DOYNE LECTURE THE PATHOLOGY OF THE OUTFLOW SYSTEM IN PRIMARY AND SECONDARY GLAUCOMA

W. R. LEE Glasgow

On behalf of my specialty of ophthalmic pathology, I noted that since the contribution of Doyne himself, am indebted to the Master of the Oxford Congress few pathologists have been honoured. Professor for the invitation to present the Doyne Memorial Norman Ashton and the late Professor David Lecture on the pathology of the outflow system in Cogan were the only ophthalmic pathologists primary and secondary glaucoma. Although I regard included in the list. In 1960, Professor Ashton3 the Doyne Lecture as an awesome responsibility, it chose to lecture on the pathology of the outflow has been an enjoyable exercise in preparing for it to channels in glaucoma and I was interested to see that go through the contributions of previous Doyne in the final paragraph of his text he wrote 'no doubt lecturers and to read the remarks made by those who this topic will again be chosen by a Doyne Lecturer'. knew Doyne personally. From the various very Thus it is appropriate that the problem of glaucoma distinguished Doyne lecturers who have preceded should again be addressed by a pathologist some 35 me, I sense that the ethos of this lecture is that years later. In this lecture I have drawn freely on the speculation is an integral requirement of the speaker. research material which my colleagues and I have This is not so easy for a pathologist, but I hope that been studying in Glasgow for the past 25 years and I some of the text will be sufficiently controversial to have used this to re-interpret conventional morphol­ stimulate further research work. ogy in the light of modern technology which uses It has also been an enjoyable task to review some molecular biology. of Doyne's published work and to read descriptions In the 1970s and early 1980s, I had the privilege of of the man and his attitudes to work and to collaborating with Ian Grierson and Paul McMena­ relaxation. The account of the life of Doyne prepared by Catford1 was of great value to those min in the investigation of the response of the sub­ who, like myself, were aware of the name only in the human primate outflow system to a variety of short­ context of this lecture. The illustrations in Catford's term stresses such as alterations in intraocular paper highlight the many facets of the character of pressure and exposure to enzymes. Non-survival Doyne - scholar, musician, clinician , fencer and experiments such as these provided valuable infor­ sailor - elegantly combining the attitudes of the mation on the functional morphology of the normal Anglo-Saxon and the Celt to recreation and tissue. contemplation. In his generation so many great The study of diseased tissue was not so easily men appeared to have a multitude of talents, which undertaken due to a paucity of suitable. specimens is something of a rarity in contemporary society. appropriately fixed and prepared for sophisticated However, the statement which I personally found pathological investigations. Accordingly for the last most endearing and of greatest encouragement when 25 years I have adopted the policy of preparing any I started to prepare for this lecture, was written in potentially interesting specimen for electron micro­ 2 Doyne's textbook in 1896, viz. 'Glaucoma was the scopy in the hope that eventually it would be possible most dreaded of all eye diseases'. to generate a cohesive description of those disease On looking through the list of Doyne medallists, I processes which test the outflow system to destruc­ tion. The intention was to integrate the results of Correspondence to: Professor W. R. Lee. Department of Ophthalmology. The University of Glasgow. Glasgow Gil 6NT. pathomorphological investigation with the massive UK. amount of factual information currently available in

Eye (1995) 9, 1-23 © 1995 Royal College of Ophthalmologists 2 W. R. LEE

Fig. 1. A giant vacuole (v) in the lining endothelium of Fig. 2. There is marked distension of the cribriform layer Schlemm's canal in a 60-year-old male suffering from of the rhesus monkey subject to an intracameral perfusion at partial angle closure with overperfusion in this part of the 50 mmHg for 1 hour prior to fixation. The collector channel meshwork. The underlying cribriform layer contains a fine (cc) is blocked and the giant vacuoles (arrowheads) are fibrillar extracellular matrix (arrowheads) and electron­ adherent to the outer wall endothelium. (Transmission dense clumps (arrows). (Transmission electron micrograph, electron micrograph, X 900) X 5000) the many excellent textbooks dealing with the clinical occurs after a hole or pore forms in the cytoplasm 4 6 aspects of glaucoma in all its forms. - which bulges into Schlemm's canal. The structure In this lecture it is appropriate to first review the (Fig. 1) then acts as a transcellular channel which current status of knowledge concerning the func­ permits rapid fluidtransport when the pressure in the tional anatomy of the outflow system. Thereafter cribriform layer is high; conversely with collapse of various pathological processes will be described in the cribriform layer at low intraocular pressure, order to provide an understanding of the mechan­ blood refluxis prevented - hence the concept of the isms which can lead to breakdown of the tissue and 'one-way valve' (for review see Tripathi and an increased resistance to aqueous outflow and, TripathiI4). ultimately, to intractable glaucoma. In the following decades the importance of the It is currently accepted that the outflow systemhas giant vacuole system in bulk flow through an the following functional properties and can behave endothelial monolayer became accepted, not only . as: 7 for transport of aqueous from the eye, but also for 1. a one-way valve; passage of cerebrospinal fluid into the intracranial 7 2. a resistor which maintains intraocular pressure; venous sinuses.14 The pressure-sensitive nature of the 3. a selective filterS which (a) removes particulate giant vacuole bulk transfer system was demonstrated materials (endogenous and exogenous), (b) per­ in 1973 by Johnstone and Grantl5 working in Boston mits the escape of cellular infiltrates and (c) acts as on human autopsy eyes and on monkey eyes, and an immunological recognition system;9 their studies emphasised the importance of intra­ 4. a pressure sensor.lO ocular pressure on the numbers of giant vacuoles in the lining endothelium of Schlemm's canal. The hypothesis of transcellular bulk flow did not THE NORMAL OUTFLOW SYSTEM go unchallenged and the possibility that the vacuolar The Outflow System as a One-Way Valve structures were autolytic artefacts persisted? How­ Morphological investigation of the mechanisms for ever, subsequent conclusive studies proved the aqueous transport across the lining endothelium of opposite.16 In the early 1970s the working hypoth­ Schlemm's canal was initiated by Garron et at. I I in esis of Grierson and myself was that, if the giant 1958 and Holmbergl2 in 1959 with the discovery that vacuoles were not artefactual, there would be an the monolayer contained large membrane-bound increase in number and size in response to pressure spaces which were subsequently described as giant and the concomitant flow of aqueous through the vacuoles. The nature of giant vacuoles was not fully trabecular tissues. Painstaking morphometric studies appreciated until Tripathil3 provided the hypothesis carried out on the monkey eye by Grierson17-l9 that a giant vacuole is preceded by the formation of provided data to prove that the number and size of an invagination and that emptying of the vacuole giant vacuoles increased with a pressure rise (as THE OUTFLOW SYSTEM IN GLAUCOMA 3

ous physiological studies of aqueous outflow dynamics. The importance of glycoproteins in resistance to aqueous outflow was investigated by McMenamin, who conducted a series of experiments in which the baboon anterior chamber was perfused in vivo with a solution containing hyaluronidase at physiological pressures (18 mmHg).27,28 The prolonged perfusion experiments with and without the addition of hyaluronidase for periods of 4-6 hours on the primate eye, i.e. with a combination of physiological measurements and morphometry,27 failed to demon­ strate any effect of the enzyme. This was because the perfusion technique caused such distension in the control tissue that an enzyme effect, if it indeed exists, was totally obscured. Our observations were Fig. 3. A light micrograph to show numerous giant confirmed by another recent study29 which also vacuoles (arrowheads) in the human lining endothelium demonstrated that the distension of the cribriform of Schlemm's canal after fixation perfusion at 30 mmHg. (X 475) layer obstructed the collector channels. Our conclu­ sion was that the pliability of the monkey outflow maintained by a manometer in vivo) in the normal system rendered our physiological data unreliable. physiological range. Thus a 'giant vacuole count' Coincidental with the experimental work, it was became a mathematical indicator of both pressure our policy to study drug effects on the human outflow and flow through the outflow system, particularly if system using eyes obtained immediately after this was associated with distension of the cribriform surgical enucleation, in most cases in treatment for layer. posterior uveal melanoma. In the early experiments Similarly the number of flow channels and pores the globes were fixed immediately by immersion in (as determined by scanning and transmission elec­ glutaraldehyde. We found that pre-enucleation 0 tron microscopy) in the lining endothelium also treatment of the patients with topical pilocarpine3 increased in response to an increase in intraocular resulted in a rotation of the scleral spur and a pressure?O-23 The morphological observations with statistically significantincrease in the number of giant regard to the response of the outflow system to vacuoles in the lining endothelium of Schlemm's intraocular pressure were confirmedand accepted by canal. We concluded that rotation of the spur opened other workers in the field and recent review articles the inter-trabecular spaces and allowed aqueous to describe the lraiseworthy contributions of Alvarado pass more easily into the cribriform layer and et al.,24 Bill2 and Rohen et al.26 through the endothelium of Schlemm's canal. The combination of studies on the sub-human This hypothesis was strengthened by experiments primate eye and the human eye revealed that the on the baboon eye31 ,32 which revealed an absence of young monkey or baboon eye had a much more a detectable effect on the giant vacuole count after pliable outflow system than the human eye. In the instillation of topical pilocarpine. This was attributed subhuman primate, the pressurelfiow vacuole rela­ to the fact that this species does not possess a scleral tionship held at relatively low pressures for fairly spur so that constriction of the muscle does not short experimental time intervals (1-2 hours), but at widen the meshwork. In another series of experi­ pressure levels of the order of 50 mmHg, large bulges ments, pretreatment with topical timolol of the in the delicate cribriform layer blocked off the human eye prior to enucleation for melanoma had collector channels, \5.22 with presumably an increase no effect on the giant vacuole count by comparison in resistance to outflow (Fig. 2). In addition, large with untreated controls.33 It is generally accepted giant vacuoles in the lining endothelium of the inner that the pressure-lowering effect of timolol is due to trabecular wall became adherent to the outer wall inhibition of aqueous inflow and it was not endothelium and it was not difficult to find signs of unexpected that there were no detectable changes rupture and disorganisation in the monolayer. We in the outflow system. speculated that disruption of the extracellular matrix These experiments were followed by an investiga­ in the distended cribriform layer would, by allowing tion into the rapidity with which the human mesh­ free passage of aqueous, act to counterbalance the work can respond to changes in pressure.34 Eighteen increased resistance due to plugging of collector enucleated human eyes were perfused with glutar­ channels. This form of disorganisation would provide aldehyde solution at a pressure of 30 mmHg on the detectable, but erroneous, information in overzeal- assumption that the adaptation to this pressure level 4 W. R. LEE

