Eye (1990) 4, 243-248.

The Choroidal Circulation and Metabolism-An Overview

WAL�ACE S. FOULDS, CBE Glasgow

In most species, including man, the outer ret­ measurement from a divided vortex vein or ina is totally dependent on the blood circula­ embolisation by radioactive spheres. I We, in tion in the for its metabolic needs. In Glasgow, using a radioactive Krypton wash­ some species, such as the which lacks out technique showed that the choridal blood retinal vessels, the choroid supports the flowis approximately ten times cerebral blood metabolic requirements of the whole thick­ flow when considered on the basis of blood ness of the . This situation is also true of flow per unit mass of tissue.2 the macular retina in including man. The high blood flowin the choroid is at least There are obvious visual benefits in having a in part related to low resistance in the chor­ retina of reduced vascularity, not the least of oidal which are very wide bore as which is the unimpeded transmission of the compared, for example, with retinal capillar­ visual to the receptor cells of the retina. ies (Fig. 1). Additionally there is a very gener­ A disadvantage is the relative remoteness of ous arterial blood supply via the ciliary the outer retina from its source of nourish­ which arise directly from the ophthal­ ment and the interpolation of various struc­ mic . The work of Hayreh on the seg­ tures between the choroid and the retina, any mental supply of blood to the choroid3 has had of which may become abnormal and interfere a profound influenceon our understanding of with the supply of nutriments to the retina or choroidal vascular insufficiency. The question with the elimination of the waste products of of homeostasis in the choroidal circulation is retinal metabolism. still debatable. Wilson and Strang4 in Glasgow An 'immediately obvious peculiarity of the choroidal circulation is the fact that it appears to be greatly in excess of even the high meta­ bolic requirements of the retina. The extrac­ tion· of metabolites from the choroidal circulation is incomplete and blood in the vor­ tex veins leaves the highly oxgenated with only a small proportion of its haemoglobin in the reduced form. There is an analogy in the kidney for blood in the renal veins like that in the vortex veins is bright red and highly oxy­ genated. Blood flow in the choroid has been measured by many workers using a variety of Fig. I. Scanning electronmicroscopy of a resin cast specimen of the retina and choroid to show the techniques. Anders Bill in Sweden has been considerable disparity in size which exists between active in this area for many years using direct retinal capillaries and the underlying chorio-capillaris.

From: Tennent Institute of , University of Glasgow, Western Infirmary, 38 Church Street, Glasgow G 11 6NT. Correspondence to: Wallace S Foulds, CBE, Ross Hall AMI Hospital, 221 Crookston Road, Glasgow. 244 WALLACE S. FOULDS

Fig. 23. Fig.2c.

Fig.2b. Fig.2d.

Fig. 2. (a) photograph and (b) Fluorescein angiogram of a patient suffering temporary occlusion of the following an orbital to show considerable dilatation of patent cilia-retinal vessels (c) Fundus phlltograph and (d) Fluorescein angiogram after spontaneous restoration of retinal circulation to show return of cilia-retinal system to its normal undilated state. (Case previously published.by Chawla J, Trans Ophthalmol Soc UK 1972; 92: 777-84) THE CHOROIDAL CIRCULATION AND RETINAL METABOLISM 245

may be needed in relation to the mechanisms involved in glomerular filtration and tubular re-absorption, is the high choroidal blood flow related to fluid exchange across the pos­ terior blood ocular barrier and, particularly, to the large outward movement of fluidwhich has been shown to exist from the vitreous cav­ ity to the choroid?6 Lastly, is the high flow related to the relative separation of the outer retina from its blood supply and the need to maintain a high concentration gradient of required metabolites (and a low concentra­ Fig. 3. Transmission eiectronmicroscope picture of the chorio-capillaris to show fenestrations. tion of waste products of metabolism) to aid diffusional exchange? many years ago showed that choroidal cir­ A striking feature of the choriocapillaris is culation could be maintained relatively con­ the presence of fenestrations which have been stant over a wide range of perfusion pressures thought to be necessary to aid the supply of achieved by alteration of lOP and of systemic metabolites to the retina (Fig. 3). Many highly blood pressure. metabolising tissues, however, do not need The choroidal circulation is exquisitely sen­ such a structural modification in their capil­ sitive to carbon dioxide tension in the circu­ lary walls. We know that the capillaries of the lating blood and probably also to the acid choroid, unlike those of the retina, readily products of metabolism, all of which cause leak plasma protein as demonstrated on flu­ choroidal vasodilatation and increased chor­ orescein angiography. Plasma proteins are oidal blood flow. This is well exemplified in a excluded from the eye in the interests of opti­ patient with a temporary obstruction of the cal clarity and the leakiness of the choroidal central retinal artery who showed a gross dila­ capillaries demands the presence of a diffu­ tation of a cilio-retinal artery while the retina sion barrier between the choroid and the ret­ was ischaemic but whose cilio-retinal system ina. It seems likely that the fenestrations in returned to normal' dimensions when blood the choroidal capillaries are needed to allow flow was re-established in the central retinal egress of plasma albumen and, more particu­ artery (Fig. 2). larly, binding protein (RBP) so that Although there must be good reasons as to the all-important vitamen A can reach the ret­ why we need such a high blood flow in the ina. An important role of the retinal choroid it is interesting that not only does it (RPE) is therefore selectively to seem surplus to retinal metabolic require­ transfer retinol from its carrier protein and ments but additionally the retina unlike the arrange for its transfer to the retina while at can withstand quite prolonged periods the same time excluding from the interior of of total ischaemia being capable of surprising the eye plasma albumin and proteins of a size morPhological and functional recovery after similar to RBP which would otherwise cloud periods of total ischaemia lasting up to one vision. hour. 5 This prompts us to ask why we need To reach the retina metabolites from the such a large flowof blood in the choroid. Is it choriocapillaris must pass progressively to cool the retina by removing heat generated through Bruch's membrane, the RPE and the during visual transduction? Is it, as recently interphotoreceptor space (IPS). Of these the suggested by Damato in my Department, a RPE is obviously of the most crucial impor­ mechanism to keep the retina warm and func­ tance. There is now ample evidence that the tioning by preventing the depression of tem­ RPE actively and selectively transfers mol­ perature which might be associated with the ecules to and from the retina and, impor­ exposed position of the eye in relation to the tantly, is involved in the continuous outward external environment? Returning to the anal­ movement from the vitreous cavity and sub­ ogy of the kidney where a large blood flow retinal space of water and other molecules 246 WALLACE S. FOULDS

