BritishJournal ofOphthalmology, 1991,75,552-557

Alterations in elastin of the optic nerve head inhuman and experimental glaucoma

Harry A Quigley, Andrew Brown, Mary Ellen Dorman-Pease

AbstractThe optic nerve heads from normal andglaucomatous eyes of humans and monkeyswere examined by light and electron micro-scopy for the presence and distribution ofelastin. Elastin densely lined the insertion ofthe lamina cribrosa into the sclera and wasprominent in the laminar beams. The long axisof elastin paralleled that of the collagen fibrilsand corresponded to the directions ofexpectedforces on the tissue. In glaucomatous eyeselastin molecules were curled instead ofstraight and seemed disconnected from theother elements ofthe connective tissue matrix.Laminar beams stretch and reorganise theirsubstructure during glaucomatous atrophy,probably leading to changed compliance.Differences in elastin function may have a partin susceptibility to glaucomatous injury.

Glaucoma Service,Wilmer Institute, JohnsHopkins UniversitySchool of Medicine,Baltimore, Maryland,USAH A QuigleyA BrownM E Dorman-PeaseCorrespondence to:Dr H A Quigley, MaumeneeBR 10, Wilmer, Johns HopkinsHospital, Baltimore,Maryland 21205, USA.Accepted for publication15 March 1991This paper was presented inpart at the 1990 meeting of theAssociation for Research inVision and Ophthalmology,Sarasota, Florida, USA

Glaucoma is best defined as an optic neuropathyassociated with a characteristic excavation of theoptic disc' and a progressive loss of visual fieldsensitivity. This neuropathy is related to thelevel of the intraocular pressure (IOP) such that,the higher the IOP the more likely the neuro-pathy is to occur.2 In some persons it develops atwhat is considered normal IOP, while in manyothers the IOP is intermittently or consistentlyabove the statistically normal level.The IOP acts on the optic nerve head and the

sclera by two force vectors. Firstly, the IOP isusually higher than optic nerve tissue pressure,generating an inside-out force. Secondly, theIOP generates a wall tension that pulls on theperimeter of the nerve head. The tissues at thiscomplex opening in the eye wall are specialised toaccount for these forces.?- This specialised com-position allows the safe passage of nerve fibresout of the eye and the maintenance of an

appropriate blood supply to them under normalconditions.

We have conducted detailed studies of theextracellular connective tissues ofthe optic nervehead. Collagen, glycosaminoglycans, and elastinare instrumental in determining the response ofthe eye to the forces produced by IOP.Anderson6 may have been the first to describe thepresence of elastin within the lamina cribrosa byelectron microscopy. We here demonstrate thedetailed distribution of elastin in the normaloptic nerve head by electron microscopy and by a

histochemical method not previously used inocular tissues. These techniques illustrate howelastin may contribute to the biomechanicalresponse of the nerve head. Features of extra-cellular structure may have an important role indetermining the susceptibility of the individualeye to pressure-induced forces. In addition we

present the eyes of human eye bank donors whohad had glaucoma and the eyes of monkeys withexperimental glaucoma. The appearance of elas-tin fibres and the surrounding connective tissueelements are dramatically changed by the IOP.

Material and methodsWe studied 10 human eyes that had no history ofeye disease and 19 human glaucomatous eyeswith a history of raised IOP and/or clinicalglaucomatous damage. Only one eye per personwas examined. Control eyes had a normal neuralarea m their optic nerve cross-section, which wasembedded in epoxy resin, thick-sectioned at1 lim, and analysed.78 Normal and glaucoma-tous donors did not differ in age (mean=74 yearsin each group, p>0 05); 90% of both donorgroups were Caucasian, and the two groups wereequally divided between men and women.Human glaucomatous eyes were fixed in buf-

fered aldehydes within 24 hours of death. Thedegree of glaucomatous damage was determinedby the visual field test result, the optic nerveneural area and/or fibre count, or both (Table 1).Nineteen glaucomatous eyes were studied bylight microscopy, at least three from each of thefive severity stages. In addition we examined byelectron microscopy six normal and sevenglaucomatous human eyes, two each from stages1, 4, and 5 and one from stage 3.