would be captured by diffusion of glutaraldehyde, colloidal gold, ferritin, latex microspheres, fibrin, which rapidly causes the tissue to become firm. and leucocyte debris, intact red cells and red cell debris, causes increasing resistance to outflOW.16 An addi­ bacteria, even fibrils from the cat tapetum35) were tional factor to be considered in hindsight is that the used by various workers in in vivo experiments.R,36.37 stabilisation of the tissue was probably protective in It was also possible to demonstrate that the terms of excessive pressure stress. trabecular endothelial cells were capable of phago­ R It was reassuring to find that the elderly human cytosis in vitro.3 -4o In such experiments it soon outflow system was much more robust in the face of became obvious that the trabecular endothelial cells this insult than was that of the sub-human primate. In in culture had a selective capacity to phagocytose this study of the trabecular meshwork in eyes certain particles, particularly melanosomes. It was perfused immediately after enucleation,34 the inci­ tempting to assume that release of large numbers of dence of giant vacuoles was 3-5 times greater in the particles into the anterior chamber in vivo would perfused eyes than in the control, immersion-fixed overload the system and that this would be eyes (Fig. 3). In the perfused eyes (age 20-60 years) deleterious because cell separation would allow the mean giant vacuole count was between 15 and 30 water influx into the trabecular beams which would per section, while in the elderly (70-80 years) the lead to occlusion of the intertrabecular spaces. The count was between 10 and 20. In the immersion-fixed nature of the particles is of importance in the specimens the giant vacuole count declined from a stimulus to phagocytosis and its effects. For exam­ maximum of 10 in the younger age groups to less ple, the system can readily remove melanin granules, than 5 per section in the later decades. It should be as evidenced by the complicated survival experi­ stressed that this form of perfusion fixationappli ed a ments which have shown that the monkey outflow 30 mmHg pressure differential directly across the system can recover from repeated injections of meshwork and the canal endothelium, because the melanin into the trabecular meshwork.41 pressure in the episcleral veins was of course zero in the enucleated eye. One of the important findings was a decrease in the number of giant vacuoles with The Outflow System as a Resistor age in both the perfused and immersion-fixed eyes. The main impetus to the intensive study of the However, it was also possible to demonstrate an age­ trabecular meshwork was the desire to identify a related decrease in the size of canal due to morphological cause for the increased resistance hyalinisation of the scleral sulcus and enlargement which occurs in primary open angle glaucoma, a of the scleral spur. When the age-related reduction in topic which is discussed later. Since there is over­ the circumference of the canal was taken into whelming evidence that in normal tissue the main account, the capacity of the endothelium to form locus of resistance is the outermost layer of the giant vacuoles no longer exhibited an age-related trabecular meshwork �the cribriform layer or decline. In none of the human tissue used for this juxtacanalicular layer ),7,3 .42-45 morphologists in the study was there any evidence of tissue prolapse into last two decades have, using light microscopy, collector channels, nor was there narrowing of the immunocytochemistry and electron microscopy, canal with contact between the inner and outer walls energetically attempted to identify those compo­ as had been observed in the rhesus monkey and nents of the extracellular matrix which might be baboon eyes. responsible for resistance to fluid flow. At the The final conclusions from the numerous studies ultrastructural level using conventional processing described above were that the number and dimen­ techniques, it was relatively easy to classify different sions of the giant vacuoles in the lining endothelium constituents on purely morphological grounds,26.37.45 were an indicator of the status of the outflow system viz. 'amorphous', 'fine fibrillar', 'wide banded', in its function as a one-way valve. The work on 'basement-membrane-like', etc., even though the human tissue also introduced the caveat that there biochemical nature was unknown. Rohen and his was an apparent decline in function as the ageing co_workers26,37.46.47 concluded that electron-dense process continues and this had to be taken into plaques of collagenous and elastic-like material, so account in the study of more advanced disease in the prominent in the cribriform layer (Fig. 1), were the outflow system. most likely materials to cause resistance to aqueous outflow. One interesting feature of the outflow system is The Outflow System as a Filter that with ageing the trabecular beams become The principal awareness of the outflow system as a thickened by deposition of a collagenous extracel­ filter came as a result of experiments intended to lular matrix on the trabeculae, this being accompa­ determine the size of the pores in the giant vacuoles nied b deposits on the outer wall of Schlemm's present in the lining endothelium of Schlemm's canal.4 I- 47 These are all morphological features canal. A variety of particles (carbon, melanosomes, which suggest that their presence could interfere THE OUTFLOW SYSTEM IN GLAUCOMA 5

with the capacity to respond to demands for transfer located in the basement membrane of the endothe­ of aqueous fluid and would lead, therefore. to an lium lining the canal of Schlemm and in the fine increased resistance. Paradoxically. it appears that in filamentous material beneath the canal endothelial individuals with a normal blood pressure the monolayer. Type III collagen was located with type I intraocular pressure declines with age pari passu colIagen in the striated collagen fibres in the inner 4H with a decline in the rate of aqueous inflow. so that part of the cribriform layer. The significance of these the system in the normal individual remains in investigations is that type I collagen imparts tensile 52 balance. strength while type III imparts resilience. Type VI 4lJ 51 In recent investigations, MarshalI et al. - have colIagen appears as fine fibres which bind the larger attempted to identify the various constituents of the fibres of types I and III. This collagen type has been extracellular matrix in the cribriform layer using identified by biochemical techniques and the finding immunocytochemical techniques for the identifica­ was substantiated by the demonstration of mRNA tion of collagens and laminin. The collagen types L for type VI collagen in the trabecular endothelial III, IV. V and VI have been located in the trabecular cells.53 The nature of the electron-dense plaques and meshwork using immunogold labels at the ultra­ wide-banded collagen remains controversial. In one structural level. Collagen types I and IV have been immunohistochemical study. type VI collagen was

(a) (c)

., ,. . ) I F: . · ..

�'.,\.' '.�\ ,. '. - . it . . � . . '. .. • • .�:, •••• .. - ' ...; . .. . , ': ' A, "" \ eo I : : \ . �. I) " • . � •. ' ,., ,,� . . 'IL .;.. ' .,;.;.:..' _--:�_ __ (b)

Fig. 4. (a) In complete angle closure rhe traheculae are compressed (arrowheads) and giant vacuoles are ahsent from the lining endothelium ofSchlemm \" canal which contains hlood. (X200) (h) In a .lector where the angle is open and the traheculae appear to he compressed. giant vacuoles are numerous. There is presumahly difjilsion of aqueous within the meshwork at high pressure. (X 2()O) (c) Partial angle closure and Jilsio/1 of the inner traheculae (arrowheads) is associated ..vith identiflahle vacuole�; alheit reduced in numher. (x200) (d) In one case rhe crihrif(Jrm layer was so distended that the collector channel (cc) was occluded hy a hulge: rhe canal is hridged hy cellular strands (arrowheads). (Toluidine hlue stained plastic section.l; X 475) 6 W. R. LEE identified in wide-banded collagen,54 but other Table I. Details of specimens used for the study 'of secondary workers have not confirmedthis. 51 angle closure on the trabecular meshwork Initially, we and others were interested in Case no. Patient Age (years) and sex glycoproteins, which can be demonstrated as a fine Ciliary body melanoma 1. 7360- JM cP 69 F filamentousnetwork in the trabecular interspaces by 2. 73 69- JO 'B 76 F their special affinities for identifiable compounds 3. 75-177 EF 63 F such as ruthenium red.55 Owing to their hydrophilic 4. 78 -66 3 JC 58 M properties, glycoproteins are considered capable of 5. 78 688- MJ 78 F 43 6. 85-778 EJ 50 creating resistance and Ethier et al. believe that the 7. 85-316 Evan B 52 seemingly electron-empty spaces are in fact filled Old choroiditis with gel which contributes significantly to resistance. 8. 81-542 JW 60 M

With better technology, more specific identification Neovascular glaucoma can be achieved and the sulphated proteoglycans, 9. 79-7 12 AB 77M both chondroitin sulphate and dermatan sulphate, 10. 80-03 AMcR 66 M have been identifiedin human trabecular tissue, with aperfused with fixative at 40 cm pressure. the latter predominating.56 More recently, sialic acids have been demonstrated in the extracellular matrix closure glaucoma have not been investigated due to 57 58 using immunogold lectins. Gong et al. claim that successful treatment and the absence of suitable an age-related progressive coalescence of collagen material for study. It is generally accepted that in the was found in the normal trabecular meshwork and primary disease angle closure starts in the upper part this was associated with a decrease in the amount of of the chamber angle and progresses downwards.59 proteoglycans. Further investigation of these incon­ Thus in this condition it is likely that, as the pressure spicuous extracellular matrix constituents may prove rises, the trabecular tissue will be protected where important in an understanding of primary open angle there is complete iridotrabecular contact, while in glaucoma. sectors where there are only narrow synechiae the meshwork will be exposed to high flow of aqueous The Outflow System as a Pressure Sensor due to the high intraocular pressure. To investigate It is not uncommon to find nerves in the human this hypothesis, I studied the chamber angle in a trabecular meshwork, particularly in the scleral spur group of 10 enucleated human eyes (Table I) in and in the cribriform layer, and it is tempting to which there was a partial angle closure in some areas assume that there are neural connections between lD as observed in routine paraffin sections. It must be eye and brain. In a recent study Tamm et al. have stressed that these globes were enucleated in identified structures which have the morphological treatment of secondary angle closure glaucoma. features of mechanoreceptors within the scleral spur. Blocks were taken for electron microscopy and The possibility that there is some form of neurolo­ giant vacuoles were counted in semithin sections gical control of the intraocular pressure raises stained with toluidine blue according to the estab­ possibilities for valuable future research. lished techniques described above.

THE EFFECTS OF VARIOUS PATHOLOGICAL Complete Angle Closure PROCESSES ON THE HUMAN OUTFLOW SYSTEM In the cases in which there was angle closure as far as the peripheral cornea, the iris was compressed The following topics have been selected to illustrate against the inner surface of the meshwork (Fig. 4a). various types of insult which can lead to irreversible The uveal and corneoscleral layers of the trabecular damage to the human outflow system: meshwork were also compressed, although the 1. Sector overperfusion, as in secondary angle trabecular endothelial cells were still identifiable. In closure glaucoma. some cases Schlemm's canal was either reduced to a 2. Cellular infiltration within the trabecular mesh­ narrow slit or was not identifiable; giant vacuoles work, as in 'trabeculitis' due to inflammatory cell were never observed in these Circumstances. In other infiltration. cases the canal was widely dilated. In this situation, 3. Deposition of inert substances in the trabecular we must assume that the total abolition of flow had meshwork, as in the exfoliation syndrome. resulted in reflux of blood from the collector channels in which pressure was high due to EFFECTS OF HIGH PRESSURE ON THE episcleral congestion secondary to raised intraocular HUMAN OUTFLOW SYSTEM IN ANGLE pressure. CLOSURE The pathological effects of angle closure on the Partial Angle Closure trabecular meshwork in the primary forms of angle In sectors in which the chamber angle was open there THE OUTFLOW SYSTEM IN GLAUCOMA 7

Fig. 6. Fragments of iris stroma (arrowheads) attached to the inner surface of the meshwork.

open and the uveal trabeculae were widely spaced, there was inner/outer wall apposition of the Fig. 5. Electron micrograph to show fusion between the endothelium of Schlemm's canal such as was seen iris stroma (arrowheads) and the uv'eal beams which are in the primate experiments (Figs. 2, 4d). This compressed. (X 100) appeared to lead to disorganisation of the canal with residues of cellular strands linking the inner and were plentiful giant vacuoles in the majority of outer walls. In one speCimen it was possible to specimens (Fig. 4b). In some cases the giant vacuole identify a bulge from the cribriform layer blocking a counts reached values of 30-50 per section, which collector channel (Fig. 4d). The various disturbances was greater than that noted in the experiments on seen in the trabecular meshwork in this series of perfusion-fixedeyes ?4 Where iridotrabecular contact secondary closed angle glaucoma cases could account was loose and interrupted, giant vacuoles were still for the well-recognised failure to control intraocular present and in some cases were numerous even pressure by surgical iridectomy or YAG laser though the uveal trabeculae were compressed. The iridotomy in a minority of cases of primary closed latter finding was not entirely consistent with the angle glaucoma.4-6 An increased resistance to working hypothesis, but could be explained by the aqueous outflow could follow fusion of uveal assumption that there was lateral diffusion of trabecular beams or adherence of remnants of iris aqueous through the looser adjacent trabecular stroma to the uveal layers. More importantly, tissues. In other specimens, in which the uveal disintegration of the lining endothelium of beams were compressed and fused (Figs. 4c, 5), the Schlemm's canal would rob the system of the giant vacuoles counts were much lower than those delicately controlled capability of transferring aqu­ sections in which there was obvious space for eous by the giant vacuole bulk flow system. aqueous to pass into the trabecular meshwork. At As a supplement to these conclusions, I have noted the ultrastructural level it was difficult to find any in a separate study of routine trabeculectomy space between the iris stroma and the collagenous specimens removed in treatment of closed angle cores of the trabeculae (Fig. 5). Uveal trabecular closure unresponsive to primary intervention (iri­ fusion occurred in the absence of iridotrabecular dectomy or iridotomy), the presence of inflammatory contact in some cases, which implies separation of cells, usually lymphocytes, in the intertrabecular pre-existing contact. spaces. This suggests that in the acute primary form In two cases it was possible to identify fragments of of closed angle glaucoma a rapid rise in intraocular iris stroma attached to the uveal trabeculae - a pressure had caused ischaemia in the iris stroma. feature which could also be interpreted as an Release of inflammatory mediators (e.g. prostaglan­ indication of separation of pre-existing synechiae dins and prostacyclins) would promote migration of (Fig. 6). Thus fragments of fibrous tissue derived inflammatory cells into the iris stroma and the from the iris stroma and firmly adherent to the inner anterior chamber with subsequent migration into surface of the meshwork could be responsible for the intertrabecular spaces. In many of the trabecu­ obstruction to aqueous outflow. lectomy specimens the presence of numerous One of the most interesting findingsin terms of the melanin granules in the trabecular endothelial cells effects of overperfusion was that where the angle was was noteworthy; this could be explained by ischaemic 8 W. R. LEE

