine in vertebrate retina:

localization, receptor characterization, and function. Cell. Mol.

Biol. 11, 463–484.

Brubaker, R.F., 1991. Flow of aqueous humor in humans. Invest.

Ophthalmol. Vis. Sci. 32, 3145–3166.

Chiou, G.C.Y., Chen, Y.J., 1993. Effects of antiglaucoma drugs on

ocular blood flow in ocular hypertensive rabbits. J. Ocul.

Pharmacol. 9, 13–24.

Cogan, D.G., Toussaint, D., Kuwabara, T., 1961. Retinal

vascular pattern. IV Diabetic retinopathy. Arch. Ophthalmol. 66,

366–378.

Collignon-Brach, J., 1992. Primary open-angle glaucoma and beta-

blocking agents. In: Drance, S.M. (Ed.), International Symposium


On Glaucoma, Ocular flow, and Drug Treatment. Williams &

Williams, Baltimore, MD, pp. 136–151.

Cooper, R.L., 1990. Blind registrations in Western Australia: a five-

year study. Aust. NZ J. Ophthalmol. 18, 421–426.

Coyle, J.T., Puttfarcken, P., 1993. Oxidative stress, glutamate, and

neurodegenerative disorders. Science 262, 689–695.

Cringle, S.J., Yu, D.Y., 2002. A multi-layer model of retinal oxygen

supply and consumption helps explain the muted rise in inner

retinal PO2 during systemic hyperoxia. Comp. Biochem. Physiol.

132, 61–66.

Cringle, S., Yu, D.Y., Alder, V., Su, E.N., 1992. Oxygen tension and

blood flow in the retina of normal and diabetic rats. Oxygen

Transp. Tissue 14, 787–791.

Cringle, S.J., Yu, D.Y., Alder, V.A., Su, E.N., 1993. Retinal blood

flow by hydrogen clearance polarography in the streptozotocin-

induced diabetic rat. Invest. Ophthalmol. Vis. Sci. 34, 1716–1721.

Cringle, S.J., Yu, D.-Y., Yu, P.K., Su, E.-N., 2002. Intraretinal oxygen

consumption in the rat in vivo. Invest. Ophthalmol. Vis. Sci. 43,

1923–1928.

Dallinger, S., Dorner, G.T., Wenzel, R., Graselli, U., Findl, O.,

Eichler, H.G., Wolzt, M., Schmetterer, L., 2000. Endothelin-1

contributes to hypoxia-induced vasoconstriction in the human

retina. Invest Ophthalmol. Vis. Sci. 41, 864–869.

Davis, M.D., 1992. Diabetic retinopathy—a clinical overview.

Diabetes Care. 15, 1844–1874.

de Schaepdrijver, L., Simoens, P., Pollet, L., Lauwers, H., de Laey,

J.J., 1992. Morphological and clinical study of the retinal

circulation in the miniature pig. B: flluorescein angiography of

the retina. Exp. Eye Res. 54, 975–985.

Defily, D.V., Chilian, W.M., 1995. Coronary microcirculation:

autoregulation and metabolic control. Basic Res. Cardiol. 90,

112–118.

Delaey, C., van de Voorde, J., 1998. Retinal arterial tone is controlled

by a retinal-derived relaxing factor. Circ. Res. 83, 714–720.

Delaey, C., Boussery, K., van de Voorde, J., 2000. A retinal-derived

relaxing factor mediates the hypoxic vasodilation of retinal arteries.

Invest. Ophthalmol. Vis. Sci. 41, 3555–3560.

Drance, S.M., 1997. Glaucoma: a look beyond intraocular pressure.

Am. J. Ophthalmol. 123, 817–819.

Duijm, H.F.A., Vandenberg, T.J.T.P., Greve, E.L., 1997. Choroidal

hemodynamics in glaucoma. Br. J. Ophthalmol. 81, 735–742.

Duling, B.R., Rivers, R.J., 1986. Isolation, cannulation, and perfusion

of microvessels. In: Anonymous (Ed.), Microcirculatory Technol-

ogy, Academic Press, New York, pp. 265–280.