Fig. 4a & 4b. Fig. 4c & 4d.

Fig. 4. Breakdown in the posterior blood ocular barrier resulting from poisoning by various doses of sodium iodate and demonstrated by the diffusion of intravenously injected horseradish peroxidase through the whole thickness of the retina via the damaged RPE (experiments carried out in rabbit by McKechnie and Foulds (unpublished)). a & c, control tissue b & d, effects of intravenous sodium iodate. which, if allowed to accumulate, would lead tracers such as horseradish peroxidase (HRP) to retinal separation. (See Steinberg for (Fig. 4). If the transfer of water from the review).7 potential subretinal space to the choroid is Ionic movement across the RPE apical and mediated by active work on the part of the basal membranes is reflected in the standing RPE one might expect that poisoning the potential of the eye and the 'c' wave of the RPE would reduce this outward movement of ERG, both of which can be increased in water. Surprisingly the reverse seems to be amplitude by the administration of substances the case, poisoning by sodium iodate leading which stimulate ionic transfer such as sodium to an increased outward movement of water azide,8,9 or reduced or even abolished by rather than a decrease as shown by Marmor poisons which selectively inhibit RPE func­ using his localised tech· tion, notably sodium iodate as demonstrated niquell and by ourselves using tritiated water many years ago by Noell. 10 as a tracer. One has to postulate that the out· Not unexpectedly, poisoning of the RPE by ward movement of water from the subretinal sodium iodate leads to severe metabolic upset space to the choroid is only partly dependent and structural change in the related outer ret­ on active transport by the RPE and that other ina and to a breakdown in the posterior blood physical factors not requiring a functionally ocular barrier as demonstrated by the use of intact RPE are also at work. These factors mE CHOROIDAL CIRCULATION AND RETINAL METABOLISM 247 may include the colloid osmotic pressure of membrane to the outward movement of mol­ protein in the extravascular tissue spaces of ecules from the RPE to the chorotid has 2 the choroid as originally suggested by Bill.1 received a great deal of attention, little has Recently the subretinal space itself has been said about the effect changes in the per­ become an area of great interest and, particu­ meability of Bruch's membrane may have on larly, the interphotoreceptor matrix which fills the inward diffusion of metabolites from the it. This matrix is formed of a mixture of com­ choroid to the retina. It may be that normally plex glycoproteins and glycosaminoglycans such metabolites are delivered to the RPE in and contains an important transport protein excess and that even a moderate reduction in for (interphotoreceptor matrix ret­ supply to the RPE related to increased resis­ inol binding protein, IRBP). The matrix is tance in Bruch's membrane might not influ­ structurally inhomogeneous being highly gly­ ence the ability of the RPE to transfer cated in the vicinity of cone outer segments. 13 sufficient to the retina to meet its metabolic The matrix may have adhesive properties requirements. important in retinal apposition.? It obviously The whole area of metabolic exchange interposes a diffusion barrier between the across the posterior blood ocular barrier to choroid and the retina and at present its role the retina is currently of great interest, not in retinal metabolism remains conjectural. only because of an increased of The highly polymerised glycosaminoglycans many choroidal vascular diseases and dis­ and glycoproteins of the matrix, however, are orders of the retinal pigment epithelium water retentive so maintain an interval which may affect it but also in relation to our between the apical RPE and the outer retina understanding of the mechanism of action of perhaps in this way helping to avoid unwanted photocoagulation in such conditions as dia­ local concentrations of metabolites or waste betic retinal neovascularisation. Undoubt­ products. 14 edly energy delivered in panretinal An essential aspect of retinal metabolism photocoagulation is absorbed at the level of is, of course, that related to the synthesis and the RPE and choroid and in ways which we do phagocytic destruction of outer segments; not yet understand leads to metabolic changes processes that require a large turnover of in the retina which are associated with regres­ phospholipid for plasma membrane synthesis sion of neovascular complexes at a distance and of retinal for the synthesis of visual pig­ from the area treated. Whether these effects ment. There is evidence that some compon­ are the result of altered blood flow and oxy­ ents of the matrix may be important in this genation in the choroid and retina or process through their inhibitory effect on mediated through the release of metabolically phagocytosis of receptor outer segments active molecules from damaged RPE is a sub­ (ROS) by the RPE.15 ject of intense debate. The supply of lipid needed for the continu­ ous synthesis of ROS plasma membrane has The scanning and transmission electron micrographs were prepared by Mrs. Dorothy Aitken and the flu­ to take place not only through the RPE but orescein angiograms by Mrs. Anne Currie to both of across Bruch's membrane which in health whom I am grateful. does pot appear to provide any barrier to the movement of even large molecules. In the elderly, however, there may be an accumu­ lation of lipid containing waste products of References 1 AIm A and Bill A. Ocular and optic blood flow phagocytosis in Bruch's membrane and these at normar and increased in may impede the outward movement of water monkeys (Macaca lrus) a study with radioactively from the RPE to the choroid and be of rel­ labelled microspheres including flow determina­ evance to the genesis of RPE detachment as tions in brain and other tissues. Exp Eye Res 1973, 15: 15. suggested on theoretical grounds by Damato 2 Wilson TM, Strang R, Wallace J, Horton PW, John­ in 198516 and as demonstrated by and ston NF. The measurement of the choroidal blood Marshall. I? It is interesting that although the flow in the rabbit using Krypton 85. Exp Eye Res effects of increased resistance in Bruch's 1973, 16: 421-55. 248 WALLACE S. FOULDS