(a) (b)

Fig. 7. Electron microscopy of polymorphonuclear leucocytes infiltrating (a) the inner part of the trabecular meshwork and (b) the outer part. In the inner part contact between dying polymorphonuclear leucocytes and trabecular endothelial cells destroys the latter (arrowheads). The dissolution of the trabecular beams (arrowheads) in the outer part of the meshwork is due to lysis of the leucocytes with release of proteolytic enzymes. (each X 1000) necrosis of the pigment epithelium or could have remainder of the outflow system. This assumption represented damage by surgical intervention or YAG has not been investigated by pathological studies and laser treatment, both of which could have disrupted the deleterious effects of massive infiltration of the the pigmented cells of the iris. trabecular meshwork by polymorphonuclear leuco­ The consequences of inflammatory cell infiltration cytes has not been documented at the ultrastructural and melanin overload are discussed in the following level. It could be assumed that release of proteolytic section. enzymes would dissolve the collagenous cores of the trabeculae if the protection afforded by the trabe­ RESPONSE OF THE OUTFLOW SYSTEM TO cular endothelial cells were lost. In the most severe CELLULAR INFILTRATION example in my files there was a considerable degree Amongst the various forms of secondary open angle of disruption of the collagenous cores where the glaucomas, those which are due to inflammatory endothelial cells had separated (Fig. 7) or had been disease are of clinical importance because of the destroyed by leucocytes in contact with the cells. In challenges they present in terms of diagnosis and addition there appeared to be an almost total loss of management. For the pathologist these disorders are matrix in the cribriform layer, presumably due to of interest in terms of the effects of the presence of release of proteolytic enzymes by the degenerate infiltrating cells on the resident cells. Although in leucocytes which were trapped there. many cases encountered in routine pathology the Detailed morphological information concerning infiltrate is of a mixed nature, consisting of the effects of the later stages of a pyogenic infection polymorphonuclear leucocytes, lymphocytes, plasma is also sparse because in those cases proceeding to cells and macrophages, in some disorders the enucleation, the rapid progress of pyogenic inflammatory response essentially involves one endophthalmitis is so destructive that ultrastructural population of cells. While the term 'trabeculitis' investigation would appear unrewarding. The ante­ should apply to all forms of cellular infiltrate, it is rior chamber in routine histopathological prepara­ most commonly applied in clinical usage to lympho­ tions contains a fibrinopurulent exudate in the plasmacytoid infiltration in chronic non-specific inferior angle and this is associated with synechia anterior uveal inflammatory disease. . formation due to hypotonia and fibrinous adhesions between the iris stroma and the trabecular mesh­ Polymorphonuclear Leucocytes: Hypopyon work. In addition the end-stage is complicated by From clinical experience it would appear that the fibrosisof the trabecular meshwork and fibrovascular outflowsystem recovers effectively after exposure to proliferation in the scleral sulcus. polymorphonuclear leucocytes. In the case of a small hypopyon any permanent damage would be Lymphoplasmacytoid Infiltration (Non-Granuloma­ restricted presumably to the inferior part of the tous Trabeculitis) chamber angle and this would not have any The clinical features of all forms of non-granuloma­ significant effect on the functional capacity of the tous trabeculitis have been comprehensively THE OUTFLOW SYSTEM IN GLAUCOMA 9

Table II. Details of specimens used for the study of non -g ranulomatous trabeculitis

Case no . Patient Age (years) and sex Clinical details

Herpes simplex keratitis 1. 84-680a JM cCF 69 F Iritis, normal lOP 2. 88-901 a JB 75F Iritis, normal lOP 3. 71 -208" TP 67 M Iridocyclitis. hypopyon 4. 80- 01 5 AS 69 F Iritis, partial angle closure 5. 86- 31 7" JR 70 M POAG, trabeculectomy, .iritis 6. 91 -1 73a JW 81 M Iritis. glaucoma 7. 87-681a CG 62 M ?HSK, acid burns, keratoplasty, iritis

Fuchs' heterochromic iridiocyclitis 8. 71-22 5 AP 60 F 9. 87-735"·b BT 48 M 10. 92-21 9 RH 51 M 11 . 93-167 SO'B 38 M

Miscellaneous 12 . 82-876 MD 71 F RD, iritis, trabeculitis 13. 85-1168a MD 70F HZO, iritis, fibrosis of Schlemm's canal rop, intraocular pressure; POAG. primary open angle glaucoma; HSK, herpes simplex keratitis; RD, retinal detachment; HZO, herpes zoster ophthalmicus . "Cases studied by electron microscopy . bTrabeculectomy specimen .

reviewed in many standard texts.4,6,60 The response lead to stripping of the beams, fluidmovement into of the outflow system to lymphocytic infiltration in the trabecular collagenous core and narrowing of anterior uveitis is said to be variable: many cases lead the intertrabecular spaces. 1 to hypotonia, only some to a raised pressure.6 3. Obstruction to outflowdue to mechanical blockage Before the advent of effective anti-viral therapy, by the cellular infiltrateon the inner surface of the one of the commonest causes of trabeculitis was meshwork, or in the intertrabecular spaces.60·70 herpes simplex keratouveitis, and a raised intraocular 4. At the end-stage, sclerosis of the meshwork would 62 pressure occurred in 28% of cases. Because of the be associated with an endothelial downgrowth and striking iritis associated with keratitis and trabeculi­ the formation of a hyaline membrane.71 tis, the clinical diagnosis of herpes simplex keratitis is easily confirmed histopathologically but cannot be There is no documentation to my knowledge of the proven at the later stages by identification of viral ultrastructural changes occurring in the trabecular particles at the electron microscopic level. In an early meshwork in lymphoplasmacytoid trabeculitis, nor light microscopic study, Hogan et al.,63 who coined has there been a report on the immunohistochemical the term 'trabeculitis', showed infiltration of the characteristics of the infiltratingcell s. In addition the intertrabecular spaces by lymphocytes and plasma changes occurring in the intrinsic cells have not been cells within fibrin and this has been confirmed investigated at the ultrastructural level. subsequently by a number of authors (see Spen­ To investigate this process further, 12 surgical cer64 for a review). Striking lymphoplasmacytoid enucleation specimens in which trabeculitis had been infiltrations are not confined, however, to herpes reported in routine paraffin sections were retrieved simplex infections, but are also a feature of herpes from file.In this series, the commonest pathology was zoster and Fuchs' heterochromic iridocyclitis. Scler­ herpetic keratouveitis (Table II). Each of the cases in itis and episcleritis are also complicated by an intense the herpes simplex group was characterised by a trabeculitis.65 severe non-granulomatous iritis and a lymphocytic The pathogenesis of cell and matrix destruction in and plasma ceil infiltratein the trabecular meshwork. the outflowsy stem in trabeculitis has not been widely In two cases (84-680 and 88-901) the globes were investigated and review articles66 are somewhat normotensive; the remainder showed pathological speCUlative. The possible mechanisms suggested for evidence of glaucoma. This is something of a outflow obstruction in uveitis include: heterogeneous group since in one of the specimens there was partial angle closure (85-015) and the 67 1. Mechanical blockage by serum or plasma seventh case (87-681) was presumed to be due to released from iris vessels in which the perme­ herpes simplex, but was complicated by the fact that ability had been altered by inflammatory media­ the cornea had been damaged by an acid burn and a 6 69 tors. 8 , corneal graft had been performed. 2. Damage to trabecular endothelial cells by enzymes Four cases of Fuchs' heterochromic iridocyclitis or cytokines released by inflammatory cells. The were available for study and one of these was a surviving trabecular endothelial cells would also be trabeculectomy specimen (87-835). Two other mis­ required to phagocytose cell debris, which could cellaneous specimens exhibiting iritis were included 10 w. R. LEE

Immunohistochemistry Ten of the enucleations specimens cases were subjected to immunohistochemical investigation in order to classify the constituents of the lymphocytic infiltrate. In two specimens (87-735 and 82-876) there was total fibrosis of the outflow system and immunohistochemistry was not considered to be justifiable. For T-Iymphocytes, the markers employed were UCHLl and MT1 (CD3 in some cases), and for B-Iymphocytes the markers were MB1 and L26. Because there was obvious segmenta­ tion of endothelial cells in the scleral sulcus (an apparent 'neovascularisation' of Schlemm's canal) in some of the cases the following markers for endothelial cells were employed: Ulex europaeus, factor VIII antibody and Q Bend 10. The antibody reaction was visualised with ethylaminocarbimazole to avoid the confusion with melanin which occurs when the peroxidase-antiperoxidase label is used. Although the primary fixation was glutaraldehyde, the staining reactions in the tissues of the trabecular meshwork were considered reliable if there was unequivocal staining of (a) the inflammatory cells in the iris and (b) the endothelium of the blood vessels in the iris, ciliary body and conjunctiva. In general the labelling for T cells was more intense than for B cells. In eight cases, blocks had been taken from the chamber angle and processed and embedded for electron microscopy: five cases were due to herpes simplex keratitis, one to post-'acid burn' keratitis, one to Fuchs' heterochromic iridocyclitis and one to post-herpes zoster ophthalmitis. The results of this investigation were as follows:

Herpes Simplex Trabeculitis By conventional light microscopy, the most consis­ tent pattern in less severe disease was a uniform infiltrationof plasma cells and lymphocytes through­ out the intertrabecular spaces, with apparent pres­ ervation of the indigenous trabecular endothelial cells: infiltrationin the outer wall of Schlemm's canal Fig. 8. Immunohistochemistry of lymphoplasmacytoid was less extensive. Vacuoles were not observed in the trabeculitis. The upper figure demonstrate an intense T lining endothelium of Schlemm's canal in any of the cell infiltrate. In the middle figure the marker for B cells specimens. By immunohistochemistry the infiltrate stains fewer cells which are located in the outer part of the was characterised by a predominant T-Iymphocyte meshwork. In the lower figure the marker for endothelial cells outlines small blood vessels within the trabeculae. The population, with only a small minority reacting with scleral sulcus is filled with fibrovascular tissue. (Original the B-Iymphocyte markers (Fig. 8); such cells when magnifications X 250) present were located more commonly in the outer part of the meshwork. The trabeculae were lined by attenuated endothelial cells and there was only for comparison: one a complicated retinal detach­ occasional evidence of melanophagocytosis. The ment, the other a burnt-out herpes zoster infection. lymphocytes in the iris labelled strongly for both B Five control specimens (enucleations for posterior and T cells; the former were more diffuse, the latter melanomas in patients of a similar age) were in focal clusters. investigated using the same morphological techni­ By electron microscopy, lymphocytes and plasma ques. cells were prominent and often occupied the surface THE OUTFLOW SYSTEM IN GLAUCOMA 11

in the cribriform layer and the outer wall became altered with the replacement of the normal hetero­ geneous mixture of fine collagenous matrix, elastic­ like material and curly collagen by a fine network of fibrillarmater ial (Fig. 9). The normal, rather stellate cells of the cribriform layer were replaced by enlarged spindle cells, which had the characteristics of fibroblasts with prominent rough endoplasmic reticulum (Fig. 9). The endothelial cells lining Schlemm's canal did not possess the pegs and attachments normally found in the lining endothe­ lium of Schlemm's canal and the sulcus in advanced disease was often filled by a series of capillary-like structures surrounded by fibrous tissue. Giant vacuoles were never observed and there was no evidence of cell migration through the endothelial monolayer. In parts of the meshwork there was Fig. 9. Electron micrograph of the scleral sulcus in advanced trabeculitis. Note the atypical blood vessels (v) fusion of swollen trabeculae and in this circumstance without cytoplasmic pegs to the abnormal underlying the endothelial cells became swollen and formed collagenous connective tissue. Note the fibroblast (arrow) circular capillary-like structures (Fig. 10) with a and the plasma cells (arrowheads). (X 1200) basement membrane. It was possible to identify cells resembling pericytes surrounding the endothe­ of denuded uveal trabeculae, which did not appear to lial cells in some examples. be swollen. Thin trabecular endothelial cells contain­ These capillary-like structures were found ing occasional melanin granules lined the intertrabe­ throughout the uveal and corneoscleral layers of cular spaces, but there was no indication of the meshwork in the five herpes simplex cases and in phagocytosis of cell debris by these cells. Degenera­ the herpes zoster case using immunohistochemical tive or cytolytic changes were not observed in the markers for mesoderm-derived endothelial cells trabecular endothelial cells and fibrin and cell debris (Table III). In keeping with the reports in recent was not identified in the intertrabecular spaces. Viral literature,8 7, 2 the cells lining the wall of Schlemm's particles were not seen in the tissues. canal were found to react in a similar manner to the In the more severely affected tissue the inner part other mesoderm-derived endothelial cells within the of the trabecular meshwork became compressed, eye (Fig. 8c). although there was no striking loss of endothelial The interpretation of this form of vascularisation cells nor were the trabecular beams excessively occurring within the trabecular meshwork can only swollen. More significantly, the extracellular matrix be speculative at this stage. What seems unlikely

(a) (b)

Fig. 10. Abnormal blood vessels in the intertrabecular spaces in trabeculitis. In (a) the endothelial cells are surrounded by prominent pericytes (arrowheads); in (b) the structure appears to be formed from endothelial cells (arrowheads). (Transmission electron micrographs; a, X 3200; b, X 4700) 12 w. R. LEE

Table m. Immunohistochemical results in trabeculitis Case no. Age (years) and sex Vascular proliferation in meshwork T-lymphocytes B-lymphocytes Herpes simplex 84-680 69 F + +++ + 88 -901 75 F ++ +++ + 85 -015 69 F + +++ + 86 -317 70M +++ ++ + 91 -17 3 81 M ++ +++ +

Acid bums 87 -6 81 62M + +

Fuchs' heterochromic iridocyclitis 92-21 9 51 M + +

Herpes zoster 85 -1 16 8 70 F + + +

from the distribution of these channels is that they 2) which stimulate migration and proliferation of were derived from a neovascular membrane on the further generations of T cells. T-Iymphocytesrelease inner surface of the meshwork. It also seems unlikely a wide spectrum of lymphokines and adhesion that the vessels represent extensions from the newly molecules which in addition to the induction of B­ formed fibrovascular tissue in the scleral sulcus, lymphocyte activity can modify the phenotypic because the endothelial cells in these vascular characteristics of cells and hence their behaviour. It channels were not accompanied by pericytes. A is known that B-Iymphocytes are a source of more plausible suggestion is that the new capillaries interleukin I-alpha and interleukin I-beta, both of were derived from ingrowths from the capillary bed which stimulate fibroblastic activity (fibroblast pro­ in the ciliary muscle. liferation in the scleral sulcus was one of the most The morphological results obtained in the tissue of interesting observations made by electron micro­ the presumed herpes simplex case (87-681) were far scopy in the present study). The B-Iymphocytes less impressive than those in which there was which were identified in the outer part of the definitive clinical information to incriminate herpes meshwork could therefore be responsible for the simplex, so to some extent this 'non-specific' case stimulation of fibroblastic activity in the scleral serves as a control and this reinforces the pattern sulcus. Alternatively, fibroblastic activity and seen in herpes simplex trabeculitis. endothelial cell proliferation could be due to the The results of this investigation can be considered presence of growth factors such as transforming in relation to the rapidly expanding understanding of growth factor beta, known to be present in the immune recognition system in the trabecular aqueous?7 Trabecular endothelial cells have been meshwork.9 7, 3-78 It is now evident that the tissue is an shown to possess receptors for transforming growth immunological filter with a resident population of factor beta and it could be aberrations in this cells in the meshwork capable of HLA class II pathway which lead to the fibrous replacement of antigen presentation?3-75 The upregulated recogni­ the canal which is so common at the end-stage of a tion system in normal trabecular endothelial cells variety of disease processes. Finally, it was encoura­ could also amplify the activity of the dendritic cells ging to discover that the presence of lymphocytes which are now known to be resident in the iris and and plasma cells in the meshwork did not appear to the trabecular meshwork of the human and lower cause significantdeleter ious changes in the cytoplasm mammals.76 of the trabecular endothelial cells. Thus, in the less In the context of viral infection, the herpes virus severely affected cases, the trabecular beams were has the capacity to upregulate resident cells in tissue not swollen. to express HLA class II through the release of gamma-interferon which also stimulates resident Miscellaneous Cases macro hages to produce cytokines of various The only useful information derived from the two ,f? types. Gamma-interferon (G-IFN) is a protein cases included as 'miscellaneous' was that in herpes that has been shown to induce class II major zoster ophthalmicus the predominant inflammatory histocompatibility antigens which inhibit normal cells in the meshwork were T-Iymphocytes (with a and transformed cell proliferation and modulate minority staining with the B markers) and there was macrophage activation including inhibition of pha­ immunohistochemical staining for mesoderm-derived gocytosis: this is effected by disturbing the actin endothelial cells. The vascular proliferation within cytoskeleton of the cells. The immune recognition the trabecular endothelial cells is obviously a system in herpes simplex infection also has the manifestation of a response to the abundant potential of presenting antigen to T cells which are lymphoplasmacytoid infiltration in the meshwork in then stimulated to secrete cytokines (e.g. interleukin herpes simplex and in herpes zoster, but I have no THE OUTFLOW SYSTEM IN GLAUCOMA 13 useful information to offer for the actual nature of trabecular meshwork by dense fibrous tissue is the immunological signalling system. Modern mole­ almost a pathognomonic histopathological feature cular pathology, however, will have the capability to of this condition. solve this problem in due course. Steroid Glaucoma Heterochromic Iridocyclitis It must be assumed that many of the cases of trabeculitis were treated with steroids and this raises The clinical features of heterochromic iridocyclitis the interesting question of steroid-induced resistance have been the subject of comprehensive reviews in in the outflow system. Morphologists have not been the recent literature.78 79, In discussing the aetiology able to provide definitive morphological evidence to and pathogenesis, La Hey et at.80 came to the explain the increased resistance in steroid glaucoma. conclusion that the disease is a secondary phenom­ In two studies,83,84 plaques of fibrillar material were enon with a spectrum of clinical features and multiple found within the outer part of the cribriform layer, causes. In the literature, pathological studies are rare, but few other authors have been able to confirm this but there is general agreement that the inflammatory and in my own material I have made the morpho­ round cell infiltrate is not as massive as in other logical diagnosis of corticosteroid glaucoma on only forms of trabeculitis and that at the end-stage one occasion. The effects of steroids on the intractable glaucoma is due to trabecular fibrosis trabecular meshwork cells in vitro have been studied and in some cases neovascular glaucoma?8 8, 1 8, 2 extensively, particularly with regard to secretion of In my own collection, in two specimens of extracellular matrix.85-9o Such studies have shown heterochromic iridocyclitis, the meshwork contained that the cultured endothelial cells become enlarged a scattered inflammatory cell infiltrate(Fig. 11) which with an increase in smooth endoplasmic reticulum on immunohistochemical analysis proved to be a and rough endoplasmic reticulum which is in keeping mixed T and B cell population. This was of a different with a demonstration of synthesis of, for example, pattern and present to a far lesser degree than the glycosaminoglycans,fibronectin , collagen and elastin. inflammatory cell infiltrationseen in herpes simplex Proliferation and phagocytic activity are reduced in keratitis. There was no evidence of capillary buds the presence of steroids in vitro, but the relationship within the trabecular meshwork. A puzzling feature between this and altered function in vivo is not clear. was that in these specimens pigment deposition was not particularly noticeable in the meshwork; this is surprising in view of -the atrophy of the iris pigment Obstruction of the In tertrabecular Spaces by epithelium and stroma which occurs in the disease. In Macrophages two cases the meshwork was fibroticand in one there For the morphologist the most visually attractive was an endothelial downgrowth, which reinforces the infiltration of the meshwork is that seen when a clinical observations concerning the difficulty in macrophagic response is activated: the mononuclear management of glaucoma in heterochromic iridocy­ cells are large and easily seen by all forms of clitis.4 5, 6, 78, 79, Indeed, total replacement of the microscopy due to the presence of phagocytosed material. The interest for immunologists is that this

Fig. 11. In Fuchs ' heterochromic iridocyclitis the inflam­ Fig. 12. In phacolytic glaucoma large spheroidal macro­ matory infiltrate is sparse in both iris and meshwork; phages line the inner surface of the meshwork and block the fibrosis begins in the outer part of the meshwork (arrow­ intertrabeclilar spaces (arrowheads). (Scanning electron heads). (Paraffin section, H&E, X 70) micrograph, X 670) 14 W. R. LEE

(a) (b)

Fig. 13. In (a) a phacomacrophage is jammed in an intertrabecular space (arrow); small clumps of lens matter can be identified in the cytoplasm of the endothelial cells (arrowheads). In (b) the electron micrograph demonstrates the wide variety of abnormal structures present in a degenerating lens cortex; these can be seen in the phacomacrophages. (a, X 3000; b, X /0 000) response does not apparently involve the lympho­ sion electron microscopy was that the enlarged plasmacytoid arm of the antigen recognition system. macrophages completely blocked the interspaces in The macrophagic reaction to endogenous particulate the inner part of the meshwork and that there was material may be a simple chemotactic response as very little infiltration into the outer part of the has been shown to occur in in vitro experiments in meshwork (Figs. 12, 13a). This was because the cells which lens protein acts as a chemoattractant to were presumably too large to insinuate themselves 9 suspensions of macrophages. 1 This begs the question through the intertrabecular spaces. Thus it would as to how the recognition arm of the reaction (see the seem unlikely that washout of the anterior chamber section on lymphoplasmacytoidinfiltration ) is primed would relieve this form of obstruction once estab­ to 'call up' the monocytes which enter the anterior lished. chamber via the iris vasculature. In these specimens there was no evidence of cell Clinical entities in which a massive macrophagic debris which could represent fragments of degen­ reaction is triggered by the immune detection system erating lens cells lying loose in the intertrabecular in the outflow system are seen when the aqueous spaces. The cytoplasm of the macrophages contained ( ) fluid contains lens matter phacolytic glaucoma , globules, membrane-bound vesicles and granular ( ) melanosomes melanomalytic glaucoma and blood bodies identical to those which are found in the with its breakdown products (haemomacrophagic degenerating lens fibres (Fig. 13b) in the senile glaucoma). human cataract.93-96 Electron microscopy revealed that the phagocytic activity of the exogenous Phacomacrophages in Phacolytic Glaucoma macrophages was more than adequate to protect The clinical entity now recognised as 'phacolytic the trabecular endothelial cells from the burden of glaucoma' was identified at the pathological level by phagocytosis of lens debris; only occasionally was Flocks et al.92 in 1955. This paper demonstrated the there any evidence that the endothelial cells importance of mechanical obstruction of the outflow contained fragments of degenerate lens matter system by macrophages which had phagocytosed (Fig. 13a). degenerate lens protein released from the liquefied Pathologists have long recognised phacolytic cortex of a cataractous lens. While there has been glaucoma as a distinct entity and clinically this can adequate documentation subsequently at the light now be diagnosed at an early stage by examination of microscopic level, ultrastructural investigations are an aspirate from the anterior chamber in a suspected sparse. With Professor Grierson, I studied three cases case. The phacomacrophages stain positively with the of phacolytic glaucoma (unpublished data) with the PAS stain and react positively with immunohisto­ purpose of investigating the involvement of the chemical markers for macrophagesY7 trabecular endothelial cells in the phagocytic pro­ It should also be noted that crystalline material is cess. What was apparent by scanning and transmis- present too in such preparations and examination THE OUTFLOW SYSTEM IN GLAUCOMA 15

Fig. 14. Melanomacrophages (arrows) infiltrating the outer part of the meshwork in a case of melanomalytic glaucoma; trabecular endothelial cells and the endothelial cells lining the canal also contain melanosomes (arrowheads). (Trans­ mission electron micrograph, X 2400) Fig. 15. Intact and fragmented red cells (arrowheads) are will reveal the presence of birefringent calcium present in the cytoplasm of macrophages after prolonged oxalate which is commonly observed in the sclerotic exposure to bleeding into the anterior chamber. Where the nucleus of a Morgagnian cataract. The biochemical endothelial cell cover is lost, the trabecular beams are derivation of calcium oxalate in a nuclear cataract is swollen (arrows). (Transmission electron micrograph, X 2000) unexplained, as is the identificationof the cholesterol crystals which have also been observed in the packed with extremely large macrophages containing aspirate.97 primary and secondary melanophagosomes (Fig. 14). Both morphologists and clinicians must take note There was only a moderate quantity of melanopha­ that the outflow system can be blocked by lens gocytosis in the trabecular endothelial cells, which protein or degenerate lens matter in the absence of a m macrophagic response. Presumably the capsular could be interpreted to mean that the melanoso es rupture occurs so rapidly that the immune recogni­ released from the necrotic tumour were already tion system does not have time to recruit the phagocytosed before the trabecular endothelial cells macrophage responseY8.99 The opportunity for could be exposed to the particles. The activated pathologists to study this eminently curable disease endothelial cells bear some resemblance to neoplas­ (by anterior chamber washout) is of course rare, but tic cells and this is relevant to the fact that in one trabeculectomy specimen I was able to meianomalytic glaucoma is not an altogether benign confirm that the interspaces of the meshwork process since it has been possible to demonstrate contained lens debris only: macrophages were not viable tumour cells interspersed between the macro­ identified. phages by use of an immunohistochemical marker 1 1 1 such as S-lOO and by electron microscopy. 00, 0 Melanomacrophages in Melanomalytic Glaucoma While melanomalytic glaucoma is an entity always Massive spontaneous necrosis in a ciliary body included in standard texts, it is a great rarity in melanoma, a ciliary body melanocytoma or a clinical and pathological practice and rather surpris­ choroidal melanoma leads to release of myriads of ingly it has not been found to occur in proton beam melanin pigment granules into the anterior chamber treated melanomas, in which there is much more and an acute rise in intraocular pressure. In one case tumour necrosis than is usually seen after other on file100 we found that the outflow system was forms of therapeutic irradiation. 16 W. R. LEE

Glaucoma Occurring in Response to Intraocular cytoplasm, were present in the interspaces of all Haemorrhage layers of the meshwork, there was no evidence of Haemomacrophages damage or degeneration in the trabecular endothelial Although malleable red cells can pass easily through cells. Indeed the endothelial cells were so 'activated' the outflowsystem in small numbers in experimental that the presence of enlarged nuclei and phagocy­ models,102 it has long been recognised that the tosed melanin granules made it difficult on occasion outflow system is obstructed when there is pro­ to distinguish the malignant cells from the activated longed bleeding into the anterior chamber.103 Red normal cells. In four cases in a relatively large cells break down readily with disruption of the cell personal series of 60 iris melanomas there was membrane and release of haemoglobin fragments. evidence of spread into Schlemm's canal, but this The cell fragments are eagerly phagocytosed by did not appear to result in a greater risk of metastatic macrophages and these large cells obstruct the death: none has been recorded as yet after 5-10 intercellular spaces in combination with fibrin and years. It is, however, important to note that in a small non-phago tosed cell debris. In our own initial proportion of cases, iris melanomas if left in situ will studieslO4,1O� we were impressed by the involvement transform into cells which have a much more of trabecular endothelial cells in the phagocytic malignant phenot pe with the appearance of large O � process and it was attractive to postulate that the epi thelioid cells.l activation of the endothelial cells leads to their While there is obvious damage to the meshwork separation from the trabecular beams and their cells in the pseudohypopyon produced when a migration out of the system. Loss of endothelial necrotic retinoblastoma releases debris into the cells by this process of stripping was thought to be anterior chamber, there is to my knowledge no important in that the collagenous cores became available information concerning the changes which swollen by the absorption of aqueous and this are observed in the trabecular endothelial cells at the could lead to a narrowing of the intertrabecular ultrastructural level. spaces (Fig. 15). Blood Products Released from a Vitreous Haemor­ OBSTRUCTION OF THE OUTFLOW SYSTEM BY rhage: Erythroclastic or Ghost Cell Glaucoma ABNORMAL EXTRACELLULAR MATERIALS Haemomacrophagic obstruction of the outflow The importance of the phagocytic activity of the system is less significant when blood passes through trabecular meshwork in maintaining the aqueous the vitreous face after a long-standing vitreous outflow pathway by clearing obstructive debris and haemorrhage. In this situation, incarcerated red particulates from the channels has already been cells undergo spontaneous lysis with the passage of discussed. Pathological investigations and in vitro time and assume a spherical shape which endows the studies have shown that the trabecular endothelial property of rigidity. Release of such cells into the cells in the natural environment are facultative and anterior chamber causes a transient glaucoma which highly discriminatory phagocytes which can remove can be diagnosed by the presence of lysed red cells or endogenous particulate material such as melanin 'ghost' red cells in an aqueous tap. The term 'ghost granules and the breakdown products of blood from cell glaucoma' is appropriate for this process, the aqueous. It has been suggested that overload of although 'erythroclastic glaucoma' is an acceptable 106 the self-cleansing biological filter mechanism may alternative. In my own filesit has not been possible contribute to the pathogenesis of certain types of to find a completely monocellular ghost cell infiltrate obstructive glaucomas.4o The intrinsic natural parti­ in the trabecular meshwork and there has always cles to be included in this context are red cell been an admixture of macrophages in cases of such fragments and lens debris (see above) , melanosomes, severity that enucleation has been required. and exfoliation material. The non-biological exogen­ ous materials to be considered include emulsified In vasion by Neoplastic Cells silicone oil and Haelon. The latter, however, causes The many mechanisms which lead to glaucoma in an only transient glaucoma when used in cataract eye containing a malignant uveal tumour have been surgery and to my knowledge this cause of a raised reviewed by Shields et al. 1O In terms of the biology of intraocular pressure has not been investigated by the trabecular meshwork it is necessary to consider morphology in human tissue. only the effects of tumour cell infiltration on the In the various pathological situations, an overload endothelial cells. My own studies of neoplastic of the system can have very serious consequences in melanocytic infiltration of the trabecular meshwork terms of the functional and anatomical preservation have been confinedto ring spread in iris melanomas of the tissue. However, in some circumstances the and ciliary body melanomas. While tumour cells, system has a surprising capacity for recovery, as in identifiedby the presence of premelanosomes in the some forms of pigmentary glaucoma. THE OUTFLOW SYSTEM IN GLAUCOMA 17

spindle. Paradoxically, good experimental evidence has been obtained to show that pigment granules per se within the extracellular spaces of the meshwork do not contribute significantly to resistance.l 1l Histopathological and electron microscopic inves­ tigations have been in agreement that after a massive overload of melanin granules to the trabecular endothelial filtering system, there is separation of the endothelial cells and collapse of the trabecular meshwork.1 l2-1 l 6

(Ps eudo) exfoliation Glaucoma The clinical6 ,1 1 7 and conventional pathological fea­ tures64 ,l 18 of the 'pseudo' exfoliation syndrome are well documented. In paraffin sections, exfoliation substance appears as eosinophilic amorphous clumps on the ciliary processes, the zonular fibres, the anterior lens capsule and on both anterior and posterior surfaces of the iris. The morphological demonstration of a dense deposition of exfoliation substance in the cribriform layer and on the inner surface of the meshwork goes some way towards Fig. 16. Exfoliation material in the cribriform layer explaining why the response to pilocarpine is less appears to invaginate the endothelial cells (arrows). Note l19 the strands of endothelial cells crossing the canal (arrow­ effective in the exfoliation syndrome. Electron microscopy is required to demonstrate heads) which suggests previous inner/outer wall contact. l (Transmission electron micrograph, X3000) the fibrillarna ture of the material, 1 8 which in some cases can be identified within giant vacuoles in the Pigmentary Glaucoma lining endothelium of Schlemm's canal (Fig. 16). Few authors have been able to confirm the original The clinical and pathological features of aqueous 1 0 outflow obstruction due to overload by melanin observations by Ringvold 2 that the trabecular granules have been reviewed comprehensively.109 ,1 10 endothelial cells have the capacity to phagocytose The main source of free melanosomes in the exfoliation substance. In my own cases, the trabe­ pigmentary glaucomas is the iris pigment epithe­ cular endothelial cell morphology has been well preserved without evidence of intracytoplasmic lium, but pigment may be released also from the exfoliation substance. The electron microscope is heavily pigmented iris stroma. Thus any form of also required to demonstrate exfoliation substance in damage to the iris stroma (e.g. contusion angle the walls of iris vessels, which are extensively altered deformity) can lead to pigment deposition in the in the disease process,121 so much so that some meshwork; indeed such deposition is a feature of the authors have suggested that deposition at this site normal ageing process in some individuals when it is leads to a lowering of the oxygen tension in the due to iris atrophy. More extensive deposition can aqueous, which could potentially have a serious occur after lens surgery when the monolayer is effect on the viability of the trabecular endothelial damaged by removal of a cataractous lens and cells. 122 insertion of an implant, but as a secondary In addition to the nomenclature ('pseudo' or 'non­ phenomenon trabecular endothelial pigmentation is pseudo-' exfoliation), controversies persist in many striking in the exfoliation syndrome, in which there is aspects of this disease. For a pathologist, the site of progressive breakdown of the iris pigment as the origin (or synthesis) and the biochemical composi­ exfoliation substance accumulates within the cells tion of exfoliation substance remains a major and on the surface. The most impressive accumula­ challenge. For the clinician, the extent to which tion of melanin in the trabecular endothelial cells exfoliation substance deposited in the outflowsystem occurs in the pigment dispersion syndrome and in contributes to resistance in open angle glaucoma is pigmentary glaucoma, two conditions in which the not yet defined.6 ,117 Undoubtedly in some parts of aetiology is unknown. There are many unexplained Europe, such as Greece, the disease is common and clinical features such as the rarity in Asians and constitutes a major part of the practice of a specialist blacks and the clinical observation that sometimes in in glaucoma.117 In addition, the management of open pigmentary glaucoma the trabecular meshwork is not angle glaucoma in cases in which the exfoliation pigmented even though there is a Kruckenberg deposits are prominent and heavy, can be more 18 W. R. LEE

A major step forward in our appreciation of the complexity of the exfoliation syndrome is the fact that this is a systemic disease. Exfoliation substance can be identified in the conjunctiva and the orbital tissues (blood vessels, extraocular muscle, optic nerve sheaths and connective tissue septa). More recently exfoliation substance has been identified in skin, lungs, heart, liver, kidney and the cerebral meninges.124,126,127 These findings help to resolve the controversy of the origin of exfoliation substance within the eye, since the material appears to be an extracellular matrix product secreted by many types of connective tissue cell. Another fruitful technique applied to the constituents of exfoliation substance has been to identify carbohydrate epitopes. For example an epitope 'HNKl' is present on subsets of glycolipids and glycoproteins which are implicated in cell adhesion. These epitopes label strongly in exfoliation material and they may serve as a binding agent in the assembly of the complex molecules which make up exfoliation substance.1 30

Silicone Oil Glaucoma Much of the published work on the effects of silicone oil on the ocular tissues has concentrated on the dispersion of emulsifiedoil within the vitreous and its presence in macrophages in preretinal mem­ Fig. 17. Silicone oil droplets in the cytoplasm of macro­ 3 phages in the intertrabecular spaces (arrowheads). Some of branes.1 ! The complication of secondary open the macrophages also contain intracytoplasmic melano­ angle glaucoma is now recognised1 32 but descrip­ somes (arrows). This unusual specimen also contains a tions are sparse and confined to posterior segment capillary (c) in the intertrabecular space. The tissue was pathology in human eyes 133 and in animal eyesP4 taken from a glaucomatous eye after treatment of retinal Various authors have speculated that there is detachment using intravitreal silicon. (Transmission electron mechanical blockage either by the oil or by micrograph, X 1000) macrophages containing oil, with the added possibi­ difficult than in cases of primary open angle lity that there may be damage to endothelial cells and glaucoma. secondary swelling of the trabecular beams. In the Histochemistry and immunogold electron micro­ two cases on file, the predominant effect was scopy have been useful tools in the identification of occlusion of the intertrabecular spaces by macro­ the constituents of exfoliation substance and the phages containing spaces once occupied by oil groups led by Streeten,118,123,124 Schloetzer-Schre­ droplets of varying size. Very few trabecular hardtI25-127 and KonstasI21,128,129 have made major endothelial cells contained oil droplets, although contributions in this field. To date the following occasional cells were found with small inclusions substances have been identified within exfoliation (Fig. 17). substance: elastin-related proteins (e.g. elastin),123,124 fibrillin, amyloid p, 129 gp 115 (an elastin-related PRIMARY OPEN ANGLE GLAUCOMA glycoprotein),129 vitronectin and the basement­ The pathogenesis of primary open angle glaucoma membrane-related proteins (e.g. heparan sulphate, remains unknown and both the clinical and patholo­ laminin,128 nido en/entactin, chondroitin sulphate gical literature is filled with speculation and uncer­ � 1 and fibronectin1 5- 27). The technology currently tainties.4,5,6,48 This is without doubt a reflectionof the employed is not yet totally reliable and the subtle and complex nature of a multifactorial disease. immunohistochemical approach requires refinement In addition many pathological reports have utilised to eliminate the discrepancies now apparent between trabeculectomy specimens, which in general are not reports. A fundamental problem in the interpretation suitable for the study of uveal and corneoscleral of the immunohistochemistry is that many of the trabecular endothelial morphology since the cells are compounds identified within the exfoliation sub­ so easily damaged by the surgical manipulations. stance are normal basement membrane constituents Another fundamental problem with the study of which could have diffused in from the aqueous. trabeculectomy specimens is that the sampling is THE OUTFLOW SYSTEM IN GLAUCOMA 19 limited in a disease which is patchy in the trabecular patients suffering from primary open angle glau- meshwork. coma. 144 The antenor. c h ambers were perfused at a pressure of 15 mmHg and the numbers and sizes of Abnormal Extracellular Matrix in the Cribriform the pores in the lining endothelium were measured. Layer Careful morphometric analysis failed to detect any significant difference between the normal and the The well-founded assumption that resistance to 144 aqueous outflow in the normal outflow system lies diseased tissue. in the cribriform layer (see above), was transferred One of the intriguing features of primary open to the outflow system in primary open angle angle glaucoma is that in spite of the high intraocular glaucoma. Before any valid observations could be pressure, Schlemm's canal is found to be open in the 140 made all the investigators in this field were aware of majority of well preserved specimens. If the the importance of documenting the changes which previously described studies on the effects of partial occur with ageing in the normal eye,8.l4,4'5-47,1 35-138 angle closure are taken into account, it would be and a large volume of information is now available, reasonable to assume that prolonged pressure would with different research groups placing emphasis on break down attachments in the cribriform layer and slightly different features of the cellular and extra­ that this would proceed to inner/outer wall apposi­ cellular constituents in the normal and disease states. tion in the canal. That this does not occur may be I am in no position to enter into the debate on the attributed to the rigidity of the extracellular matrix in pathogenesis of primary open angle glaucoma with the cribriform layer, but this assumption was not regard to the importance of the obstruction by confirmed by the distension of the cribriform layer abnormal extracellular matrix in the cribriform which occurred in elderly eyes subject to high­ layer: this has been extensively investigated by pressure fixation?4 others?6,45,46,115,137,1 39,140 It would therefore seem reasonable to propose The fundamental problem with which morpholo­ that narrowing of the collector channels would in gists have struggled is that of differentiating between some cases increase resistance to such a level that the changes seen as part of the normal ageing process distension of the cribriform layer would be prevented and those observed in primary open angle glaucoma. by a raised pressure in Schlemm's canal. According Interest has centred on 'plaque' materiali6 ,37,45-47,54 to Poiseuille's law, a reduction in channel diameter to and on the fine fibrillarmater ial located beneath the 80% of the original makes it necessary to double the lining endothelium of Schlemm's canal (this material pressure to maintain the same flow rate. In a light was recently shown to contain elastin)139, but the microscopic study of 100 trabeculectomy specimens, evidence obtained so far has been inconclusive in selected on the basis of good-quality morphology, I pinpointing a specific abnormality in primary open found that the canal was open in 94 % and that angle glaucoma. It is salutary to note that the most fibrous replacement of the trabeculae with enlarge­ comprehensive study of the constituents of the ment of the scleral spur and narrowing of the canal extracellular matrix has been unable to make a occurred in 50%. When collector channels were clear distinction between the diseased tissue and the identified 60% were found to be narrowed by control.140 deposition of fibrous tissue beneath the endothe­ lium: a feature which was emphasised by Ashton3 35 Decline in Cell Population some years ago. These non-specific and rather inconspicuous It is tempting to speculate that the constant demand abnormalities lead me to suggest that the holistic on the trabecular endothelial cells to behave as approach should be taken to the pathology of the phagocytes leads to a gradual depletion in the cell trabecular meshwork in primary open angle glau­ population - a morphological feature which is now coma - the relatively minor changes which occur , 141-143 we 11- d ocumente d . . A good case has been made throughout the outflow system in the disease state to show that cell depletion makes a significant are probably too subtle for detection by morpholo­ contribution in primary open angle glaucoma, the gical investigation. We must now rely on the hypothesis being that the disease is an exaggeration molecular biologists, who can investigate the bio­ the normal ageing process, but the evidence is not bf chemical characteristics and synthetic capabilities8 of totally convincing. Certainly the abnormality does the normal and abnormal trabecular endothelial cells 110t appear to lie in the capacity of the lining at a level undreamed of 35 years ago. In 1960 endothelium of Schlemm's canal to produce invagi­ Ashton3 ended his Doyne lecture with the following nation-vacuoles. A study was made of the lining words: endothelium of the canal by scanning electron microscopy to estimate the numbers of pores (i.e. Today there can be no brilliant conquests to lay in representations of flow channels) in normal eyes tribute to our founder for as I have tried to show obtained at autopsy and eyes similarly obtained from research in this field is still developing, and a� 20 W. R. LEE

orthodox concepts dissolve under the impact of mechanoreceptors in the human scleral spur. Invest new discoveries so further possibilities unfold. We Ophthalmol Vis Sci 1994;35:1157-66. are on shifting ground and this has been a progress 11. Garron L, Feeney ML, Hogan MJ, McEwan WK. Electron microscopic studies of the human eye. Am J report, but no doubt this topic will again be chosen OphthalmoI 1 958;46:27-56. by a Doyne Lecturer. To him across the years I 12. Holmberg AS. The fine structure of the inner wall of bequeath these words, probably to smile at - just Schlemm's canal. Arch OphthalmoI 1959;62:956-8. possibly to build upon - but we must leave him 13. Tripathi RC. Mechanisms of the aqueous outflow entirely to write the conclusions. across the trabecular wall of Schlemm's canal. Exp Eye Res 1971;11:116-21. I can only end by saying that I echo these words. 14. Tr ipathi RC, Tr ipathi BJ. Functional anatomy of the chamber angle. In: Jakobiec FA, ed. Ocular anatomy, It is a great pleasure to record the names of the research embryology and teratology. Philadelphia: Harper and workers and the technicians who have collaborated over Row, 1982:197-284. the past decades in the study of the normal and the 15. Johnstone MA, Grant WM. Pressure dependent diseased outflow system. These include Professor Ian changes in structures of the aqueous outflow system Grierson, Dr Paul G. McMenamin, Dr George E. of human and monkey eyes. Am J Ophthalmol Marshall, Dr Anastassos 0. Konstas and Dr John 1973;75:365-83. Ainsworth. Dr Ali Amanat of Great Yarmouth deserves 16. Grierson I, Johnson NF. The postmortem vacuoles GI special recognition for providing untraumatised trabecu­ Schlemm's canal. In: Lee WR, ed. Current research in lectomy specimens over the last decade. In addition I am ophthalmic electron microscopy. Berlin: Springer­ most grateful to Mrs Dorothy Aitken and Mr James A. W. Ve rlag, 1981:41-56. Ralston for their highly skilled technical assistance. Mrs 17. Lee WR, Grierson I. Relationships between intra­ Sophia Cameron has provided invaluable collaboration for ocular. pressure and the morphology of the outflow the recent studies of trabeculitis and closed angle apparatus. Tr ans Ophthalmol Soc UK 1974;94:430-49. glaucoma. Miss Sandra Howat is to be thanked for expert 18. Grierson I, Lee WR. Pressure-induced changes in the secretarial assistance. ultrastructure of the endothelium lining Schlemm's Without the environment and the stimulation provided canal. Am J OphthalmoI 1975;80:863-84. by Professor Wallace Foulds the research would not have 19. Grierson I, Lee WR. Light microscopic quantitation been possible. of the endothelial vacuoles in Schlemm's canal. Am J This work has been supported by the Medical Research Ophthalmol 1977;84:234-46. Council, the Wellcome Trust, the Scottish Home and 20. Lee WR, Grierson I. Pressure effects on the Health Department, the Guide Dogs for the Blind endothelium of the trabecular wall of Schlemm's Association and the Ross Foundation for the Prevention canal: a study by scanning electron microscopy. of Blindness. Graefes Arch Klin Exp OphthalmoI 1975;196:255-65. 21. Grierson I, Lee WR. The fine structure of the Keywords: Glaucoma: Morphology, Tr abecular meshwork. trabecular meshwork at graded levels of intraocular pressure. I. Pressure effects within the near physiolo­ gical range (8-30 mmHg). Exp Eye Res 1975;20: 505-21. REFERENCES 22. Grierson I, Lee WR. The fine structure of the trabecular meshwork at graded levels of intraocular 1. Catford GY. Our founder, Robert Doyne: a short pressure. II. Pressure outside the physiological range biographical tribute. Tr ans Ophthalmol Soc UK 1976; (0 and 55 mmHg). Exp Eye Res 1975;20:523-30. 96:285-9. 23. Grierson I, Lee WR, Abraham S. The pathways for 2. Doyne RW. Notes on the more common diseases in the drainage of aqueous humour into Schlemm's the eye. London: HK Lewis, 1896:1-41. canal. Tr ans Ophthalmol Soc UK 1977;7:719-25. 3. Ashton N. The exit pathways of the aqueous. Tr ans 24. Alvarado J, Murphy C, Juster R. Tr abecular mesh­ Ophthalmol Soc UK 1968;80:397-419. work cellularity in primary open angle glaucoma and 4. Cairns JE. Glaucoma. London: Harcourt Brace, 1986. nonglaucomatous normals. Ophthalmology 1984;91: 5. Rich R, Shields MM, Krupin T (eds). The glaucomas. 564-79. Vols 1 and 2. St Louis: CV Mosby, 1989. 25. Bill A. Some aspects of aqueous humour drainage. 6. Shields BM. Te xtbook of glaucoma. 3rd ed. Balti­ Eye 1993;7:14-9. more: Williams and Wilkins, 1992. 26. Rohen JW, Lutjen-Drecoll E. Morphology of aqueous 7. Bill A, Maepea 0. Mechanisms and routes of aqueous outflow pathways in normal and glaucomatous eyes. humour and drainage. In: Albert DM, Jakobiec FA, In: Rich R, Shields MB, Krupin T, eds. The eds. Principles and practice of ophthalmology. glaucomas. St Louis: Mosby, 1989:41-7. Philadelphia: WB Saunders, 1994:206-26. 8. Polansky JR, Alvarado JA. Cellular mechanisms 27. McMenamin PG, Lee WR. Effects of prolonged influencing the aqueous humour outflow facility. In: intracameral perfusion with mock aqueous humour Albert DM, Jakobiec FA, eds. Principles and practice on the morphology of the primate outflow apparatus. of ophthalmology. Philadelphia: WB Saunders, 1994: Exp Eye Res 1986;43:129-41. 226-5 1. 28. McMenamin Po. Functional and morphological 9. Coupland SE, Penfold P, Billson F, Hoffmann F. studies of the primate outflow apparatus. PhD Immunohistochemistry of the glaucomatous and Thesis, University of Glasgow, 1981. normal human trabecular meshwork. German J 29. Sawaguchi S, Yue B, Yeh P, Tso MOM. Effects of OphthalmoI 1994;3:168-74. intracameral injection of chondroitinase ABC in vivo. 10. Tamm ER, Flugel C, Stefani FH, Lutjen-Drecoll E. Arch OphthalmoI 1992;110:110-7. Nerve endings with structural characteristics of 30. Grierson I, Lee WR, Abraham S. Effects of THE OUTFLOW SYSTEM IN GLAUCOMA 21

pilocarpine on the human outflow apparatus. Br J 48. Shiose Y. Intraocular pressure: new perspectives. Surv OphthalmoI 1978;62:302-13. OphthalmoI 1990;34:413-5 . 31. Lee WR, Grierson I, Abraham S. Comparative 49. Marshall GE, Konstas AG, Lee WR. Immunogold anatomy of the angular tissues in the baboon and ultrastructural localisation of collagens in the aged human eye in relation to the pilocarpine response. human outflow system. Ophthalmology 1991;98: Tr ans Ophthalmol Soc UK 1977;97:726-31. 692-700. 32. Grierson I, Lee WR, Abraham S. The effects of 50. Marshall GE, Konstas AG, Lee WR. Immunogold topical pilocarpine on the morphology of the outflow localisation of type IV collagen and laminin in the apparatus of the baboon. Invest Ophthalmol aging outflow system. Exp Eye Res 1990;51 :691-9. 1979;18:346-55. 51. Marshall GE, Konstas AG, Lee WR. Collagens in 33. McMenamin PG, Lee WR, Grierson 1, Grindle J. ocular tissues. Br J OphthalmoI 1993;77:515-24. Giant vacuoles in the lining endothelium of 52. Montes GS, Bezerra MSF, Junquiera LCU. Collagen Schlemm's canal in the human eye after topical distribution in tissues. In: Ruggieri A, Motta PM, eds. timolol maleate. Invest Ophthalmol Vis Sci Ultrastructure of the connective tissue matrix. Boston: 1983;24:339-42. Martinus Nijhoff, 1984:65-88. 34. Ainsworth JR, Lee WR. Effects of age and high 53. Tr ipathi BJ, Hansen M, Li J, Tr ipathi RC. Identifica­ pressure fixatior on the lining endothelium of tion of type VI collagen in the trabecular meshwork Schlemm's caL",. Invest Ophthalmol Vis Sci and expression of its mRNA by trabecular cells. Exp 1990;31:745-50. Eye Res 1994;58:181-8. 35. Lee WR. The study of the passage of particles 54. Lutjen-Drecoll E, Rittig M, Rauterberg EA, Jander through the endothelium of the outflow apparatus of R, Mollenhauer J. Immunomicroscopical study of type the monkey eye by scanning and transmission VI collagen in the trabecular meshwork of normal and electron microscopy. Tr ans Ophthalmol Soc UK glaucomatous eyes. Exp Eye Res 1989;48:139-147. 1971;91: 687-705. 55. Grierson I, Lee WR. Acid mucopolysaccharides in the outflow apparatus. Exp Eye Res 1975;21:417-31. 36. Rohen JW, van der Zypen E. The phagocytic activity 56. Tawara A, Va rner HH, Holleyfield Jo. Distribution of the trabecular meshwork endothelium: an electron and characterisation of sulfated proteoglycans in the microscopic study of the vervet (Cercopitheus human trabecular tissue. Invest Ophthalmol Vis Sci aethiops). Graefes Arch Klin Exp Ophthalmol 1989;20:2215-31. 1968;175:143-60. 57. Chapman SA, Bonshek RE, Stoddart RW, Mackenzie 37. Rohen JW, Lutjen-Drecoll E. Biology of the trabe­ KR, McLeod D. Localisation of alpha (2,3) and alpha cular meshwork. In: Lutjen-Drecoll E, ed. Basic (2,6) linked terminal sialic acid groups in human aspects of glaucoma research. Stuttgart: Schattauer, trabecular meshwork. Br J Ophthalmol 1994; 1981:141-66. 78:632-7. 38. Grierson I, Day J, Unger WG, Ahmed A. Phagocy­ 58. Gong H, Freddo TF, Johnson M. Age-related changes tosis of latex microspheres by bovine meshwork cells of sulfated proteoglycans in the normal human in culture. Graefes Arch CIin Exp Ophthalmol 1986; trabecular meshwork. Exp Eye Res 1992;55:691-709. 224:536-44. 59. Fourman S. Diagnosing acute angle closure glaucoma: 39. Shirato S, Murphy CG, Bloom E� Franse-Carmen L, a flowchart. Surv OphthalmoI 1989;33:491-4. Maglio MT, Polansky JR, Alvarado JR. Kinetics of 60. Krupin T, Feitl ME. Glaucoma associated with uveitis. phagocytosis in trabecular meshwork cells: flow In: Rich R, Shields MM, Krupin T, eds. The cytometry and morphometry. Invest Ophthalmol Vis glaucomas. St Louis: CV Mosby, 1989: 1205-23. Sci 1989;30:2499-511. 61. Cantor LB, Schlaegel TF, Abrams JH, Weber Jc. 40. Park CH, Latina MA. Effects of gamma interferon on Intraocular pressure in anterior uveitis. Glaucoma human trabecular meshwork cell phagocytosis. Invest 1992;14:155-8. Ophthalmol Vis Sci 1993;34:2228-36. 62. Falcon MG, Williams HP. Herpes simplex kerato­ 41 . Epstein DL, Freddo TF, Anderson PJ, Patterson MM, uveitis and glaucoma. Tr ans Ophthalmol Soc UK Bassett-Chu S. Experimental obstruction to aqueous 1978;98:101-4. outflow by pigment particles in living monkeys. Inv 63. Hogan MJ, Kimura SJ, Thygeson P. Pathology of Ophthalmol Vis Sci 1986;27:387-95. herpes simplex keratoiritis. Am J Ophthalmol 42. Maeepa 0, Bill A. Pressures in the juxtacanalicular 1964;57:551-64. tissue and Schlemm's canal in monkeys. Exp Eye Res 64. Spencer WH. Pathological physiology of the second­ 1992;54:879-83. ary glaucomas. In: Cairns JE, ed. Glaucoma. Vo l. 1. 43 . Ethier C, Kamm R, Palazewski B. Calculations of flow London: Harcourt Brace, 1986:chap 20. resistance in the juxtacanalicular meshwork. Invest 65. Wilhelmus K, Grierson T, Watson po. Histopatholo­ Ophthalmol Vis Sci 1986;127:1741-50. gic and clinical associations of scleritis and glaucoma. 44. Seiler T, Wollansek J. The resistance of the trabecular Am J OphthalmoI 1981;91 :697-705. meshwork to aqueous humour outflow. Graefes Arch 66. Panek WC, Holland GN, Lee DA , Christensen RE. Clin Exp OphthalmoI 1985;223:88-91. Glaucoma in patients with uveitis. Br J Ophthalmol 45 . Grierson I. What is open angle glaucoma? Eye 1990;74:223-7. 1987;1:15-28. 67. Epstein DL, Hashimoto JM, Grant WM. Serum 46. Rohen JW, Futa R, Lutjen-Drecoll E. The fine obstruction of aqueous outflow in enucleated eyes. structure of the cribriform meshwork in normal and Am J OphthalmoI 1978;86:101-6. glaucomatous eyes as seen in tangential sections. 68. Kass MA, Podos SM, Moses RA, Becker B. Invest Ophthalmol Vis Sci 1981;21:574-85 . Prostaglandin E, and aqueous humour dynamics. 47 . Lutjen-Drecoll E, Shimuzu R, Rohrbach M, Rohen Invest Ophthalmol Vis Sci 1972;11:1022-7. Jw. Quantitative analysis of plaque material in the 69. Howes EL Jr, Cruse VK. The structural basis of inner and outer wall of Schlemm's canal in normal altered vascular permeability following intraocular and glaucomatous eyes. Exp Eye Res 1986;42:443-5 . inflammation. Arch OphthalmoI 1978;96:1668-76. 22 W. R. LEE

70. Roth M, Simmons R1. Glaucoma associated with induce a sialated glycoprotein (Cort-GP) in trabecular precipitates on the trabecular meshwork. Ophthal­ cells in vitro. Exp Eye Res 1990;51:735-7. mology 1979;86:1613-8. 90. Johnson DH, Bradley JMB, Acott TS. The effect of 71. Schlaegel TF. Complications of uveitis. Int Ophthal­ dexamethasone on glycosaminoglycans of human mol Clin 1977;17:65-74. trabecular meshwork in perfusion organ culture. 72. Foets B, van den Oord J, Engelmann K, Missotten Invest Ophthalmol Vis Sci 1990;31:2568-71. LA. Comparative immunohistochemical study of 91 . Rosenbaum JT, Samples JR, Seymour B, Langlois L, human corneotrabecular tissue. Graefes Arch Clin David L. Chemotactic activity of lens proteins and the Exp Ophthalmol 1992;230:269-74. pathogenesis of phacolytic glaucoma. Arch Ophthal­ 73. Lynch MG, Peeler JS, Brown RH, Niederkorn JY. mol 1987;105:1582-4. Expression of HLA class I and II antigens on cells of 92. Flocks M, Littwin CS, Zimmerman LE. Phacolytic the human trabecular meshwork. Ophthalmology glaucoma: a clinicopathologic study of one hundred 1987;94:851-7. and thirty-eight cases of glaucoma associated with 74. Tripathi BJ, Tr ipathi RC, Wong P, Raja S. Expression hypermature cataract. Arch Ophthalmol of HLA by the human trabecular meshwork and 1955;54:37-45 . corneal endothelium. Exp Eye Res 1992;51 :269-76. 93. Bron AJ, Habgood JO. Morgagnian cataract. Tr ans 75. Forrester JV, McMenamin PG, Holthouse I, Lumsden Ophthalmol Soc UK 1976;96:265-77. L, Liversidge 1. Localisation and characterisation of 94. Vrensen G, Kappelhof J, Willekens B. Morphology of major histocompatibility complex class II-positive cell the aging human lens. Lens Eye Toxicity Res in the posterior segment of the eye: implications for 1990;7:1-30. induction of auto-immune uveoretinitis. Invest 95 . Costello MJ, Oliver TN, Cobo LM. Cellular architec­ Ophthalmol Vis Sci 1994;35:64-77. ture in age related human nuclear cataracts. Invest 76. McMenamin PG, Crewe J, Morrison S, Holt PG. Ophthalmol Vis Sci 1992;33:3209-27. Immunomorphic studies of macrophages and MHC 96. Vrenson G, van Marie JV, van Been H, Willekens B. class II-positive dendritic cells in the iris and ciliary Membrane architecture as a function of lens fibre body of the rat, mouse and human eye. Invest maturation: a freeze fracture and scanning electron Ophthalmol Vis Sci 1994;35:3234-50. microscopic study in the human lens. Exp Eye Res 77. Missotten L. Immunology and herpetic keratitis. Eye 1992;54:433-46. 1994;8:12-21. 97. Brooks AMV, Grant G, Gillies WE. Comparison of specular microscopy and examination of aspirate in 78. O'Connor GR. Heterochromic iridocyclitis. Tr ans phacolytic glaucoma. Ophthalmology 1990;97;85-9. Ophthalmol Soc UK 1985;104:219-31. 98. Epstein DL, Jedziniak JA, Grant WM. Identification 79. Jones NP. Fuchs' heterochromic uveitis: an update. of heavy molecular weight soluble protein in aqueous Surv OphthalmoI 1993;37:253-72. humor in human phacolytic glaucoma. Invest 80. La Hay E, de Jong PTVM, Kij lstra A. Fuchs' Ophthalmol Vis Sci 1978;17:398-402. heterochromic. cyclitis: review of the literature on 99. Richter C, Epstein DL. Lens-induced open angle the pathogenetic mechanisms. Br J Ophthalmol glaucoma. In: Ritch R, Shields MB, Krupin T, eds. The 1994;78: 307-12. glaucomas. St Louis: CV Mosby, 1989:1017-26. 81. Goldberg MF, Erozan YS, Duke JR, Frost JK. 100. McMenamin PG, Lee WR. Ultrastructural pathology Cytopathologic and histopathologic aspects of Fuchs' of melanomalytic glaucoma. Br J Ophthalmol 1986; heterochromic iridocyclitis. Arch Opthalmol 1965;74: 70:895-906. 604-9. 101. El Baba F, Hagler WS, DeLa Cruz A, Green WR. 82. Perry HD, Yanoff M, Scheie HG. Rubeosis in Fuchs' Choroidal melanoma with pigment dispersion in heterochromic iridocyclitis. Arch Ophthalmol 1975; vitreous and melanomalytic glaucoma. Ophthalmol­ 93:337-9. ogy 1988;95:370-7. 83. Rohen HJW, Linner E, Witmer R. Electron micro­ 102. lnomata H, Bill A, Smelser GK. Aqueous humor scopic studies on the trabecular meshwork in two pathways through the trabecular meshwork and into cases of corticosteroid glaucoma. Exp Eye Res 1973; Schlemm's canal in the cynomolgus monkey (Macacus 17:19-31. irus): an electron microscopic study. Am J Opthalmol 84. Grierson I, Lee WR, McMenamin PG. The morpho­ 1972;73:760-89. logical basis of drug action on the outflow system of 103. Fenton RH, Zimmerman LE. Hemolytic glaucoma, an the eye. Res Clin Forums 1981;3:7-25. unusual cause of acute open angle secondary 85. Wilson K, McCartney MD, Miggans ST, Clark AF. glaucoma. Arch Ophthalmol 1963;70:236-9. Dexamethasone induced ultrastructural changes in 104. Grierson I, Lee WR. Erythrocyte phagocytosis in the cultured human trabecular meshwork cells. Curr Eye human trabecular meshwork. Br J Ophthalmol 1973; Res 1993;12:783-94. 57:400-15. 86. Steely HT, Browder SL, Julian MB, Miggans ST, 105. Grierson I, Lee WR, Abraham SA. Further observa­ Wilson KL, Clark AF. The effects of dexamethasone tions on the process of haemophagocytosis in the on fibronectin expression in cultured human trabe­ human outflow system. Graefes Arch Klin Exp cular meshwork cells. Invest Ophthalmol Vis Sci OphthalmoI 1978;208:49-64. 1992;33:42-50. 106. Campbell DG. Ghost cell glaucoma. In: Ritch R, 87. Partridge CA, Weistein BI, Southren AL, Gerritsen Shields MB, eds. The secondary glaucomas. St Louis: ME. Dexamethasone induces specific proteins in CV Mosby, 1982:320-7. human trabecular meshwork cells. Invest Ophthal­ 107. Shields CL, Shields JA , Shields MB, Augsburger JJ. mol Vis Sci 1989;30:1843-7. Prevalence and mechanisms of secondary intraocular 88. Yun AJ, Murphy CG, Polansky JR, Newsome DA, pressure elevation in eyes with intraocular tumours. Alvarado JA . Proteins secreted by human trabecular Ophthalmology 1987;94:839-46. cells. Invest Ophthalmol Vis Sci 1989;30:2021-2. 108. Bechrakis N, Lee WR. Dedifferentiation potential of 89. Tr ipathi BJ, Millard CB, Tr ipathi RC. Corticosteroids iris melanomas. Fortschtr Ophthalmol 1991;88:651-6. THE OUTFLOW SYSTEM IN GLAUCOMA 23

109. Farrar SM, Shields BM. Current concepts in pigmen­ clinico-histological correlation. German J Ophthalmol tary glaucoma. Surv OphthalmoI 1993;37:233-52. 1993;2:51-60. 110. Richardson TM. Pigmentary glaucoma. In: Ritch R, 128. Konstas AG, Marshall GE, Lee WR. Immunogold Shielcis MB, Krupin T, eds. The glaucomas. St Louis: localisation of laminin in the normal and exfoliation CV Mosby, 1989:981-95. iris. Br J OphthalmoI 1990;74:450-7. 111. Murphy CG, Johnson M, Alvarado JA . Juxtacanali­ 129. Vogiatzis A, Marshall GE, Konstas AGP, Lee WR. cular tissue in pigmentary and primary open angle Immunogold study of non-collagenous matrix compo­ glaucoma: the hydrodynamic role of pigment and nents in normal and exfoliative iris. Br J Ophthalmol other constituents. Arch Ophthalmol 1992;110:79-85. 1994;78: 850-8. 112. Fine BJ, Yanoff M, Scheie HG. Pigmentary glaucoma: 130. Uusitalo M, Kivela T, Tarkkanen A. Immunoreactiv­ a histologic study. Tr ans Am Acad Ophthalmol ity of exfoliation material for the cell adhesion related OtolaryngoI 1975;78:314-25 . HNK1-carbohydrate epitope. Arch Ophthalmol 113. Richardson TM, Hutchinson BT, Grant WM. The 1993;111:1419-23. outflow tract in pigmentary glaucoma: a light and 131. Ni C, Wang WJ, Albert DM, Schepens CL. Intravit­ electron microscopic study. Arch Ophthalmol reous silicone injection: histologic findings in a human 1977;195:1015-25. eye after 12 years. Arch Ophthalmol 1983;101: 114. Alvarado JA, Murphy CG. Outflow obstruction in 1399-401. pigmentary and primary open angle glaucoma. Arch 132. Nguyen QH, Lloyd MA, Heuer DK, Baerveldt G, OphthalmoI 1992;110:1769-78. Minckler DS, Lean JS, Liggett PE. Incidence and 115. Rodrigues MM, Spaeth GL, Sivalingam E, We inreb S. management of glaucoma after intravitreal silicone oil Histopathology of 150 trabeculectomy specimens in inj ection for complicated retinal detachments. glaucoma. Tr ans Ophthalmol Soc UK 1976;96:245-55. Ophthalmology 1992;99:1520-6. 116. Iwamoto T, Witmer R, Landolt E. Light and electron 133. Kirchof B, Tavakolian H, Paulmann H,. Heimann K. microscopy in absolute glaucoma with pigment Histopathological findings in eyes after silicone oil dispersion phenomena and contusion angle defor­ inj ection. Graefes Arch Clin Exp Ophthalmol mity. Am J OphthalmoI 1971;72:420-34. 1986;224:34-7. 117. Konstas AG, Jay JL, Marshall GE, Lee WR. 134. Suzuki M, Ukada I, Takeuchi S, Ishii Y, Yamashita H, Prevalence, diagnostic features and response to Hori S. Effect of silicone oil on ocular tissues. Jpn J trabeculectomy in exfoliation glaucoma. Ophthalmol­ Ophthalmol 1991;35:282-91. ogy 1993;100:619-27. 135. McMenamin PG, Lee WR. Age-related changes in 118. Streeten BAW, Dark Al Pseudoexfoliation syn­ extracellular materials in the inner wall of Schlemm's drome. In: Garner A, Klintworth GK, eds. Pathobiol­ canal. Graefes Arch Klin Exp Ophthalmol 1980; ogy of ocular disease. New York: Marcel Dekker, 212:1 59-72. 1994:591-630. 136. McMenamin PG, Lee WR, Aitken DN. Age-related 119. Brinchmann-Hansen 0, Albreksten T, Anmarkrud N. changes in the human outflow apparatus. Ophthal­ Pilocarpine drops do not reduce intraocular pressure mology 1986;93:194-209. sufficiently in pseudoexfoliation glaucoma. Eye 137. Lutjen-Drecoll E, Futa R, Rohen Jw. Ultrahisto­ 1993;7:51 1-6. chemical studies on tangential sections of the 120. Ringvold A. Electron microscopy of the trabecular trabecular meshwork in normal and glaucomatous meshwork in eyes with exfoliation syndrome (pseu­ eyes. Invest Ophthalmol Vis Sci 1981;21:563-73. doexfoliation of the lens capsule). Virch Arch (Pathol Anat) 1971;353:110-27. 138. Oates DC, Belcher CD. Aging changes in trabecular 121. Konstas AG, Marshall GE, Cameron SA, Lee WR. meshwork, iris and ciliary body. In: Albert DM, Morphology of iris vasculopathy in exfoliation Jakobiec FA, eds. Principles and practice of ophthal­ glaucoma. Acta Ophthalmol (Copenh) 1993;71:751-9. mology: basic sciences. Philadelphia: WB Saunders, 122. Helbig H, Schloetzer-Schrehardt U, Noske W, Kellner 1994:697-701. U, Foerster MH. Anterior chamber hypoxia and iris 139. Umihara J, Nagata S, Nohara M, Hanai T, Usuda N, vasculopathy in pseudoexfoliation syndrome. German Segawa K. Localisation of elastin in the normal and J OphthalmoI 1994;3:148-53. glaucomatous human trabecular meshwork. Invest 123. Li Z-Y, Streeten BW, Wallace RN. Association of Ophthalmol Vis Sci 1994;35:486-94. elastin with pseudoexfoliative material: an immuno­ 140. Rohen JW, Lutjen-Drecoll E, Flugel C, Meyer M, electron microscopic study. Curr Eye Res 1988; Grierson I. Ultrastructure of the trabecular meshwork 7:1163-72. in untreated cases of primary open angle glaucoma. 124. Streeten BW, Li ZY, Wallace RN, Eagle RC, Exp Eye Res 1993;56:683-92. Keshgegian AA. Pseudoexfoliative fibrillopathy in 141. Alvarado J, Murphy C, Polansky I Age-related visceral organs of a patient with pseudoexfoliation changes in trabecular meshwork cellularity. Invest syndrome. Arch OphthalmoI 1992;110:1757-62. Ophthalmol Vis Sci 1981;21:714-27. US.Schloetzer -Schrehardt U, Doerfler S, Naumann GOR. 142. Alvarado J, Murphy C, Juster R. Tr abecular mesh­ Immunohistochemical localisation of basement mem­ work cellularity in primary open angle glaucoma and brane components in pseudoexfoliation material of non-glaucomatous controls. Ophthalmology 1984;91: the lens capsule. Curr Eye Res 1992;11:343-55. 564-79. 126. Schloetzer-Schrehardt UM, Koca MR, Naumann 143. Grierson I, Howes R. Age related depletion of the cell GOH, Volkholz H. Pseudoexfoliation syndrome: population in the human trabecular meshwork. Eye ocular manifestations of a systemic disorder. Arch 1987;1:204-10. OphthalmoI 1992;110:1752-6. 144. Allingham RR, De Kater AW, Ethier CR, Anderson 121. Schloetzer-Schrehardt U, Kuchle M, Dorfler S, PJ, Hertzmark E, Epstein DL. The relationship Naumann GOH. Pseudoexfoliative material in the between pore density and outflow facility in human eyelid skin of pseudoexfoliation-suspect patients: a eyes. Invest Ophthalmol Vis Sci 1992;33:1661-9.