In cynomolgus monkeys (Macaca fascicularis)we caused a raised IOP in one eye by lasertrabecular treatment.9 This produces an opticneuropathy that is indistinguishable fromhuman glaucoma in clinical and histopatho-logical details. The animals were anaesthetisedwith intravenous pentobarbitone, killed byexsanguination, the eyes fixed by vascular perfu-sion of aldehyde fixatives, and the optic nerveheads of their damaged and normal eyes dis-sected as described below. We studied the nerveheads of five glaucomatous monkey eyes andtheir fellow eye controls. The degree of injury inmonkey eyes was judged by the degree of fibreloss in the optic nerve cross sections in similarfashion to human eyes. Two of the monkey eyeshad stage 3 damage and three had stage 5

Table I Criteriafor severity staging ofhuman glaucoma eyes

Visualfield Optic nerve area

Stage 1* Normal >85% of normalStage 2 Normal 75-85% of normalStage 3 Usually abnormal 50-75% of normalStage 4 Abnormal 25-50% of normalStage 5 Advanced loss <25% ofnormal

In seven of the 19 eyes only one of the two parameters wasavailable.*Stage 1 consists of eyes suspected of being glaucomatous that hadstatistically abnormal intraocular pressure.

552

on February 23, 2021 by guest. P

rotected by copyright.http://bjo.bm

j.com/

Br J O

phthalmol: first published as 10.1136/bjo.75.9.552 on 1 S

eptember 1991. D

ownloaded from

Alterations in elastin ofthe optic nerve head in human and experimental glaucoma

damage. This research complied with the ARVOResolution on Animals and was approved andmonitored by the Johns Hopkins Animal CareCommittee.

For light microscopy we removed the opticnerve head and trimmed the surrounding sclerato identify the superior position. The tissue wasrinsed in phosphate buffer, and the retina wasremoved. Nerve heads were placed in 30%sucrose in distilled water for cryoprotection,frozen in liquid nitrogen, and serially sectionedat 10 [tm thickness parallel to the interface of thenerve head with the vitreous cavity. This methodgave similar tissue preservation to paraffinembedding. Frozen, unfixed sections werestained by Luna's method, whose major stepsare: (1) 30 minutes in aldehyde fuchsin; (2)alcohol rinse; and (3) 1 minute in haematoxylinto stain the alpha elastin fibres.'0 By stainingalternate sections we found the Luna techniqueto be superior to Verhoeff-van Gieson and otherhistochemical methods for highlighting thethree-dimensional elastin network. We con-firmed by electron microscopic methods descri-bed below that the position and relative numberof stained elements with the Luna methodcorrespond to alpha elastin. In addition ourprevious light microscopic studies of staining ofalpha elastin by antibodies labelled by theimmunoperoxidase method indicate that thedistribution of stained material is similar to that

obtained with the Luna method. The nerveheads were analysed by region in the normal eyes(superior, inferior, nasal, temporal). In theglaucomatous eyes the degree of nerve fibre losswithin an optic nerve cross section7'8 was corre-lated with the appearance of elastin in eachregion. The findings in human and monkey eyeswere similar, but only human examples arepresented here.For electron microscopy nerve heads were

postfixed in 1% osmium tetroxide and stainedwith tannic acid, uranyl acetate, and lead citrateafter thin sectioning, to define better the alphaelastin." We have demonstrated previously thatthe elements referred to as elastin fibres'2 stainwith antibodies to alpha elastin.'3

ResultsOur normal material provided an importantcomparison with the glaucomatous tissues pre-pared by the same methods. There was a largeamount of elastin in the optic nerve head region.The specificity of the Luna stain was confirmedby staining of Bruch's membrane and arteriolarwalls. The peripapillary sclera had more elastinthan the sclera distant from the nerve. Byelectron microscopy the long axes ofelastin fibreswere parallel to the orientation of the collagenbundles in which they resided (Fig 1). Elastinstained darkly with tannic acid, and the specifi-city of this staining had been previously con-firmed by labelling with antibodies against alphaelastin conjugated with gold particles (Fig 2).