3 Hayreh SS. The segmental of the choroidal 11 Marmor MF, Abdul Rahim AS, Cohen DS. The vasculature. Br J Ophthalmol1975,59: 631-48. effect of matabolic inhibitors on retinal adhesion 4 Wilson TM, MacKenzie ET, Strang R. Factors influ­ and subretinal fluid absorption. Invest Ophthal­ encing ocular blood flow. In: Vision Circulation, mol Vis Sci 1980; 19: 893-903. 12 Cant JS (Ed) St Louis 1976. The C V Mosby Co Bill A. Some aspects of tissue fluid dynamics in the pp45-49. eye. In: Vision and Circulation, Cant JS (Ed) St 5 Johnson NF. Electronmicroscopy of acute retinal Louis 1976. The C V Mosby Co 333-8. ischaemia in the rabbit and a study of the pattern 13 Johnson LV and Hageman GS. Enzymatic charac­ of recovery. In: Vision and Circulation, JS Cant terisation of peanut agglutinin binding compon­ (Ed) St Louis. The C V Mosby Co 1976 pp 79-92. ents in the retinal interphotoreceptor matrix. Exp 6 Moseley H, Foulds WS, Allan D, Kyle PM. Routes Eye Res 1987, 44:553-65. of clearance of radioactive water from the rabbit 14 Foulds WS. Is your vitreous really necessary? The vitreous. Br J Ophthalmol1984,68: 145-51. role of the vitreous in the eye with particular 7 Steinberg RH. Research update: report from a reference to retinal attachment, detachment and workshop on of retinal detachment. the mode of action of vitreous substitutes. Eye Exp Eye Res 1986, 43: 696--706. 1987,1: 641-64. 8 Noell WK. Azide sensitive potential differences 15 Gregory CY, Converse CA, Foulds WS. Effect of across the eye bulb. Am J Physiol 1952, 170: glycoconjugates on rod outer segment phagocyto­ 217-38. sis by retinal pigment epithelial explants in vitro 9 Foulds WS and Ikeda H. The effects of retinal assessed by a specific double radioimmunoassay detachment on the induced and resting ocular procedure. Current Eye Res. (In Press). potentials in the rabbit. Invest Ophthalmol Vis Sci 16 Damato BE. The subretinal to neovascular­ 1966,5: 93-108. isation in Mackenzie symposium on retinal neo­ 10 Noell WK. Studies on the and the vascularisation. Glasgow 1985 (unpublished). metabolism of the retina. USAF School of Avi­ 17 Bird AC and Marshall J. Retinal pigment epithelial ation . Project No.211120110004. Report detachments in the elderly. Trans Ophthalmol Soc No.1, 122 1953. UK 1986,105: 674-820.