Ernest, J.T., Goldstick, T.K., 1983. Response of choroidal vascular

resistance to hyperglycemia. Int. Ophthalmol. 6, 119–124.

Ferrari-Dileo, G., Davis, E.B., Anderson, D.R., 1991. Angiotensin II

Binding receptors in retinal and optic nerve head blood vessels.


Ferrari-Dileo, G., Ryan, J.W., Rockwood, E.J., Davis, E.B.,

Anderson, D.R., 1988. Angiotensin-converting enzyme in bovine,

feline, and human ocular tissues. Invest. Ophthalmol. Vis. Sci. 29,

876–881.

Flammer, J., 1994. The vascular concept of glaucoma. Surv.


Flammer, J., 1996. To what extent are vascular factors involved in the

pathogenesis of glaucoma? In: Kaiser, H.J., Flammer, J., Hen-

drickson, P. (Eds.), Ocular Blood Flow. New Insights into the

Pathogenesis of Ocular Diseases. Karger, Basel, pp. 12–39.

Forster, B.A., Ferrari-Dileo, G., Anderson, D.R., 1987. Adrenergic

alpha1 and alpha2 binding sites are present in bovine retinal blood

vessels. Invest. Ophthalmol. Vis. Sci. 28, 1741–1746.

Frank, R.N., 1995. Diabetic retinopathy. Prog. Retina. Eye. Res. 14,

361–392.

Friedman, E., 1970. Choroidal blood flow. Pressure-flow relationships.

Arch. Ophthalmol. 83, 95–99.

Frishman, W.H., Fuksbrumer, M.S., Tannenbaum, M., 1994. Topical

ophthalmic B-adrenergic blockade for the treatment of glaucoma

and ocular hypertension. J. Clin. Pharmacol. 34, 795–803.

Geier, S.A., Rolinski, B., Sadri, I., Kronawitter, U., Bogner, J.R.,

Klauss, V., Goebel, F.D., 1995. Ocular microangiopathy syndrome

in patients with AIDS is associated with increased plasma levels of

the vasoconstrictor endothelin-1. Klin. Monatsbl. Augenheilk. 207,

353–360.

Gidday, J.M., Park, T.S., 1993a. Adenosine-mediated autoregulation

of retinal arteriolar tone in the piglet. Invest. Ophthalmol. Vis. Sci.

34, 2713–2719.

Gidday, J.M., Park, T.S., 1993b. Microcirculatory responses to

adenosine in the newborn pig retina. Pediatr. Res. 33, 620–627.

Gouras, P., Hoff, M., 1970. Retinal function in an isolated, perfused

mammalian eye. Invest. Ophthalmol. Vis. Sci. 9, 388–399.

Granstam, E., Wang, L., Bill, A., 1992. Ocular effects of endothelin-1

in cats. Curr. Eye Res. 11, 325–332.

Grunwald, J.E., 1986. Effect of topical timolol on the human retinal

circulation. Invest. Ophthalmol. Vis. Sci. 27, 1713–1719.

Grunwald, J.E., Riva, C.E., Brucker, A.J., Sinclair, S.H., Petrig, B.L.,

1984a. Altered retinal vascular response to 100% oxygen breathing

in diabetes mellitus. Ophthalmology 91, 1447–1452.

Grunwald, J.E., Riva, C.E., Petrig, B.L., Sinclair, S.H., Brucker, A.J.,

1984b. Effect of pure O2-breathing on retinal blood flow in normals

and in patients with background diabetic retinopathy. Curr. Eye

Res. 3, 239–242.

Grunwald, J.E., Riva, C.E., Martin, D.B., Quint, A.R., Epstein, P.A.,

1987. Effect of an insulin-induced decrease in blood glucose

on the human diabetic retinal circulation. J. Ophthalmol. 94,

1614–1620.

Hardy, P., Abran, D., Li, D.Y., Fernandez, H., Varma, D.R.,

Chemtob, S., 1994. Free Radicals in Retinal and Choroidal Blood

Flow Autoregulation in the Piglet: interaction with prostaglandins.