In the immediate 200-300 iim zone of inser-tion of the lamina cribrosa, elastin fibrils ringedthe nerve head in a dense band (Fig 3). From thisring elastin fibrils originated at nearly a rightangle, entering the base of each beam of thelamina cribrosa. Each laminar beam had severalelastin fibres, with collagen spaced betweenthem (Fig 4). The elastin fibres in the normalhuman or monkey lamina were invariablystraight in appearance, bending only slightly asthey passed from one beam to another. Therewere only minimal differences between regionsofthe nerve head cross section in elastin distribu-tion. However, anteroposteriorly elastin densitywas highest at the level of the sclera and lower atthe level of the retina and posterior to the line ofmyelination (Fig 5).

In the stage 1 glaucomatous eyes no definitedifferences from the normal pattern of elastinstaining were observed with light or electronmicroscopy. In all the stage 2 nerve heads theappearance of elastin was abnormal. In somebeams the normally straight elastin moleculesappeared curvilinear, like small sine waves (Fig4). This was most prominent in atrophic zones ofthe nerve. At damage stages 3 and 4 every eyehad curved elastin profiles in many beams. Ineyes with severe damage (stage 5) the interpreta-tion of the findings was made difficult by theextreme change in laminar architecture.We correlated the sectors of substantial neural

loss with the presence of curled elastin. Six eyeswere not included in this analysis because theirdamage was uniform. Moreover, we excludedthe four stage 5 eyes that had nearly total loss offibres. In five of the remaining nine eyes the most

Figure I Transmissionelectron micrographs oflamina cribrosa beamstructure. A: Collagenfibrils(C) are seen in cross sectionand elastin complexes(arrows) are parallel tocollagen, therefore alsoappearing in cross section. B:Collagen (C) and elastin(arrows) are both inlongitudinal section.(Normal monkey nervehead, x 17 000.)

553

on February 23, 2021 by guest. P

rotected by copyright.http://bjo.bm

j.com/

Br J O

phthalmol: first published as 10.1136/bjo.75.9.552 on 1 S

eptember 1991. D

ownloaded from

Quigley, Brown, Dorman-Pease

Figure 2 Collagen andelastin from normal monkeylamina cribrosa (upper left)shows the normally closeapposition ofcollagenfibrilsto the microfibrillarcomponent ofthe elastincomplex (arrows). The upperright micrograph from anormal human lamina showsthat the darkly stained core(white E in upper left) islabelled by antibodiesagainst alpha elastin that Vwere conjugated with 15 nm .Scolloidal gold particles.4 Thetwo lower micrographs arefrom lamina ofmonkeys withexperimental glaucoma. riNote the muchgreater)separation between elastin'smicrofibrils (arrows), aL _ Acompared with the normalappearance upper left. (Theseparation present in thegold-labelled specimen,upper right, is due topreparation conditionsnecessaryfor immunoelectron 'Wmicroscopy.) (Upper left,lower right and left: monkey,x48000; upper right,normal human, x37 000).

prominent change in elastin appearance was inthe area(s) of nerve loss.The number of elastin fibres in the glaucoma-

tous eyes seemed qualitatively normal. It provedimpossible to quantify elastin fibres by lightmicroscopy. We are at present conducting quan-tifications by other techniques.By electron microscopy the normal laminar

Figure 3 Zone ofscleraimmediately adjoining thenerve head (insertion zone).Note dense ring ofelastinsurrounding nerve headopening (arrow) and elastinofmost peripheral laminacribrosa beam perpendicularto the ring passingfromscleral rim into nerve head -A.Vbetween neural pores (1).(Normal human eye,Luna/me ,frozen section, W

Fiuex540fcer t. .)~-

,j~ ~ ;ad

beams were surrounded by astrocytes and theirbasement membrane. The beam core consistedof fibrillar collagen (types I, III, and VI), elastin,glycosaminoglycans, and the laminar capillarieswith their associated basement membranes (Fig1). Even in the mildly damaged glaucomatouseyes the lamina cribrosa beam structure wasdisrupted, leading to a lower density of those

554

on February 23, 2021 by guest. P

rotected by copyright.http://bjo.bm

j.com/

Br J O

phthalmol: first published as 10.1136/bjo.75.9.552 on 1 S

eptember 1991. D

ownloaded from

555Alterations in elastin ofthe optic nerve head in human and experimental glaucoma

Figure 4 Light micrographs showing lamina cribrosa beams stainedfor elastin in nerve headsclassified as normal (upper), mildly damaged (middle), and severely damaged (lower). Thenormal elastin appearance (upper) consists ofstraight elastinfibres spaced evenly across thebeams. In the mildly and severely damaged glaucomatous eyes the elastinfibres arecurved (arrows) rather than straight. (Human normal and glaucoma stage 2 and stage 4; Lunamethod, frozen sections, x490.)

Figure S Lightmicrograph ofelastin atscleral insertion ofnervehead (S). Elastin in laminarbeams (L) runs straightacross the nerve head opening(left to right), while that inthe meninges (M) is parallelto the long axis ofthe opticnerve. The septa ofthe opticnerve behind the line ofmyelination contain elastinwith various directionalorientations (arrow).(Normal human eye; Lunamethod with epoxy resinembedding, x 350.)

components visible by electron microscopy (Fig6). The better preservation of the monkey eyesprovided a clearer view of this change, but it waspresent in the human material as well. Thebeams underwent a movement toward themyelinated nerve from their original position.This was evident in our specimens as a loss intheir astrocytic covering, and the development ofelectron lucent spaces between collagen andelastin fibrils (Fig 7). In the normal laminarbeam collagen and elastin are tightly packedtogether. In the glaucomatous nerve heads,however, elastin fibres were less closely apposedto fibrillar or basement membrane collagen thanin normal tissue (Figs 2, 6, 7). The ultrastructuralsubstructure of the elastin fibres was not dif-ferent in normal and glaucomatous eyes.

DiscussionHernandez and coworkers'4 found that the areaof the sclera immediately adjacent to the laminais rich in elastin and that there is a circumferen-tial orientation to the molecules. They also foundelastin within laminar beams, orientated pre-ferentially across the nerve head opening. Weconfirm that the densest accumulation of elastinat the normal nerve head is at the insertion ringimmediately surrounding the lamina, at the baseofthe most peripheral beams of the lamina. Herethe long axis of elastin fibres is orientatedcircumferentially. Biomechanical analysis of ahole in a walled structure indicates that the stressat the edge ofthe hole might be several times thatof the wall tension itself.'5 Both elastin andcollagen are concentrated in amount and positionat this location in a way that appears to resistforces that would tend to enlarge the nerve headopening.The long axis of elastin fibres is typically

parallel to that of the collagen fibrils in thelamina. Throughout the nerve head both fibrillarspecies run across the lamina from one edge ofthe nerve head to the other. Elastin fibres areusually straight in appearance in the normallamina. It is logical that the maximum resistanceto the deformation of laminar beams is in theplane across the nerve head, compared withforces exerted from inside outward. This mayexplain why the most prominent change in thestructure of nerve head in glaucoma is notenlargement of its diameter at the level ofBruch's membrane,'6 but an anteroposteriorcompression of the laminar plates.7

Laboratory investigations show that theprimary function of elastin is to resist tissuedeformation and to return its shape to normal atlow levels of stress.'7 Elastin consists of com-plexes of a 72 000 molecular weight poly-peptide,'8 deposited on a microfibrillar matrix(fibrillin).'3 Elastin stretches under stress andreturns to its original length after release oftension. Collagen appears to function in loadbearing with moderate and severe stress.'9 It iscapable of resistance to elongation but does notrestore the original position or length of thetissue as effectively as elastin does.

In both human and monkey glaucoma wefound that the elastin within the lamina cribrosaappeared wavy in outline rather than straight, as

on February 23, 2021 by guest. P

rotected by copyright.http://bjo.bm

j.com/

Br J O

phthalmol: first published as 10.1136/bjo.75.9.552 on 1 S

eptember 1991. D

ownloaded from

Quigley, Brown, Dorman-Pease

Figure 6 Lamina cribrosabeam in a monkey eye withexperimental glaucomadamage. