SCLERA

The sclera accounts for approximately 93% of the outer coat of the eye. Anteriorly, it is continuous with the cornea at the corneoscleral junction (limbus) (Fig. 40.1). It is punctured by a number of foramina containing nerves and blood vessels, most notably the optic foramen, which lies 3 mm medial to the midline and 1 mm below the horizontal and houses the optic nerve. Smaller openings contain anterior ciliary arteries which penetrate anteriorly, vortex veins which cross the sclera equatorially, and the long and short ciliary nerves and arteries which enter posteriorly. The sclera is thickest posteriorly (1–1.35 mm), decreasing gradually towards the equator to reach a minimum of 300 μm under the tendons of the recti. Going forwards from the insertions of the recti, it gradually increases in thickness once more to 800 μm at the limbus.

The external surface of the sclera is covered by a delicate episcleral lamina of loose fibro-vascular tissue, which contains sparse blood vessels and is in contact with the inner surface of the fascial sheath of the eyeball (see p. 657). Anteriorly the external scleral surface is covered by conjunctiva which is reflected onto it from the deep surfaces of the eyelids. The scleral internal surface adjacent to the choroid is attached to it by a delicate fibrous layer, the suprachoroid lamina, which contains numerous fibroblasts and melanocytes. Anteriorly, the inner sclera is attached to the ciliary body by the lamina supraciliaris. Posteriorly, the sclera is pierced by the optic nerve. Here the outer half of the sclera turns back to become continuous with the dura mater, while the inner half is modified to form a perforated plate, the lamina cribrosa sclerae. The optic nerve fascicles pass through these minute orifices, while the central retinal artery and vein pass through a larger, central aperture. The lamina cribrosa sclerae is the weakest part of the sclera and bulges outwards (a cupped disc) when intraocular pressure is raised chronically as in glaucoma (see p. 686).

Like the cornea, the scleral stroma is composed mainly of densely packed collagen embedded in a matrix of proteoglycans, which are mixed with occasional elastic fibres and fibroblasts. However, in contrast to the cornea, scleral collagen fibrils show a large variation in diameter and spacing, and the lamellae branch and interlace extensively. This arrangement of fibres results in increased light scatter, which is responsible for the opaque dull white appearance of the sclera, and also imparts a high tensile strength to the sclera to resist the pull of the extraocular muscles and contain the intraocular pressure. Collagen fibre bundles are arranged circumferentially around the optic disc and the orifices of the lamina cribrosa. The fibres of the tendons of the recti intersect scleral fibres at right angles at their attachments, and then interlace deeper in the sclera. Collagen fibres of the scleral spur are orientated circularly, and there is an increased incidence of elastic fibres here. Although the sclera acts as a conduit for blood vessels, scleral vessels are few and mainly disposed in the episcleral lamina, especially close to the limbus. Its nerve supply is surprisingly rich, accounting for the intense pain associated with scleral inflammation (Watson & Young 2004).

Filtration angle and aqueous drainage

Aqueous humour is produced by the ciliary epithelium, passes through the pupil and circulates within the anterior chamber supplying the avascular cornea and lens with nutrients. It drains from the eye mainly through the trabecular meshwork into the canal of Schlemm at the iridocorneal (filtration) angle (Figs 40.2, 40.3), formed between the posterior aspect of the corneoscleral limbus and the periphery of the iris.

Fig. 40.2 Meridional section through the anterior eye highlighting the ciliary body and drainage apparatus.

Fig. 40.3 Tissues bordering the anterior chamber angle. The trabecular meshwork is an anulus of tissue spanning the angle: its meshes are shown in transverse section opposite the canal of Schlemm and partly attached to the scleral spur. A single iris process bridges the angle, connecting the trabecular meshwork to the anterior tissues of the iris.

The trabecular meshwork consists of loose trabecular tissue which is continuous anteriorly with Descemet’s membrane and the endothelium of the cornea. Aqueous humour filters from the anterior chamber through interconnected spaces among trabecular fibres. Most of these fibres are attached to the anterior, external aspect of the scleral spur. Of the remainder, most are continuous with longitudinal fibres of the ciliary muscle, some of which attach to the posterior internal aspect of the scleral spur. The trabecular meshwork provides sufficient resistance to aqueous humour outflow to generate an intraocular pressure of approximately 15 mmHg. It also acts as a filter and has the capacity to phagocytose particulate matter, although overloading may contribute to the pathogenesis of various forms of obstructive secondary glaucomas.

The canal of Schlemm (sinus venosus sclerae) is an anular endothelial canal located near the internal surface of the sclera close to the limbus. In section, the canal appears as an oval cleft, with an outer wall that grooves the sclera. Posteriorly, the cleft extends as far as a rim of scleral tissue, the scleral spur, which in section forms a triangle with its apex directed forwards (Figs 40.2, 40.3). The canal of Schlemm may be double or multiple in part of its course, and its walls are constructed of a continuous single thin endothelial layer. Passage of aqueous humour to the canal probably occurs via giant pinocytotic vacuoles, which form on the inner face of the endothelium and discharge into the canal at the outer face, and through intercellular clefts. Aqueous humour passes through a plexus of fine intrascleral vessels which connect the canal of Schlemm with anterior ciliary veins (Fig. 40.4). Normally the canal does not contain blood: pressure gradients prevent the reflux of blood even though the channels between the canal and veins have no valves. However, in venous congestion, blood may enter the canal: the continuous endothelial outer wall of the canal prevents further reflux.

Fig. 40.4 Tissues of the filtration angle in section. The approximate borders of the limbal region are indicated by the solid lines. Aqueous humour drains from the anterior chamber at T through the trabecular meshwork into the canal of Schlemm from where it flows through collector channels either to the intrascleral venous plexus, where it mixes with blood from the ciliary venous plexus, or directly to the surface tissues of the eye as an aqueous vein. The aqueous vein is shown joining an episcleral vein at the surface to form a laminated vessel. Aqueous and blood drain posteriorly in episcleral veins which in turn join anterior ciliary veins.

An alternative route for aqueous outflow, the uveo-scleral pathway, has been described. Since there is no epithelial barrier between the anterior chamber and the ciliary body, aqueous humour is able to enter the loose connective tissue in front of the ciliary muscle and pass between the muscle fibres into the supraciliary and suprachoroidal spaces, where potentially it can be absorbed by vessels that drain the uvea.

CORNEA

The avascular cornea is the anterior transparent part of outer coat. Convex anteriorly, it projects from the sclera as a dome-shaped elevation with an area of 1.1 cm², forming approximately 7% of the external tunic area. Since it is more curved (average radius, r = 7.8 mm) than the sclera (r = 11.5 mm), a slight sulcus sclerae marks the corneoscleral junction (limbus). The cornea is approximately 670 μm thick close to the corneoscleral junction, and 520 μm at its centre. At the nasal and temporal limbus the transition from cornea to sclera occurs in a line that is approximately perpendicular to the cornea: this transition occurs more obliquely superiorly and inferiorly, with the sclera overlapping the cornea to a greater extent anteriorly. Consequently, when viewed from in front, the corneal perimeter is slightly elliptical, as its horizontal diameter (11.7 mm) is a little greater than its vertical (10.6 mm). Its posterior perimeter is circular (diameter 11.7 mm).

Microscopically, the cornea consists of five layers, namely, corneal epithelium, anterior limiting lamina (Bowman’s layer), substantia propria (stroma), posterior limiting lamina (Descemet’s membrane) and endothelium, arranged anteroposteriorly (Fig. 40.5A).

Fig. 40.5 A, Low power light micrograph of the five layers of the cornea. B, High power light micrograph of the corneal epithelium and Bowman’s layer. C, High power light micrograph of the corneal endothelium, Descemet’s layer and part of the stroma. The darkly staining cells in the stroma are keratocytes.

Substantia propria (stroma)

The substantia propria is approximately 500 μm thick and forms the bulk of the cornea. It is a compact and transparent layer, composed of 200–250 sequential lamellae, each made up of fine parallel collagen fibrils mainly of type I collagen. Flat dendritic interconnecting fibroblasts (keratocytes) form a coarse mesh between the lamellae. Alternate lamellae are typically orientated at large angles to each other (Fig. 40.6). X-ray diffraction studies indicate that they run along two preferred directions, superior-inferior and nasal-temporal, to account for the additional tensile stress exerted by the recti along these meridians (Boote et al 2005). Each lamella is approximately 2 μm thick and of variable width (10–250 μm, or, rarely, more). In the central cornea fibrils within the lamellae have a similar diameter of approximately 31 nm. This increases slightly with age and approaching the limbus. The small size of the fibrils (much smaller than the wavelength of light), along with the regularity of their spacing (maintained by collagen: proteoglycan interactions), and the careful control of corneal hydration, are the principal factors which determine stromal transparency. Light scattered by the collagen fibrils is eliminated by destructive interference in all directions other than forwards.

Fig. 40.6 Stroma of the human cornea showing the geometric precision of the alternation in direction of adjacent layers of fibres.

(By courtesy of John Marshall, St Thomas’ Hospital, London).

CORNEOSCLERAL JUNCTION (LIMBUS)

The limbus marks the transitional zone between the cornea and sclera. Here the corneal epithelium merges with the epithelium of the conjunctiva, which thickens (up to 12 cells). Bowman’s layer terminates and the corneal stromal collagen loses its regularity. As the cornea contains neither blood nor lymphatic vessels, the capillaries of the conjunctiva and episclera end in loops near the limbus. Internally, Descemet’s membrane disperses into the fibres of trabecular meshwork and the corneal endothelium is continuous with that covering the trabeculae.

The cornea does not possess epithelial stem cells and cell replacement depends on the centripetal migration of cells from the edges of the cornea which are the progeny of mitotic limbal stem cells.

CHOROID

The choroid is a thin (60–160 μm, thickest behind macula) highly vascular, pigmented layer which lines almost five-sixths of the eye posteriorly. It is pierced by the optic nerve where it is firmly adherent to the sclera. Elsewhere its external surface is only loosely connected to the sclera by the suprachoroid layer (lamina fusca). Internally it is attached to the retinal pigment epithelium, and at the optic disc is continuous with the pia-arachnoid tissues around the optic nerve.

Four layers can be identified in transverse section (Fig. 40.7, see also Fig. 40.23): suprachoroid, vascular stroma, choriocapillaris and lamina vitrea (Bruch’s membrane).

Fig. 40.7 Drawing of a histological section through the choroid close to the posterior pole.

Fig. 40.23 Transverse section of the retina and choroid. The ten layers of the retina are: 1. Pigment epithelial layer. 2. Rod and cone layer. 3. External limiting membrane. 4. Outer nuclear layer. 5. Outer plexiform layer. 6. Inner nuclear layer. 7. Inner plexiform layer. 8. Ganglion cell layer. 9. Nerve fibre layer. 10. Internal limiting membrane. CC, choriocapillaris; SC, suprachoroid.

Vessel layer (stroma)

Internal to the suprachoroid lies a layer composed mainly of arteries and veins, but also some loose connective tissue containing scattered pigment cells. These melanocytes limit the passage of light through the sclera to the retina. More importantly, like the retinal pigment epithelium (see p. 689), they also absorb light traversing the retina that is not absorbed by the photoreceptors, so preventing internal reflection.

The blood supply of the choroid, as well as the rest of the uvea, is summarized in Figure 40.10. Short posterior ciliary arteries enter the eye through the sclera near the optic disc and supply the posterior choroid. These vessels branch and gradually decrease in size as they approach the retinal border. The choroidal stroma can be divided into layers based on the change in the calibre of these vessels; an outer layer of larger vessels (Haller’s layer), an inner layer of smaller vessels (Sattler’s layer) which eventually give rise to the choriocapillaris (Fig. 40.7). Long posterior ciliary arteries and recurrent branches of anterior ciliary arteries supply the anterior part of the choriocapillaris.

Fig. 40.10 The vasculature of the uveal tract. The long posterior ciliary arteries, one of which is visible (A), branch at the ora serrata (b) and feed the capillaries of the anterior part of the choroid. Short posterior ciliary arteries (C) divide rapidly to form the posterior part of the choriocapillaris. Anterior ciliary arteries (D) send recurrent branches to the choriocapillaris (e) and anterior rami to the major arterial circle (f). Branches from the circle extend into the iris (g) and to the limbus. Branches of the short posterior ciliary arteries (C) form an anastomotic circle (of Zinn) (h) round the optic disc, and twigs from this (i) join an arterial network on the optic nerve. The vortex veins (J) are formed by the junctions (k) of suprachoroidal tributaries (l). Smaller tributaries are also shown (m, n). The veins draining the scleral venous sinus (o) join anterior ciliary veins and vorticose tributaries.

(By permission from Hogan MJ, Alvarado JA, Weddell JE 1971 Histology of the Human Eye. Philadelphia: WB Saunders.)

Veins within the choroid converge spirally onto four, or very occasionally more, principal vortex veins. These pierce the sclera behind the equator to reach tributaries of the ophthalmic veins.

The vessels of the choroid have a rich autonomic vasomotor supply.

The blood flow through the choroid is high, a feature probably associated with an intraocular pressure of 15–20 mm Hg, which means that a venous pressure above 20 mm Hg is required to maintain circulation. The choroidal perfusion rate exceeds that required to supply nutrients and may serve to cool the retina during exposure to bright light.

CILIARY BODY

The ciliary body serves to anchor the lens via suspensory ligaments and by the contraction of its smooth muscle changes the refractive power of the lens (accommodation). Its anterior internal surface is also the source of aqueous humour, while posteriorly its inner surface is contiguous with the vitreous humour, and secretes several of its components. The anterior and the long ciliary arteries meet in the ciliary body and the major nerves to all the anterior tissues of the eyeball pass through it.

Externally, the ciliary body may be represented by a line which extends from approximately 1.5 mm posterior to the limbus of the cornea (corresponding also to the scleral spur) to a line approximately 7.5–8 mm posterior to this on the temporal side, and 6.5–7 mm on the nasal side. The ciliary body is thus slightly eccentric. It projects posteriorly from the scleral spur, which is its attachment, with a meridional width varying from 5.5 to 6.5 mm. Internally, it exhibits a posteriorly crenated or scalloped periphery, the ora serrata, where it is continuous with the choroid and retina (Figs 40.8, 40.9). Anteriorly, it is confluent with the periphery of the iris, and externally bounds the iridocorneal angle of the anterior chamber.

Fig. 40.8 The ciliary region seen from the ocular interior. Above is the periphery of the lens, attached by the fibres of the zonule (suspensory ligament) to the processes of the corona ciliaris (pars plicata) of the ciliary body (a). The orbiculus ciliaris or pars plana ciliaris (b) has a scalloped boundary, the ora serrata (c), which separates it from the retina (d). Flanking the ‘bays’ (e) of this are the dentate processes (f), with which linear ridges or striae (g) are continuous: the striae extend forwards between the main ciliary processes, providing an attachment for the longer zonular fibres. The posterior aspect of the iris shows radial (h) and circumferential (i) sulci.

(By permission from Hogan MJ, Alvarado JA, Weddell JE 1971 Histology of the Human Eye. Philadelphia: WB Saunders.)

Fig. 40.9 The posterior aspect of the anterior half of the eye showing the termination of the neural retina at the ora serrata and the ciliary body. The lens has retained sufficient transparency to reveal the iris border: the crenated perimeter of the lens is due to tension imposed by the attached suspensory ligaments (unseen).

In cross section the ciliary body is composed of four layers (from internal to external), namely a double layer of epithelial cells, the stroma, ciliary muscle, and a supraciliary layer (Fig. 40.2).

Meridionally, the ciliary body can be divided into two parts (Figs 40.2, 40.9). Anteriorly, the ridged anterior pars plicata (corona ciliaris) surrounds the base of the iris and accounts for about ⅓ rd of the ciliary body. Posteriorly, the relatively smooth and thin pars plana (orbicularis ciliaris) lies adjacent to the ora serrata, and forms a convenient access point for instruments during vitreo-retinal surgery and for intraocular injection. The whole ciliary body is covered by a double epithelium, in which the inner layer is unpigmented, while the outer layer contains melanin. At the ora serrata the retinal pigment epithelium is continuous with the outer pigmented epithelium of the pars plana, while the neural retina is replaced by inner unpigmented ciliary epithelium. Anteriorly, this double epithelium continues over the pars plicata until it merges with the double epithelium on the posterior surface of the iris (where the inner layer of cells also accumulates melanin) (see p. 682). The anterior pars plicata is ridged meridionally by 70–80 ciliary processes radiating from the base of the iris (Figs 40.2, 40.9). In the young eye these process are approximately parallel-sided structures, but in the adult their flanks become less regular and appear thicker. A minor ridge, or ciliary plica, lies in the valley between most of the processes. The crests of the processes are less pigmented, giving them the appearance of white (or light) striae, from which the name ciliary is derived. Fibres of the zonule (the suspensory ligaments of the lens) extend into the valleys. They pass beyond the ciliary processes to fuse with the basal lamina of the superficial epithelial layer of pars plana. Their sites of attachment are marked by striae which pass back from the valleys of the pars plicata, across the pars plana, almost as far as the apices of the dentate processes of the ora serrata (Fig. 40.8).

Ciliary stroma

The ciliary stroma is composed largely of loose bundles of collagen, which form a considerable mass between the ciliary muscle and overlying processes, and extend into both of them. This inner layer of connective tissue contains numerous larger branches of the ciliary vessels. A dense reticulum of large (up to 35 μm diameter) fenestrated capillaries is concentrated in the ciliary processes. This facilitates the passage of substances from the blood plasma during aqueous production (see p. 685). Anteriorly, near the periphery of the iris, the major arterial circle (Fig. 40.10; see also Fig. 40.14) is formed chiefly by long posterior ciliary branches of the ophthalmic artery. These enter the eye some distance behind the ocular equator and pass between the choroid and sclera to the ciliary body. Ciliary veins, also draining the iris, pass posteriorly to join the vortex veins of the choroid.

Fig. 40.14 Meridional section of an iris. The posterior surface is lined with a double epithelium (E, E’): the arrow points to the processes of the anterior layer of the epithelium forming the dilator pupillae muscle. ST, Stroma. C, Collarette. RT, Root. A, Anterior border layer. PF, Folds in anterior surface towards the root. R, Pupillary ruff. S, Sphincter pupillae muscle. P, Pupillary region of the iris. The iris is in a contracted state, as shown by the thickened epithelium opposite the dilator muscle and thinning opposite the sphincter, and by the shortness of the pupillary zone. CM, Ciliary muscle. CP, Ciliary processes. I, Major iridic circle.

Ciliary muscle

Ciliary muscle is composed of smooth muscle cells, most of which are attached to the scleral spur and arranged in three different orientations. The outermost fibres are meridional or longitudinal, and pass posteriorly into the stroma of the choroid. The innermost fibres swerve acutely from the spur to run circumferentially as a sphincter near the periphery of the lens. Obliquely interconnecting radial fibres run between these two muscular strata, frequently forming an interweaving lattice (Figs 40.2, 40.11).

Fig. 40.11 The ciliary muscle and its components. The meridional or longitudinal (1), radial or oblique (2), and circular or sphincteric (3) layers of muscle fibres are displayed by successive removal towards the ocular interior. The cornea and sclera have been removed, leaving the canal of Schlemm (a), collecting venules (b) and scleral spur (c). The meridional fibres (1) often display acutely angled junctions (d) and terminate in epichoroidal stars (e). The radial fibres meet at obtuse angles (f) and similar junctions, at even wider angles (g), occur in the circular ciliary muscle.

(By permission from Hogan MJ, Alvarado JA, Weddell JE 1971 Histology of the Human Eye. Philadelphia: WB Saunders.)

Accommodation reflex

At rest distant objects are focused on the retina in an emmetropic eye (see p. 687). In order to focus closer objects the dioptric power of the eye has to be increased, which is achieved by increasing the curvature of the lens. At rest the lens is under tension from the zonular ligaments and hence flattened. On accommodation, the ciliary muscle contracts, moving the ciliary body forwards and inwards towards the optic axis. All parts of the muscle act in concert, and tension on the zonular ligaments is relaxed. As the lens is covered by an elastic capsule (see p. 686), once tension on it is released it assumes a more convex shape suitable for focusing closer objects. The radius of curvature of the anterior lens surface changes most during accommodation.

Information from the retina passing to the visual cortex does not constitute the afferent limb of a simple reflex in the usual sense of the term, but permits the visual areas to assess the clarity of objects in the visual field. Cortical efferent information passes to the pretectal area and thence to the Edinger–Westphal nucleus, which contains preganglionic parasympathetic neurones whose axons travel in the oculomotor nerve (Fig. 40.12). Efferent impulses pass in the oculomotor nerve to the orbit where they synapse in the ciliary ganglion. Postganglionic fibres (short ciliary nerves) innervate the ciliary muscle causing it to contract. There is also a sparse sympathetic innervation of ciliary muscle, which has a very limited capacity to relax the muscle.

Fig. 40.12 The neural pathways of the pupillary light reflex (left) and the accommodation reflex (right).

(From Oxford Textbook of Functional Anatomy, Vol 3 Head and Neck, MacKinnon P, Morris J (eds) 1990. By permission of Oxford University Press.)

Accommodation is usually accompanied by constriction of the pupil brought about by contraction of the sphincter pupillae (see p. 682) and convergent eye movements caused by contraction of the medial, superior, and inferior recti (all innervated by the oculomotor nerve). This is the ‘near triad’ which may become disrupted in various diseases.

IRIS

The iris is an adjustable diaphragm around a central aperture (slightly medial to true centre), the pupil. It lies between the cornea and lens and is immersed in aqueous fluid (Fig. 40.1), partially dividing the anterior segment into an anterior chamber, enclosed by the cornea and iris and a posterior chamber, which lies between the iris and the lens anterior to the vitreous. The efficacy of the iris as a light stop is mainly due to a densely pigmented posterior double epithelium. The pupillary aperture is adjusted by the action of two muscles, dilator and sphincter pupillae.

Seen from the front the iris is divided into a large ciliary zone adjacent to the ciliary body and a smaller, inner, pupillary zone (Fig. 40.13). The two regions join at the collarette. The anterior surface of the iris often contains large depressions and at the pupillary margin the posterior pigmented epithelium is visible as the pupillary ruff (Fig. 40.13).

Fig. 40.13 Composite view of the surfaces and internal strata of the iris. In a clockwise direction from above, the pupillary (A) and ciliary (B) zones are shown in successive segments. The first (brown iris) shows the anterior border layer and the openings of crypts (c). In the second segment (blue iris), the layer is much less prominent and the trabeculae of the stroma are more visible. The third segment shows the iridial vessels, including the major arterial circle (e) and the incomplete minor arterial circle (f). The fourth segment shows the muscle stratum, including the sphincter (g) and dilator (h) of the pupil. The everted ‘pupillary ruff’ of the epithelium on the posterior aspect of the iris (d) appears in all segments. The final segment, folded over for pictorial purposes, depicts this aspect of the iris, showing radial folds (i and j) and the adjoining ciliary processes (k).

(By permission from Hogan MJ, Alvarado JA, Weddell JE 1971 Histology of the Human Eye. Philadelphia: WB Saunders.)

In transverse section several subdivisions of the iris are evident (Fig. 40.14). From anterior to posterior they are an anterior border layer, the stroma (which contains the sphincter pupillae), and two pigmented epithelial layers, the most anterior of which contains the dilator pupillae. It is thinnest at its root (approximately 200 μm), where the ciliary body is attached, and thickest at the collarette.

Epithelial layers

The epithelial surface covering the iris posteriorly is a continuation of the bilaminar epithelium of the ciliary body and is formed from the two layers of the optic cup. The pupil, through which this epithelium curves for a short distance on to the anterior surface as the pigment ruff or ‘border’, corresponds to the opening of the optic cup.

The layer of epithelial cells nearest the stroma is, somewhat confusingly, termed the anterior epithelium, although it lies posterior to the stroma. Its cells are pigmented, as are those of the same layer in the ciliary epithelium. They give rise to the dilator pupillae (see below). Vitreal to this stratum is a layer of heavily pigmented cells, the posterior epithelium, which is continuous with the inner non-pigmented layer of the ciliary epithelium. Its free surface bears numerous radial ridges which facilitate the movement of aqueous humour from the posterior to the anterior chamber (because the front surface of the lens rests on the iris). Adjacent epithelial cells are extensively joined by various junctions which ensure that the layer can withstand the excursions of the iris during changes in pupillary size (Fig. 40.15).

Fig. 40.15 Posterior structures of the iris. The dilator pupillae is shown in transverse section on the right and in longitudinal section on the left (arrow shows a rarer deeper nerve terminal).

AQUEOUS HUMOUR

Aqueous humour is derived from the plasma within the fenestrated capillaries of the ciliary processes. The major component of aqueous, like plasma, is water and the composition of the two fluids is broadly similar, although they do differ in the concentration of some electrolytes and organic solutes. In the interests of optical clarity, the blood–aqueous barrier also ensures a very low concentration of protein in the aqueous (generally <1 per cent of the level in plasma). Inflammation of the anterior uvea can lead to a breakdown of this barrier and the presence of protein in the aqueous, resulting in light scatter that is manifest clinically as ‘flare’.

The aqueous is actively secreted into the posterior chamber by the epithelium overlying the ciliary processes. It passes around the equator of the lens and flows through the pupil into the anterior chamber, where it circulates as a result of convection currents derived from temperature differences between the cornea and iris. Most aqueous is drained from the eye through the trabecular meshwork into the canal of Schlemm from where it drains into episcleral veins. However, some exits through the ciliary muscle into the supraciliary and suprachoroidal spaces (uveo-scleral pathway).

Any interference with the drainage of aqueous into the canal of Schlemm increases intraocular pressure and leads to glaucoma. Glaucoma is either ‘primary’ or secondary to a specific anomaly or disease of the eye. Primary glaucoma can be either ‘closed angle’, where the filtration angle is narrowed by the proximity of the root of the iris to the cornea, or more commonly ‘open angle’, where aqueous access to angle tissues is unimpeded. In the most common form of primary open angle glaucoma, pathological changes within the trabecular meshwork reduce the facility of aqueous drainage, thus raising intraocular pressure. Sustained raised pressure leads to progressive defects in the visual field, either because of direct mechanical damage to retinal ganglion cell axons (particularly at the optic nerve head), or impairment of the optic nerve head blood supply, or both.

LENS

The lens is a transparent, encapsulated, biconvex body bathed in aqueous humour, which serves to adjust the focus of the eye. Posteriorly, it contacts the hyaloid fossa of the vitreous body. Anteriorly, it forms a ring of contact with the posterior border of the iris, but further away from the axis of the lens the gap between the two increases to form the posterior chamber of the eye (Figs 40.1, 40.16). The lens is encircled by the ciliary processes, and is attached to them by the zonular fibres which issue mainly from the pars plana of the ciliary body. Collectively, the fibres form the zonule which holds the lens in place and transmits the forces which stretch the lens.

Fig. 40.16 Sagittal section through part of the lens and related structures. The cortex is distinguished from the nucleus, which is composed of embryonic, fetal, infantile and adult parts. The capsule is drawn in blue, exaggerated in thickness 100 times and is based on observations from a 35-year-old.

The anterior convexity of the lens is less steep than that of the posterior surface. The central points of these surfaces are the anterior and posterior poles of the lens and a line connecting them is its axis. The marginal circumference of the lens is its equator.

At birth the lens is colourless and transmits all wavelengths from the infrared to the near UV well. However, throughout life, the amount of short-wave radiation transmitted diminishes until in old age the lens takes on an amber tinge as it absorbs visible short-wave radiation, decreasing blue sensitivity in older people. In cataract, the lens gradually becomes opaque.

The dimensions of the lens are optically and clinically important, but change with age as a consequence of continuous growth. Its equatorial diameter at birth is 6.5 mm, increasing rapidly at first, then more slowly to 9.0 mm at 15 years of age, and even more gradually to reach 9.5 mm in the ninth decade. Its axial dimension increases from 3.5–4.0 mm at birth to 4.75–5.0 mm at age 95. Average adult radii of the anterior and posterior surfaces are 10 mm and 6 mm respectively. These radii reduce throughout life, the anterior surface increasing most in curvature as the lens thickens, so that in old age the lens pushes the iris forward slightly.

The lens consists of three layers (Fig. 40.17). Most of the lens is composed of elongated cells known as lens fibres. The anterior surface, as far as the equator, is covered by a layer of epithelial cells and the whole is surrounded by the lens capsule. The lens is avascular and devoid of nerve fibres or other structures which might affect its transparency. Its surface forms a very effective barrier against invasion by cells or elements of the immune system, and so creates an immunologically sequestered environment. It is unique because it retains all the cells formed throughout its life.

Fig. 40.17 Section through the anterior layers of the lens. The thin capsule covers the single row of epithelial cells and the lens substance is composed of regularly stacked younger fibres and more densely stained and complex deeper fibres.

Lens fibres

The inwardly displaced epithelial cells elongate further in both an anterior and posterior direction, becoming up to 12 mm long, depending on age and position in the lens. Fibres near the surface at the equator are nucleated; the nuclei form a short S-shaped bow which extends inwards from the surface. The deeper fibres lose their nuclei and most other organelles.

Variations in lens fibre structure and composition make it possible to distinguish a softer cortical zone, representing younger fibres and a firmer central part, representing the older nucleus (Fig. 40.16). The nucleus can be further subdivided into layers representing the age at which the fibres within them were formed.

In cross-section, individual fibres are flattened hexagons measuring approximately 10 μm by 2 μm (Fig. 40.17). They are tightly packed, and fibres are firmly attached to their neighbours by a variety of mechanical junctions (Fig. 40.18). Lens fibres are also in contact through desmosomes and numerous gap junctions.

Fig. 40.18 The mechanical junctions between lens fibres. There are no junctions immediately below the epithelium. Ball and socket junctions develop in deeper layers: these are subsequently eliminated towards the nucleus, where tongue and groove junctions gradually deepen. Angle joints are present at most levels.

All lens fibres cross the equator (or the plane passing through it) and terminate on both the anterior and posterior lens surfaces at lens ‘sutures’. These radiate out from the poles towards the equator and represent lines of linearly registered interlocking junctions between terminating lens fibres. In fetuses, the sutures on the anterior surface of the lens form a triradiate pattern centred on the anterior pole resembling the limbs of an upright letter Y (Fig. 40.19A). Posteriorly, the sutural configuration is similar but inverted. The sutures increase in number and complexity as a consequence of lens growth and other changes in the arrangement of lens fibres (Fig. 40.19B). Fibres which start near the central axis of the lens anteriorly, terminate posteriorly on a suture near the periphery, and vice versa.

Fig. 40.19 The structure of the fetal (A) and adult (B) human lens, showing the major details of arrangement of the lens fibres. The anterior (a) and posterior (b) triradiate sutures are shown in the fetal lens. Fibres pass from the apex of an arm of one suture to the angle between two arms at the opposite pole, as shown in the coloured segments. Intermediate fibres show the same reciprocal behaviour. The suture pattern becomes much more complex as successive strata are added to the exterior of the growing lens, and the original arms of each triradiate suture show secondary and tertiary dichotomous branchings.

(By permission from Hogan MJ, Alvarado JA, Weddell JE 1971 Histology of the Human Eye. Philadelphia: WB Saunders.)

Lens fibres contain crystallins, proteins which are responsible for the transparency and refractile properties, and for much of the elasticity, of the lens. At least three varieties coexist, α, β and γ, their relative proportions changing throughout life. They occur in very high concentrations, and form up to 60% of the lens fibre mass. Variations in their concentration in different parts of the lens give rise to regional differences in refractive index, correcting for the spherical and chromatic aberrations which might otherwise occur in a homogeneous lens.

VITREOUS HUMOUR

The vitreous body occupies about four-fifths of the eyeball. Posteriorly it is in contact with the retina, while further forward it abuts the ciliary body, zonule and lens. Its anterior surface is hollowed into a deep concavity, the hyaloid fossa, fitting the shape of lens (Fig. 40.1). It is colourless, consisting of approximately 99% water, but is not entirely structureless. At its perimeter it has a gel-like consistency (100–300 μm thick); nearer the centre it contains a more liquid zone. Hyaluronan, in the form of long glycosaminoglycan chains, fills the whole vitreous. In addition, the peripheral gel or cortex contains a random loose network of type II collagen fibrils which are occasionally grouped into fibres. The cortex also contains scattered cells, the hyalocytes, which possess the characteristics of mononuclear phagocytes and may contribute to the production of hyaluronan. While they are normally in a resting state, they have the capacity to be actively phagocytic in inflammatory conditions. Hyalocytes are not present in the cortex bordering the lens.

The liquid vitreous is absent at birth, appears first at 4 or 5 years, and increases to occupy half the vitreous space by the seventh decade. The cortex is most dense at the pars plana of the ciliary body adjacent to the ora serrata, where attachment is strongest, and this is often referred to as the base of the vitreous. Here the vitreous is thickened into a mass of radial (zonular) fibres which form the suspensory ligament of the lens. Apart from the vitreous base, the vitreous also has a firm (peripapillary) attachment at the edge of the optic disc. This adherence of the vitreous to the retina can result in traction on the retina if the vitreous shrinks, such as occurs in old age, resulting in macular holes or peripheral breaks, possibly leading to retinal detachment.

A narrow hyaloid canal runs from the optic nerve head to the central posterior surface of the lens (Fig. 40.1). In the fetus this contains the hyaloid artery which normally disappears about 6 weeks before birth. The canal persists in adult life as a very delicate fibrous structure and is of no functional importance.

MICROSTRUCTURE

The retina is composed of a variety of epithelial, neural and glial cell types whose distribution conventionally divides it into 10 layers (Fig. 40.22). These are usually apparent in conventional histological sections (Fig. 40.23), but can also be seen in vivo using techniques such as optical coherence tomography, which uses backscattered light to visualize layers by differences in their optical scattering properties (Fig. 40.24). Embryologically, the retina is derived from the two layers of the invaginated optic vesicle. The outer layer becomes a layer of cuboidal pigment cells which separates the choroidal lamina vitrea from the neural retina, and therefore forms the outermost layer of the retina, the retinal pigment epithelium (RPE – layer 1). The other nine strata of the retina develop from the inner layer of the optic vesicle and form the neural retina.

Fig. 40.22 Neural cells whose cell bodies and interconnections account for the layered appearance of the retina in histological section (compare with Fig. 40.23). Also shown are the two principal types of neuroglial cell in the retina (although microglia are also present they are not shown).

Fig. 40.24 High definition optical coherence tomography (OCT) in vivo image of the human retina. The image has approximately 2 μm axial resolution, is 8 mm long and consists of 10,000 axial scans. NFL, Nerve fibre layer. GCL, Ganglion cell layer. IPL, Inner plexiform layer. INL, Inner nuclear layer. OPL, Outer plexiform layer. ONL, Outer nuclear layer. ELM, External limiting membrane. IS/OS, Boundary between the photoreceptor inner and outer segments. RPE/CH, Retinal pigment epithelium and choriocapillaris. The image is expanded in the vertical direction to permit better visualization of retinal layers.

(By courtesy of Professor James Fujimoto, Department of Electrical Engineering and Computer Science, MIT, Boston, USA.)

The outermost layer of the neural retina contains the light sensitive parts of the photoreceptors, which convert the optical image into neural activity. From the photoreceptors, neural activity flows radially to bipolar and ganglion cells, and laterally via horizontal cells in the outer retina and amacrine cells in the inner retina. Photoreceptors synaptically contact each other and bipolar and horizontal cells in the outer plexiform layer (OPL – layer 5), while bipolar, amacrine and ganglion cells synapse in the inner plexiform layer (IPL – layer 7). The axons of ganglion cells run towards the optic disc in the nerve fibre layer (NFL – layer 9), where they leave the retina as the optic nerve, which transmits the retinal output to the visual areas of the brain where visual processing is completed. Although most neural activity flows from the photoreceptors towards the brain, some information flow occurs in the opposite direction via centripetal fibres in the optic nerve and interplexiform cells in the retina which connect the inner and outer plexiform layers.

The classic ten layered appearance of the retina is absent in the optic nerve head, the fovea and foveola, and the ora serrata. At the optic nerve head, the axons of the retinal ganglion cells leave the retina to form the optic nerve and all the other neural cell types of the retina are missing. At the fovea and foveola, the inner five layers of the retina are ‘pushed aside’. At the ora serrata, where the retina borders the ciliary body (Fig. 40.9), the retinal pigment epithelium merges with the outer pigmented epithelium of the ciliary body, while the neural retina borders the inner unpigmented ciliary epithelium: the retina is thinnest at this point. The normal layered arrangement of the neural retina approaching the ora serrata is frequently disrupted by cysts in older individuals (Fig. 40.25).

Fig. 40.25 Junction between the retina and ciliary body (ora serrata). The retinal pigment epithelium is continuous with the outer pigmented epithelium of the ciliary body, while the neural retina abuts the inner unpigmented epithelium of the ciliary body. The layered appearance that is apparent elsewhere in the neural retina is disrupted adjacent to the ora serrata by cystic degeneration.

Cells of the retina

Retinal pigment epithelium

The retinal pigment epithelium, RPE, is composed of approximately cuboidal cells that form a single continuous layer extending from the periphery of the optic disc to the ora serrata, where it continues as the outer ciliary epithelium. The cells are flat in radial section and hexagonal or pentagonal in surface view, and number 4–6 million in the human retina. Their cytoplasm contains numerous melanosomes. Apically (towards the rods and cones), the cells bear long (5–7 μm) microvilli which contact, or project between, the outer segments of rods and cones. The tips of rod outer segments are deeply inserted into invaginations in the apical membrane of the RPE. The different embryological origins of the RPE and neural retina mean that the attachments between these two layers are unsupported by junctional complexes; the neural retina and RPE are therefore easily parted in the clinical condition of retinal detachment arising from trauma or disease.

RPE cells play a major role in the turnover of rod and cone photoreceptive components. Their cytoplasm contains the phagocytosed tips of rods and cones undergoing lysosomal destruction. The final products of this process are lipofuscin granules, which accumulate in these cells with age.

Light reaching the outer retina but missing the photoreceptors is absorbed by the RPE, which, like melanin elsewhere in the eye, prevents such stray light degrading image quality. The zone of tight junctions between adjacent cells also allows the epithelium to function as an important blood-retinal barrier between the retina and the vascular system of the choroid. The RPE is required for the regeneration of bleached visual pigment and may have antioxidant properties. It also secretes a variety of growth factors necessary for the integrity of the choriocapillaris endothelium and the photoreceptors, and produces a number of immunosuppressive factors. A failure of any of the diverse functions of the RPE could result in compromised retinal function and eventual blindness (Strauss 2005).

Rods and cones

Rods and cones are the ‘image forming’ photoreceptors of the outer retina and function at low (scotopic) and higher (photopic) light levels respectively. Both are long, radially orientated structures with a similar organization, although details differ (Fig. 40.26). From the choroidal end inwards, the cells consist of outer and inner segments connected by a thin connecting cilium (together making up layer 2 of the retina), a cell body containing the nucleus, and a synaptic terminal (either a more complex pedicle for cones or a simpler rod spherule) where they make synaptic connections with adjacent bipolar and horizontal cells and with other cone or rod cells within the OPL.

Fig. 40.26 The major features of a retinal rod cell (A) and a retinal cone cell (B). The relative size of the pigment epithelial cells has been exaggerated for illustrative purposes.

The nuclei of the rods and cones form the outer nuclear layer (ONL – layer 4). The cone nuclei are relatively large, oval and generally form a single layer that often penetrates the external limiting membrane (ELM – layer 3). They also contain less heterochromatin and thus usually stain more lightly. Rod nuclei are round and smaller, stain more darkly and form several layers vitreal to the cone nuclei.

The ELM is in fact not a membrane at all, although it appears as such in the light microscope, but a series of zonula adherentes between photoreceptors and the glial (Müller) cells that separate them. These junctions most likely serve to anchor the photoreceptors and prevent leakage of the interphotoreceptor matrix that surrounds the photoreceptor outer and inner segments.

Rod outer segments are cylindrical and consist of around 1000 flattened lobulated membranous discs. These form as deep infoldings of the plasma membrane at the base of the outer segment; they ‘bud off’ after formation so that the discs are not attached to the plasma membrane and are free floating within the outer segment. Cone outer segments are generally shorter and, as their name implies, often conical (especially in the peripheral retina). Cone discs do not bud off after formation and remain as infoldings of the plasma membrane.

The inner segment of both rods and cones is divided into an outer, mitochondria-rich, ellipsoid and an inner myoid which contains endoplasmic reticulum (Fig. 40.26). In most of the retina these inner segments are much wider in cones (5–6 μm at their widest point) than rods (1.5 μm) (Fig. 40.27). In both rods and cones proteins are manufactured within the myoid and incorporated into the newly formed discs at the base of the outer segment. In rods, as new discs are added, old discs are pushed up the outer segment and eventually phagocytosed by the RPE. Cone discs are also phagocytosed but the incorporation of new proteins within the discs is more diffuse. While all rods within the retina have a similar structure, the cones at the foveola are highly modified compared to those situated more peripherally, and in many ways resemble rods with a longer outer segment and a thinner inner segment.

Fig. 40.27 Tangential section of a human retina in the parafovea cut at the level of the inner/outer segments, showing the ‘mosaic’ of both the larger cones and the more numerous, but smaller, rods. Photoreceptors at the top of the image are sectioned at a level closer to the retinal pigment epithelium than receptors lower in the figure. The reduction in size of the cones at the top of the figure is explained by the conical shape of their outer segment. If the figure were continued upwards, representing sections closer to the RPE, the size of the cones would continue to decrease and the amount of surrounding white space increase. Rod outer segment diameter, however, would change little.

Light is absorbed by rhodopsins, visual pigments consisting of a protein, opsin, that spans the membrane of the outer segment discs, bound to a light absorbing chromophore, retinal, which is an aldehyde of vitamin A₁. Such rhodopsins have a smooth bell-shaped absorption profile with a point of maximum absorbance (λ_max), indicating the wavelength at which they are most sensitive. Humans possess four different opsins resulting in four spectrally distinct visual pigments, one located within the rods (λ_max 498 nm) and three within different populations of cones absorbing maximally at the short (λ_max 420 nm); middle (λ_max 534 nm) and longer wave (λ_max 563 nm) end of the visible spectrum. The three cone classes are sometimes referred to as the blue, green and red cones, but are better classed as S, M and L cones. The action of light is to isomerize the retinal separating it from the opsin, a process which, via a G-protein coupled enzyme cascade and a second messenger system, results in the closure of cation channels in the receptor outer segment membrane, a hyperpolarization of the photoreceptor, and a consequent decrease in the release of the neurotransmitter glutamate from its synapses.

The human retina contains on average 4.6 million cones and 92 million rods, although there is significant inter-individual variation (Curcio et al 1990). Although cones populate the whole retina, their density is highest in the foveola, where approximately 7000 cones reach an average density of 199,000 cones/mm²: this area is entirely rod free. Going outwards from the foveola, rod numbers rise, reaching a peak density in a horizontal elliptical ring at the eccentricity of the optic disc, before declining once more toward the periphery. Cone density is 40–45% higher in the nasal compared to the temporal retina, and slightly higher inferiorly than superiorly.

The number of S cones in all human retinae is similar, making up less than 10% of all cones (Curcio et al 1991; Hofer et al 2005). The distribution of S cones is relatively even throughout the retina, although they are absent from the central fovea. The relative proportions of L and M cones shows a much greater degree of variation between individuals, the L : M cone ratio varying from close to unity to over 10. The distribution of L and M cones is more irregular than that of S cones, and appears random with some indication of clumping (Bowmaker et al 2003; Hofer et al 2005).

The high packing density of cones at the foveola, achieved by decreasing inner segment size, ensures maximal resolution, while the presence of more than one spectral cone type allows colour vision. S cones probably contribute little to spatial resolution because they are absent from the foveola. Rod-based vision provides high sensitivity, but with relatively low spatial discrimination and no ability to distinguish wavelengths. Although many of the functional differences between rods and cones rely on the different properties of the photoreceptors themselves, their connectivity to other retinal neurones is equally important.

Horizontal cells

Horizontal cells are inhibitory interneurones. Their dendrites and axons extend laterally within the OPL, making synaptic contacts with cone pedicles and rod spherules, and, via gap junctions at the tips of their dendrites, with each other. Their cell bodies lie in the outer part of the inner nuclear layer (INL – layer 6). Three morphological types of horizontal cell can be distinguished in the human retina (Kolb et al 1992). The dendrites of HI and HIII cells contact cones, and their axons terminate on rods. Both the axons and dendrites of HII cells synapse only with cones.

Bipolar cells

Bipolar cells are radially orientated neurones. Their dendrites synapse on photoreceptors, horizontal cells and interplexiform cells in the OPL. Their somata are located in the INL, and axonal branches in the IPL synapse with dendrites of ganglion cells or amacrine cells. Golgi staining has identified nine distinct types of bipolar cell in the human retina (Kolb et al 1992), eight of which contact cones exclusively, and the remaining type synapses only on rods.

Cone bipolars are of three major types, midget, S (blue) cone and diffuse, according to their connectivity and size. Midget cone bipolar cells either invaginate the cone pedicle or synapse on its base (flat subtype). In the central retina, each midget bipolar cell contacts only a single cone (2–3 in the periphery) forming part of a one-to-one channel from cone to ganglion cell that mediates high spatial resolution. S cones form part of a short wavelength mediating channel, while the larger diffuse cone bipolars are connected to up to ten cones and are thought to signal luminosity rather than colour.

Cone bipolar cells can be of two types, according to their response to the light-induced decrease in glutamate release from the photoreceptors to which they are synaptically connected. If illumination of the photoreceptors with a point of light causes a depolarization of the connected bipolar cell, it is said to be an ‘ON’ bipolar, connected to the cone by ‘sign-inverting’ synapses with metabotropic receptors. However, if cones are connected to an ‘OFF’ bipolar cell via ‘sign-conserving’ synapses with ionotropic receptors, illumination of the photoreceptor will result in hyperpolarization of the bipolar cell. Illumination of a concentric area of surrounding photoreceptors causes the opposite response in the bipolar cell. This inhibition is mediated via horizontal cells and gives rise to the antagonistic centre-surround type receptive field that is characteristic of all levels of the visual system up to and including the occipital cortex.

The single morphological type of rod bipolar cell contacts 30–35 rods in the central retina, increasing to 40–45 rods in the periphery. Such convergence serves to increase the absolute sensitivity of the rod system. All rod bipolar cells are ‘ON-centre’ and do not contact ganglion cells directly, but synapse with a class of amacrine cell (AII) which then contacts cone bipolar cells.

The IPL can be divided into two main layers, an outer layer containing the synaptic endings of ‘OFF’ cone bipolar cells, and an inner layer of ‘ON’ cone and rod bipolar cell synapses.

Amacrine cells

Most amacrine cells lack typical axons and consequently their dendrites make both incoming and outgoing synapses. Each neurone has a cell body either in the INL near its boundary with the IPL, or on the outer aspect of the ganglion cell layer, when it is known as a displaced amacrine cell. The processes of amacrine cells make a variety of synaptic contacts in the IPL with bipolar and ganglion cells as well as with other amacrine cells.

The various classes of amacrine cell serve a number of important functions in vision. AII cells play an essential role in the rod pathway (see above). Other cells appear to be important modulators of photoreceptive signals, and serve to adjust or maintain relative colour and luminosity inputs under changing light conditions. They are probably also responsible for some of the complex forms of image analysis known to occur within the retina, such as directional movement detection. Up to 24 different morphological types are recognized in humans (Kolb et al 1992): coupled to their neurochemical complexity, this makes them perhaps the most diverse neural cell type in the body.

Interplexiform cells

Interplexiform cells, often regarded as a subclass of amacrine cells, generally have cell bodies in the INL. They are postsynaptic to cells in the inner retina, and send signals against the general direction of information flow in the retina, synapsing with bipolar, horizontal and photoreceptor cells in the OPL. Although their function is uncertain, it is likely that they adjust some aspect of retinal function such as sensitivity.

Ganglion cells

The human retina contains 0.7–1.5 million ganglion cells, the output neurones of the retina (Curcio & Allen 1990). Their dendrites synapse with processes of bipolar and amacrine cells in the IPL. Ganglion cell bodies, together with displaced amacrine cells, form the ganglion cell layer of the retina (GCL – layer 8). Throughout most of the retina they form a single layer; they become progressively more numerous near the macula, where they are ranked in up to 10 rows, reaching a peak density of up to 38,000/mm² in a horizontally orientated elliptical ring 0.4–2.0 mm from the foveal centre. Their number diminishes again towards the fovea, from which they are almost totally excluded.

Up to 15 ganglion cell types have been identified in the mammalian retina based on morphology, physiology, and target area in the brain, each of them presumably functionally distinct. For example, some project to different regions of the LGN and form three parallel visual pathways involved in conscious visual perception, namely the magnocellular and parvocellular systems and a pathway carrying the S cone signal (Wässle 2004). Midget ganglion cells (P cells) contact only single midget bipolar cells in the central retina, which in turn connect to single cones, giving each cone a ‘private line’ out of the retina and ensuring optimal acuity. The large dendritic field of parasol (M cells) is consistent with a role in motion detection. Parasol and midget ganglion cells together make up around 80% of human retinal ganglion cells. The remaining cells (approximately 200,000) project to the superior colliculus of the midbrain, the thalamic pulvinar, the pretectum and the accessory optic system, and contribute to various subconscious visual reflexes such as the pupillary and accommodation responses. In addition, a population of around 3000 large, intrinsically light-sensitive ganglion cells form a network composed of extensive overlapping dendrites (Dacey et al 2005). Such ‘inner retinal photoreceptors’ contain a retinal-based visual pigment (melanopsin; λ_max approximately 479 nm) which resembles an invertebrate-type visual pigment in many of its characteristics. These light-sensitive ganglion cells are part of a pathway parallel to the rod and cone-mediated ‘image-forming’ system that monitors overall levels of illumination. This ‘non-imaging’ pathway is the major route by which the eye influences circadian rhythms via the suprachiasmatic nucleus; it also contributes to light-evoked pupillary constriction via projections to the olivary pretectal nucleus.

Ganglion cell axons, which form the NFL on the inner surface of the retina, run parallel to the surface of the retina, and converge on the optic nerve head where they leave the eye as the optic nerve. Fibres from the medial (nasal) retina approach the disc in a simple radial pattern (Fig. 40.28), whereas axons from the lateral (temporal) retina take an arcuate route as they avoid the fovea. Axons from the macula form a papillomacular fasciculus which passes almost straight to the disc. The thickness of the NFL increases dramatically near the optic disc as fibres from the peripheral retina traverse more central areas. Towards the edge of the disc the other retinal layers thin, and at the disc all neural elements of the retina other than ganglion cell axons are excluded.

Fig. 40.28 Directions of axons (dashed lines) and blood vessels of the nerve fibre layer of the retina. Axons pass radially on the nasal side of the optic disc, whereas fibres on the temporal side avoid crossing the fovea by arching around it. Some of the fibres from the fovea and central region pass straight to the optic disc and others arch above and below the horizontal; together these form the papillomacular bundle. A raphé is formed by the central fibres temporal to the fovea. Venules are shown crossing in front of arteries: the reverse relationship is probably the more common pattern. All vessels issue from the disc on the right of the figure; the larger temporal branches tend to arch around the central region of the retina and do not approach the fovea. The peripheral retina and most of the nasal retina are not shown.

Axons of ganglion cells are surrounded by the processes of radial glial cells and retinal astrocytes and are almost always non-myelinated within the retina, which is an optical advantage because myelin is refractile. Although a few small myelinated axons may occur, myelination does not generally start until axons enter the optic disc to become the optic nerve.

Retinal glial cells

There are three types of retinal glial cells, namely, radial Müller cells, astrocytes and microglia. Müller cells form the predominant glial element of the retina; retinal astrocytes are largely confined to the ganglion cell and nerve fibre layers; microglial cells are scattered throughout the neural part of the retina in small numbers.

Müller cells span almost the entire thickness of the neural retina, ensheathing and separating the various neural cells except at synaptic sites. They constitute much of the total retinal volume, and almost totally fill the extracellular space between neural elements. Their nuclei lie within the INL, and from this region each cell body extends a single thick fibre that runs radially outwards, giving off complex lateral lamellae which branch among the processes of the OPL. Apically, each central process terminates at the ELM from which microvilli project for a short distance into the space between the rod and cone inner segments (fibre baskets) (Fig. 40.22). On the inner surface of the retina, the main Müller cell process expands into a terminal foot plate which contacts those of neighbouring glial cells and forms part of the internal limiting membrane (ILM; see below).

Like astrocytes, Müller cells also contact blood vessels, especially capillaries of the INL, and their basal laminae fuse with those of perivascular cells or vascular endothelia, contributing to the formation of the blood–retinal barrier. They also maintain the stability of the retinal extracellular environment by, for example; regulation of K⁺ levels, uptake of neurotransmitter, removal of debris, storage of glycogen, electrical insulation of receptors and other neurones, and mechanical support of the neural retina.

The cell bodies of retinal astrocytes lie within the NFL and their processes branch to form sheaths around ganglion cell axons. The close association between astrocytes and blood vessels in the inner retina suggests that they contribute to the blood–retinal barrier. Retinal microglia are scattered mainly within the IPL. Their radiating branched processes spread mainly parallel to the retinal plane, giving them a star-like appearance when viewed microscopically from the surface of the retina. They can act as phagocytes, and their number increases in the injured retina.

The inner border of the retina is formed by the internal limiting membrane (ILM – layer 10) which consists of collagen fibres and proteoglycans from the vitreous, a basement membrane (which is continuous with the basal lamina of the ciliary epithelium), and the plasma membrane of expanded Müller cell terminal foot plates. It is 0.5–2 μm thick in the posterior retina and thickens with age. The ILM is involved in fluid exchange between the vitreous and the retina and, perhaps through the latter, with the choroid. It also has various other functions including anchorage of retinal glial cells, and inhibition of cell migration into the vitreous body.

Modifications of the central retina

The central retina, clinically referred to as the macula, is composed of four concentric areas, which, starting with the innermost, are: the foveola (0.35 mm diameter, equivalent of an angular subtense at the nodal point of around 1.25°), the fovea (1.5 mm, 5.2°), the parafovea (2.5 mm, 8.6°), and the vaguely defined perifovea (approximately 5–6 mm, 20°). The foveola, which contains no rods or S cones, is centred about 3 mm temporal and 1 mm inferior to the optic disc (Fig. 40.20). In the foveola and surrounding fovea all the inner layers of the neural retina beyond the ONL have been displaced peripherally resulting in a retinal thickness around half of that elsewhere in the retina (Figs 40.24, 40.29). This foveal pit is created by the cone ‘axons’, known here as Henle fibres, running almost parallel to the retinal surface before connecting to post-receptoral retinal neurones outside the fovea. The Henle fibres contain two xanthophyll carotenoid pigments (lutein and zeaxanthin) which create an elliptical yellowish area (approximately 2 mm horizontally and 1 mm vertically), the macula lutea. Macular pigment density varies by more than an order of magnitude between individuals, is influenced by several environmental factors including diet, and is negligible in the central foveola.

Fig. 40.29 Section through the fovea centralis

(By permission from Young B, Heath JW 2000 Wheater’s Functional Histology. Edinburgh; Churchill Livingstone.)

The absence of the inner retinal layers, including blood vessels (Figs 40.20, 40.21), reduces light scatter, which, along with the increased packing density of cones in the foveola and their lack of convergence with ganglion cells, ensures visual resolution is highest in this part of the retina. Acuity may be further enhanced by the macular pigment, which apart from having antioxidant properties and removing potentially harmful short-wave radiation, will absorb those wavelengths most prone to chromatic aberration and Rayleigh scatter.

VASCULAR SUPPLY

The retina has a dual arterial supply and both parts are necessary to maintain retinal function. The outer five layers of the retina are avascular and rely on an indirect supply from the choriocapillaris in the choroid. The inner retina receives a direct blood supply through capillaries connected to branches of the central retinal artery and vein. Only the inner retinal circulation is described here. The choroidal blood supply is discussed on p. 680.

The central retinal artery enters the optic nerve as a branch of the ophthalmic artery approximately 1.2 cm behind the eyeball, and travels within the optic nerve to its head, where it passes through the lamina cribrosa. At this level the central artery divides into equal superior and inferior branches, which, after a few millimetres, divide into superior and inferior nasal, and superior and inferior temporal, branches, each supplying a ‘quadrant’ of the retina (Fig. 40.28). Although similar retinal veins unite to form the central retinal vein, the courses of the arteries and veins do not correspond exactly. These vessels mainly run within the nerve fibre and ganglion cell layers of the retina, accounting for their clarity when seen through an ophthalmoscope (Fig. 40.20). Arteries often cross veins, usually lying superficial to them: in severe hypertension the arteries may press on the veins and cause visible dilations distal to these crossings. The vitreal location of arteries, and their lighter, bright red colouration and smaller diameter in comparison to veins, allows the two vessel types to be distinguished ophthalmoscopically.

From the four major arteries subsequent dichotomous branches run from the posterior pole to the periphery, supplying the whole retina (Zhang 1994). Arteries and veins ramify in the nerve fibre layer, near the ILM, and arterioles pass deeper into the retina to supply capillary beds. Venules return from these beds to larger superficial veins which converge towards the disc to form the central retinal vein.

Retinal capillary networks can occur in three different layers, the number of layers depending on location (Zhang 1994). Radial peripapillary capillaries are the most superficial of the capillary networks and lie within the inner nerve fibre layer. A layer of inner capillaries runs within the nerve fibre and ganglion cell layers, while an outer capillary layer is located in the inner plexiform and inner nuclear layers (Fig. 40.30). Approaching the fovea, capillaries are restricted to two layers, and terminal capillaries eventually join to form a single-layered macular capillary ring, producing a capillary-free zone 450–500 μm in diameter at the fovea. This avascular region is clearly visible in a fluorescein angiogram (Fig. 40.21). Capillaries become less numerous in the peripheral retina and are absent from a zone approximately 1.5 mm wide adjoining the ora serrata.

Fig. 40.30 Tangential section of the retina at the level of the inner nuclear layer, highlighting the dense network of capillaries.

The territories of the arteries which supply a particular quadrant do not overlap, nor do the branches within a quadrant anastomose with each other, and consequently a blockage in a retinal artery causes loss of vision in the corresponding part of the visual field. The only exception to this end-arterial pattern is in the vicinity of the optic disc. Here, the posterior ciliary arteries enter the eye near the disc (Figs 40.10, 40.31), and their rami not only supply the adjacent choroid, but also form an anastomotic circle in the sclera around the head of the optic nerve (Fig. 40.31). Branches from this ring join the pial arteries of the nerve, and small cilioretinal arteries from any arteries in this region may enter the eye and contribute to the retinal vasculature, possibly resulting in the preservation of visual function following central retinal artery occlusion. Similarly, small retinociliary veins may sometimes also be present.

Fig. 40.31 Vessels of the optic nerve head. The dotted areas represent the principal glial membranes.

The structure of retinal blood vessels resembles that of vessels elsewhere, except that the internal elastic lamina is absent from the arteries, and muscle cells may appear in their adventitia. Capillaries are non-fenestrated and endothelial cells are joined by complex tight junctions, fulfilling the requirements of a functional blood–retinal barrier. Within the optic nerve, the central artery is innervated by sympathetic and parasympathetic fibres: this innervation does not extend to the vessels in the retina. The parasympathetic supply is derived from the pterygopalatine ganglion, and the fibres are predominantly VIP-ergic (and probably NO-ergic) and vasodilatory. The sympathetic supply is derived from the plexus around the internal carotid and ophthalmic arteries.

RETINOPETAL INNERVATION OF THE RETINA

The human retina receives modulating input from the brain via fibres within the optic nerve which emerge at the optic disk as large axons, and then course through the NFL, gradually diminishing in size. Although there are only a small number of these retinopetal axons (usually less than 10), they branch extensively before terminating in the IPL (Reperant et al 2006). One set of such axons arises from perikarya in the posterior hypothalamus and uses histamine as a neurotransmitter, while other, serotoninergic, fibres arise from cell bodies in the dorsal raphé. These neurones project to many targets in the central nervous system including the retina, forming components of an ascending arousal system. Such fibres allow higher central areas to modulate aspects of retinal activity; in man they may, among other things, affect sensitivity and modulate retinal blood flow (Gastinger et al 2006).

OPTIC NERVE HEAD

The axons of more than a million retinal ganglion cells converge at the optic nerve head (ONH) and leave the eye by penetrating the sclera to form the optic nerve (ON). The ONH represents that part of the optic nerve lying within the bulb of the eye. Since all retinal neural elements, apart from ganglion cell axons, are absent from this region it is insensitive to light and forms the ‘blind spot’.

Histologically the ONH can be divided into three zones (Fig. 40.32). These are prelaminar (the anterior part terminating at the vitreous); laminar (formed by the lamina cribrosa); postlaminar (continuous with the retrobulbar optic nerve). The surface view of the ONH, usually seen with an ophthalmoscope, is referred to as the optic disc.

Fig. 40.32 The optic nerve head, showing the distribution of collagenous tissue (grey) and neuroglial nuclei (solid blue circles). 1a, retinal internal limiting membrane; 1b, inner limiting membrane of Elschnig; 2, central meniscus of Kuhnt; 3, spur of collagenous tissue separating the anterior lamina cribrosa (6) from the choroid; 4, border tissue of Jacoby; 5, intermediary tissue of Kuhnt; 7, posterior lamina cribrosa; Sep, connective tissue septa from pia mater; Gl. M, astroglial membrane; Gl. C, astrocytes and oligodendrocytes among the fibres in their fascicles; Du, Ar, Pia: dura, arachnoid and pia mater, respectively. The dotted lines represent the borders of the lamina cribrosa.

(By permission from Anderson DR, Hoyt W 1969 Ultrastructure of intraorbital portion of human and monkey optic nerve. Arch Ophthalmol 82: 506–30.

VISUAL FIELD DEFECTS

The basis for clinical assessment of damage to the visual pathway is an understanding of the retinotopic projections within the pathway. Moreover, plotting visual field loss frequently reveals the approximate location of the causative lesion and sometimes its nature (Fig. 40.33). Since retinal lesions can be visualized with an ophthalmoscope, field testing might appear to be redundant for such defects, but visual field measurement is still helpful in assessing the extent of the damage and may be the key factor in confirming a diagnosis. Glaucoma serves as an example. Field defects in glaucoma, occurring as a consequence of damage to the nerve fibre bundles at the optic nerve head, may be detectable ophthalmoscopically, but confirmation of the diagnosis frequently depends on field assessment. Early defects consist of one or more areas of paracentral focal field loss, progressing to arcuate scotomas. The shape of the defect corresponds to the anatomical arrangement of ganglion cell axons (Fig. 40.28).

So far as the location of lesions central to the retina is concerned, deficits in the vision of one eye are usually attributable to optic nerve lesions. Lesions of the optic chiasma, involving crossing nerve fibres, produce a bilateral field loss as exemplified by a pituitary adenoma (Fig. 40.33). The tumour expands upwards from the pituitary fossa, compressing the inferior midline of the chiasma, and eventually produces bitemporal hemianopia, starting with an early loss in the upper temporal quadrants. Field defects in the rare instances of optic tract lesions are distinctive. The tract contains contralateral nasal and ipsilateral temporal retinal projections and damage will cause a homonymous contralateral loss of field with substantial incongruity (dissimilar defects in the two fields). Incongruity probably results from a delay in achieving coincidence between retinal topographical projections of the two inputs of the visual pathway, as contiguous projections adjust their location, gradually achieving coincidence. It also probably reflects the reorganization of fibres which occurs normally in the optic tracts, as some fibres leave the tract in the superior brachium and others progress to the lateral geniculate nucleus. Incongruity is most marked in defects of the optic tract, less obvious in optic radiation defects, and is usually absent in cortically induced field defects, thus providing an additional clue in assessing location of the cause.

Lesions of the optic radiations are usually unilateral, and commonly vascular in origin. Field defects therefore develop abruptly, in contrast to the slow progression of defects associated with tumours, and the resulting hemifield loss follows the general rule that visual field defects central to the chiasma are on the opposite side to the lesion. Little or no incongruity is seen in visual cortical lesions, but they commonly display the phenomenon of macular sparing, the central 5–10° field being retained in an otherwise hemianopic defect.