Harris, A., Arend, O., Chung, H.S., Kagemann, L., Cantor, L.,

Martin, B., 2000. A comparative study of betaxolol and

dorzolamide effect on ocular circulation in normal-tension

glaucoma patients. Ophthalmology 107, 430–434.

Headrick, J.P., Northington, F.J., Hynes, M.R., Matherne, G.P.,

Berne, R.M., 1992. Relative responses to luminal and adventitial

adenosine in perfused arteries. Am. J. Physiol. 263, H1437–H1446.

Hendriks, M.G.C., Kam, K.L., Pijl, A.J., Pfaffendorf, M., van

Zwieten, P.A., 1993. The effects of hypertension and diabetes

mellitus on the vascular reactivity of resistance arteries. Blood

Pressure. 2, 69–76.

Hester, R.K., Chen, Z., Becker, E.J., McLaughlin, M., DeSantis, L.,

1994. The direct vascular relaxing action of betaxolol, carteolol and

timolol in porcine long posterior ciliary artery. Surv. Ophthalmol.

38, S125–S134.

Holstein-Rathlou, N.H., Marsh, D.J., 1994. A dynamic model of renal

blood flow autoregulation. Bull. Math. Biol. 56, 411–429.

Hoste, A.M., 1997. Reduction of IOP with latanoprost. Ophthalmol-

ogy. 104, 895–896.

Hoste, A.M., 1999. In vitro studies of the effects of beta-adrenergic

drugs on retinal and posterior ciliary microarteries 1. Surv.

Ophthalmol. 43 (Suppl. 1), S183–S190.

Hoste, A.M., Andries, L.J., 1991. Contractile responses of isolated

bovine retinal microarteries to acetylcholine. Invest. Ophthalmol.

Vis. Sci. 32, 1996–2005.

Hoste, A.M., Boels, P.J., Brutsaert, D.L., de Laey, J.J., 1989. Effect of

alpha-1 and beta agonists on contraction of bovine retinal

resistance arteries in vitro. Invest. Ophthalmol. Vis. Sci. 30, 44–50.

Hoste, A.M., Boels, P.J., Andries, L.J., Brutsaert, D.L., de Laey,

J.J., 1990. Effects of beta-antagonists on contraction of bovine

retinal microarteries in vitro. Invest. Ophthalmol. Vis. Sci. 31,

1231–1237.


Hoste, A.M., Sys, S.U., 1998. Ca2+ channel-blocking activity of

propranolol and betaxolol in isolated bovine retinal microartery 2.

J. Cardiovasc. Pharmacol. 32, 390–396.

Iannaccone, A., Letizia, C., Pazzaglia, S., Vingolo, E.M., Clemente,

G., Pannarale, M.R., 1998. Plasma endothelin-1 concentrations in

patients with retinal vein occlusions. Br. J. Ophthalmol. 82, 498–

503.

Kaiser, H.J., Flammer, J., Wenk, M., Luscher, T., 1995. Endothelin-1

plasma levels in normal-tension glaucoma: abnormal response to

postural changes. Graefes. Arch. Clin. Exp. Ophthalmol. 233, 484–

488.

Kiel, J.W., 2000. Endothelin modulation of choroidal blood flow in the

rabbit. Exp. Eye Res. 71, 543–550.

Kiel, J.W., Shepherd, A.P., 1992. Autoregulation of Choroidal Blood

Flow in the Rabbit. Invest. Ophthalmol. Vis. Sci. 33, 2399–2410.

Kohler, K., Wheeler-Schilling, T., Jurklies, B., Guenther, E., Zrenner,

E., 1997. Angiotensin II in the rabbit retina. Vis. Neurosci. 14, 63–

71.

Kohner, E.M., 1976. The problems of retinal blood flow in diabetes.

Diabetes. 25, 839–844.

Konno, S., Feke, G.T., Yoshida, A., Fujio, N., Goger, D.G., Buzney,

S.M., 1996. Retinal blood flow changes in type I diabetes. A long-

term follow-up study. Invest. Ophthalmol. Vis. Sci. 37, 1140–1148.

Kulkarni, P.S., Hamid, H., Barati, M., Butulija, D., 1999. Angiotensin

II-induced constrictions are masked by bovine retinal vessels.

Invest Ophthalmol. Vis. Sci. 40, 721–728.

Lachkar, Y., Migdal, C., Dhanjil, S., 1998. Effect of brimonidine

tartrate on ocular hemodynamic measurements. Arch. Ophthal-

mol. 116, 1591–1594.

Lanigan, L.P., Clark, C.V., Allawi, J., Hill, D.W., Keen, H., 1990.

Impaired autoregulation of the retinal vasculature and micro-

albuminuria in diabetes mellitus. Eye. 4, 174–180.

Laties, A.M., 1967. Central retinal artery innervation. Arch. Ophthal-

mol. 77, 405–409.

Laties, A.M., Jacobowitz, D., 1966. A comparative study of the

autonomic innervation of the eye in monkey, cat, and rabbit. Anat.

Rec. 156, 383–389.

Lew, M.J., Duling, B.R., 1990. Arteriolar reactivity in vivo is

influenced by an intramural diffusion barrier. Am. J. Physiol.

(Heart Circ. Phys.). 28, H574–H581.

Lew, M.J., Duling, B.R., 1992. Access of Blood-Borne Vasoconstric-

tors to the Arteriolar Smooth Muscle. J. Vasc. Res. 29, 341–346.

Linsenmeier, R.A., 1986. Effects of light and darkness on oxygen

distribution and consumption in the cat retina. J. Gen. Physiol. 88,

521–542.

Liu, J., Chen, R., Casley, D.J., Nayler, W.G., 1990. Ischemia and

reperfusion increase 125I-labeled endothelin-1 binding in rat cardiac

membranes. Am. J. Physiol. 258, H829–H835.

Masaki, T., Yanagisawa, M., 1992. Physiology and pharmacology of

endothelins. Med. Res. Rev. 12, 391–421.

Matsuki, T., Hynes, M.R., Duling, B.R., 1993. Comparison of conduit

vessel and resistance vessel reactivity: influence of intimal perme-

ability. Am. J. Physiol. 33, H1251–H1255.

Mazmanian, G.M., Baudet, B., Pannier-Poulain, C., Choudat, L.,

Dulmet, E., Perrin, A., Weiss, M., Herve, Ph., 1993. Pressor

responses of rat isolated tail arteries to contractile stimulation after

methylene blue treatment: effect of adventitial versus intimal entry.

J. Vasc. Res. 30, 250–256.

McKibbin, M., Menage, M.J., 1999. The effect of once-daily

latanoprost on intraocular pressure and pulsatile ocular blood

flow in normal tension glaucoma. Eye 13, 31–34.

Messmer, C., Flammer, J., Stumpfig, D., 1991. Influence of betaxolol

and timolol on the visual fields of patients with glaucoma. Am. J.


Meyer, P., Flammer, J., Luscher, T.F., 1993. Endothelium-dependent

regulation of the ophthalmic microcirculation in the perfused

porcine eye: role of nitric oxide and endothelins. Invest. Ophthal-

mol. Vis. Sci. 34, 3614–3621.

Meyer, P., Flammer, J., Luscher, T.F., 1995a. Local action of the renin

angiotensin system in the porcine ophthalmic circulation: effects of

ace-inhibitors and angiotensin receptor antagonists. Invest.


Meyer, P., Lang, M.G., Flammer, J., Luscher, T.F., 1995b. Effects of

calcium channel blockers on the response to endothelin-1,

bradykinin and sodium nitroprusside in porcine ciliary arteries.

Exp. Eye Res. 60, 505–510.

Michelson, G., Langhans, M.I., Groh, M.J.M., 1996. Perfusion of the

juxtapapillary retina and the neuroretinal rim area in primary open

angle glaucoma. J. Glaucoma. 5, 91–98.

Morgan, W.H., Yu, D.Y., Cooper, R.L., Cringle, S.J., Alder, V.A.,

1995. The influence of cerebrospinal fluid pressure upon the lamina

cribrosa tissue pressure gradient - correspondence. Invest. Ophthal-

mol. Vis. Sci. 36, 2163–2164.

Morgan, W.H., Yu, D.Y., Alder, V.A., Cringle, S.J., Cooper, R.L.,

House, P.H., Constable, I.J., 1998. The correlation between

cerebrospinal fluid pressure and retrolaminar tissue pressure.


Murata, T., Nagai, R., Ishibashi, T., Inomata, H., Ikeda, K.,

Horiuchi, S., 1997. The relationship between accumulation of

advanced glycation end products and expression of vascular

endothelial growth factor in human diabetic retinas. Diabetologia

40, 764–769.

Neufeld, A.H., Bartels, S.P., 1982. Receptor mechanisms for

epinephrine and timolol. In: Lutjen-Drecoll, E. (Ed.), Basic aspects

of Glaucoma Research. Schattauer, Stuttgart, Germany, pp. 113–

122.

Ngai, A.C., Winn, H.R., 1993. Effects of adenosine and its analogues

on isolated intracerebral arterioles—extraluminal and intraluminal

application. Circ. Res. 73, 448–457.

Nielsen, P.J., Nyborg, N.C.B., 1989a. Adrenergic responses in isolated

bovine retinal resistance arteries. Int. Ophthalmol. 13, 103–107.

Nielsen, P.J., Nyborg, N.C.B., 1989b. Calcium antagonist-induced

relaxation of the prostaglandin-F2 response of isolated calf retinal

resistance arteries. Exp. Eye Res. 48, 329–335.

Nielsen, P.J., Nyborg, N.C.B., 1990. Contractile and relaxing effects of

arachidonic acid derivatives on isolated bovine retinal resistance

arteries. Exp. Eye Res. 50, 305–311.

Niemeyer, G., 2001. Retinal Research using the perfused mammalian

eye. Prog. Retina. Eye. Res. 20, 289–318.

Nishimura, T.,Okamoto, N., 1998. Changes in ocular blood circula-

tion after six months treatment with isopropyl unoprostone

(Resculas). in patients with normal tension glaucoma. Book of

Abstracts 28th, 140(P6.086). ICO, Amsterdam, (GENERIC) (Ref

Type: Abstract).

Nyborg, N.C.B., Nielsen, P.J., 1990. Angiotensin-II contracts isolated

human posterior ciliary arteries. Invest. Ophthalmol. Vis. Sci. 31,

2471–2473.

Nyborg, N.C.B., Nielsen, P.J., 1995. Beta-adrenergic receptors

regulating vascular smooth muscle tone are only localized to the

intraocular segment of the long posterior ciliary artery in bovine

eye. Surv. Ophthalmol. 39, S66–S75.

Nyborg, N.C., Nielsen, P.J., Prieto, D., Benedito, S., 1990. Angio-

tensin II does not contract bovine retinal resistance arteries in vitro

10. Exp. Eye Res. 50, 469–474.

Nyborg, N.C.B., Prieto, D., Benedito, S., Nielsen, P.J., 1991.

Endothelin-1-induced contraction of bovine retinal small arteries

is reversible and abolished by nitrendipine. Invest. Ophthalmol.

Vis. Sci. 32, 27–31.

Ogawa, R., Ohta, T., Tsuji, M., Mori, M., 1993. Role of the

endothelium on extraluminal and intraluminal vasoactive

mechanisms in the perfused rabbit basilar artery. Neurol. Res.

15, 154–159.


Ogo, T., 1996. Effect of Isopropyl unoprostone on choroidal

circulation I. Changes in choroidal blood flow and intraocular

pressure with topically application. Folia Ophthlamol. 47, 268–272.

Ohkubo, H., Chiba, S., 1987. Vascular reactivities of simian

ophthalmic and ciliary arteries. Curr. Eye Res. 6, 1197–1203.

Park, T.S., Gidday, J.M., 1990. Effect of dipyridamole on cerebral

extracellular adenosine level in vitro. J. Cerebral Blood Flow

Metabolism. 10, 424–427.

Pillunat, L.E., Bohm, A.G., Koller, A.U., Schmidt, K.G., Klemm, M.,

Richard, G., 1999. Effect of topical dorzolamide on optic nerve

head blood flow. Graefes. Arch. Clin. Exp. Ophthalmol. 237, 495–

500.

Polak, K., Petternel, V., Luksch, A., Krohn, J., Findl, O., Polska, E.,

Schmetterer, L., 2001. Effect of endothelin and BQ123 on ocular

blood flow parameters in healthy subjects. Invest. Ophthalmol. Vis.

Sci. 42, 2949–2956.

Pournaras, C.J., Riva, C.E., Tsacopoulos, M., Strommer, K., 1989.

Diffusion of O2 in the retina of anesthetized miniature pigs in

normoxia and hyperoxia. Exp. Eye Res. 49, 347–360.

Prunte, C.H., Flammer, J., 1989. Choroidal angiography findings

in patients with glaucoma-like visual field defects. In: Heijl,

A. (Ed.), Perimetry Update. Kugler & Ghedini, Amstelveen,

pp. 325–327.

Randall, M.D., 1991. Vascular activities of the endothelins. [Review]

[148 refs]. Pharmacol. Ther. 51, 73–93.

Riva, C.E., Sinclair, S.H., Grunwald, J.E., 1981. Autoregulation of

retinal circulation in response to decrease of perfusion pressure.


Riva, C.E., Grunwald, J.E., Sinclair, S.H., Petrig, B.L., 1985. Blood

velocity and volumetric flow rate in human retinal vessels. Invest.


Rockwood, E.J., Fantes, F., Davis, E.B., Anderson, D.R., 1987. The

response of retinal vasculature to angiotensin. Invest. Ophthalmol.

Vis. Sci. 28, 676–682.

Rudolphi, K.A., Schubert, P., Parkinson, F.E., Fredholm, B.B., 1992.

Adenosine and Brain Ischemia. Cerebrovasc. Brain Metabolism

Rev. 4, 346–369.

Sciotti, V.M., Park, T.S., Berne, R.M., Vanwylen, D.G.L., 1997.

Changes in extracellular adenosine during chemical or electrical

brain stimulation. Brain Res. 613, 16–20.

Seong, G.J., Lee, H.K., Hong, Y.J., 1999. Effects of 0.005%

latanoprost on optic nerve head and peripapillary retinal blood

flow. Ophthalmologica 213, 355–359.

Simoens, P., de Schaepdrijver, L., Lauwers, H., 1992. Morphologic

and clinical study of the retinal circulation in the miniature pig. A:

Morphology of the retinal microvasculature. Exp. Eye Res. 54,

965–973.

Small, K.W., Stefansson, E., Hatchell, D.L., 1987. Retinal blood flow

in normal and diabetic dogs. Invest. Ophthalmol. Vis. Sci. 28, 672–

675.

Stjernschantz, J., Selen, G., Astin, M., Resul, B., 2000. Microvascular

effects of selective prostaglandin analogues in the eye with special

reference to latanoprost and glaucoma treatment. Prog. Retina

Eye. Res. 19, 459–496.

Stowe, D.F., O’Brien, W.C., Chang, D., Knop, C.S., Kampine, J.P.,

1997. Reversal of endothelin-induced vasoconstriction by

endothelium- dependent and -independent vasodilators in

isolated hearts and vascular rings. J. Cardiovasc. Pharmacol. 29,

747–754.

Su, E.N., Yu, D.Y., Alder, V.A., Cringle, S.J., 1994. Effects of

extracellular pH on agonist-induced vascular tone of the cat

ophthalmociliary artery. Invest. Ophthalmol. Vis. Sci. 35, 998–

1007.

Su, E.N., Yu, D.Y., Alder, V.A., Yu, P.K., Cringle, S.J., 1995. Altered

vasoactivity in the early diabetic eye: measured in the isolated

perfused rat eye. Exp. Eye Res. 61, 699–712.

Su, E.N., Yu, D.Y., Alder, V.A., Cringle, S.J., Yu, P.K., 1996. Direct

vasodilatory effect of insulin on isolated retinal arterioles. Invest.


Su, E.N., Yu, D.Y., Cringle, S.J., Alder, V.A., Yu, P.K., Buckland, L.,

1998. Preservation of vasoactive properties of human retinal

arteries after cryopreservation. Aust. N. Z. J. Ophthalmol. 26,

S59–S61.

Su, E.-N., Alder, V.A., Yu, D.-Y., Yu, P.K., Cringle, S.J., 2000.

Continued progression of retinopathy despite spontaneous recov-

ery to normoglycemia in a long-term study of streptozotocin-

induced diabetes in rats. Graefes. Arch. Clin. Exp. Ophthalmol.

238, 163–173.

Sugiyama, T., Azuma, I., 1995. Effect of UF-021 on optic nerve head

circulation in rabbits. Jpn. J. Ophthalmol. 39, 124–129.

Sugiyama, T., Goldman, W.F., 1995. Measurement of SR free Ca2+

and Mg2+ in permeabilized smooth muscle cells with use of

furaptra. Am. J. Physiol. 38, C698–C705.

Sugiyama, T., Moriya, S., Oku, H., Azuma, I., 1995. Association of

endothelin-1 with normal tension glaucoma: clinical and funda-

mental studies. Surv. Ophthalmol. 39, S49–S56.

Sugiyama, K., Haque, M.S.R., Onda, E., Taniguchi, T., Kitazawa, Y.,

1996. The effects of intravitreally injected endothelin-1 on the iris-

ciliary body microvasculature in rabbits. Curr. Eye Res. 15,

633–637.

Tesfamariam, B., Halpern, W., 1988. Asymmetry of responses to

norepinephrine in perfused resistance arteries. Eur. J. Pharmacol.

152, 167–170.

Tilton, R.G., Chang, K., Pugliese, G., Eades, D.M., Province, M.A.,

Sherman, W.R., Kilo, C., Williamson, J.R., 1989. Prevention of

hemodynamic and vascular albumin filtration changes in diabetic

rats by aldose reductase inhibitors. Diabetes 37, 1258–1270.

Vanderkooi, J.M., Erecinska, M., Silver, I.A., 1991. Oxygen in

mammalian tissue: methods of measurement and affinities of

various reactions. Am. J. Physiol. 260, C1131–C1150.

Wagner, A.J., Holstein-Rathlou, N.H., Marsh, D.J., 1996. Endothelial

Ca2+ in afferent arterioles during myogenic activity. Am. J.

Physiol. 39, F170–F178.

Wolf, S., Arend, O., Sponsel, W.E., Schulte, K., Cantor, L.B., Reim,

M., 1993. Retinal hemodynamics using scanning laser ophthalmo-

scopy and hemorheology in chronic open-angle glaucoma.

Ophthalmology 100, 1561–1566.

Yan, H.Y., Chiou, G.C.Y., 1987. Effects of L-timolol,

D-timolol, haloperidol and domperidone on rabbit retinal

blood flow measured with laser doppler method. Ophthalmic

Res. 19, 45–48.

Yanoff, M., Fine, B.S., 1989. In: Yanoff, M., Fine, B.S. (Eds.), Ocular

Pathology a Text and Atlas. J.B. Lippincott Company, Philadel-

phia, pp. 1–737.

Yu, D.Y., Cringle, S.J., 2001. Oxygen distribution and consumption

within the retina in vascularised and avascular retinas and in

animal models of retinal disease. Prog. Retina. Eye. Res. 20, 175–

208.

Yu, D.Y., Cringle, S.J., 2002. Outer retinal anoxia during dark

adaptation is not a general property of mammalian retinas. Comp.

Biochem. Physiol. 132, 47–52.

Yu, D.Y., Alder, V.A., Cringle, S.J., Brown, M.J., 1988. Choroidal

blood flow measured in the dog eye in vivo and in vitro by

local hydrogen clearance polarography: validation of a technique

and response to raised intraocular pressure. Exp. Eye Res. 46,

289–303.

Yu, D.Y., Alder, V.A., Cringle, S.J., 1990. In vitro characterization of

the mechanical properties of canine ophthalmociliary artery. Exp.

Eye Res. 51, 729–734.

Yu, D.Y., Alder, V.A., Su, E.N., Cringle, S.J., 1992a. Relaxation

effects of diltiazem, verapamil, and tolazoline on isolated cat

ophthalmociliary artery. Exp. Eye Res. 55, 757–766.


Yu, D.Y., Alder, V.A., Su, E.N., Mele, E.M., Cringle, S.J., Morgan,

W.H., 1992b. Agonist response of human isolated posterior ciliary

artery. Invest. Ophthalmol. Vis. Sci. 33, 48–54.

Yu, D.Y., Su, E.N., Alder, V.A., Cringle, S.J., Mele, E.M., 1992c.

Pharmacological and mechanical heterogeneity of cat isolated

ophthalmociliary artery. Exp. Eye Res. 54, 347–359.

Yu, D.Y., Alder, V.A., Cringle, S.J., Su, E.N., Yu, P.K., 1994a.

Vasoactivity of intraluminal and extraluminal agonists in

perfused retinal arteries. Invest. Ophthalmol. Vis. Sci. 35,

4087–4099.

Yu, D.Y., Cringle, S.J., Alder, V.A., Su, E.N., 1994b. Intraretinal

oxygen distribution in rats as a function of systemic blood pressure.

Am. J. Physiol. 36, H2498–H2507.

Yu, D.Y., Su, E.N., Cringle, S.J., Alder, V.A., Yu, P.K., DeSantis, L.,

1997. Effect of beta blockers and Ca2+ entry blockers on ocular

vessels. In: Drance, S. (Ed.), Glaucoma Ocular Blood Flow and

Drug Treatment. Kugler Publications, Amsterdam.

Yu, D.Y., Su, E.N., Cringle, S.J., Alder, V.A., Yu, P.K., DeSantis, L.,

1998. Effect of betaxolol, timolol and nimodipine on human and

pig retinal arterioles. Exp. Eye Res. 67, 73–81.

Yu, D.-Y., Su, E.-N., Cringle, S.J., Alder, V.A., Yu, P.K., DeSantis,

L., 1999a. A review and comparison of the systemic and ocular

vascular roles of the antiglaucoma agent b-adrenergic antagonists

and the calcium entry blockers. Surv Ophthalmol. 43, S214–S222.

Yu, D.Y., Cringle, S.J., Alder, V.A., Su, E.N., 1999b. Intraretinal

oxygen distribution in the rat with graded systemic hyperoxia and

hypercapnia. Invest. Ophthalmol. Vis. Sci. 40, 2082–2087.

Yu, D.Y., Cringle, S.J., Su, E.N., Yu, P.K., 2000a. Retinal oxygen

distribution in a rat model of retinal ischemia. Invest. Ophthalmol.

Vis. Sci. 41, S19.

Yu, P.K., Yu, D.Y., Cringle, S.J., Su, E.N., 2000b. Acetylcholine-

induced relaxation in rat ocular vasculature. J. Ocular Pharmacol.

and Therap. 16, 447–454.

Yu, D.Y., Cringle, S.J., Su, E.N., Yu, P.K., Jerums, G., Cooper, M.E.,

2001a. Pathogenesis and intervention strategies in diabetic retino-

pathy. Clin. Exp. Ophthalmol. 29, 164–1919166.

Yu, D.Y., Su, E.N., Cringle, S.J., Schoch, C., Percicot, C.P., Lambrou,

G.N., 2001b. Comparison of the vasoactive effects of the docosaoid

unoprostone and selected prostanoids on isolated perfused retinal

arterioles. Invest. Ophthalmol. Vis. Sci. 42, 1499–1504.

Yu, P.K., Yu, D.Y., Cringle, S.J., Su, E.N., 2001c. Tetrahydrobiopter-

in reverses the impairment of actylcholine-induced vasodilatation

in diabetic ocular microvasculature. J. Ocular Pharmacol. and

Therap. 17, 123–130.


Date post:	20-Feb-2023
Category:	Documents
Upload:	uwa
View:	0 times
Download:	0 times

Isolated preparations of ocular vasculature and their applications in ophthalmic research

Documents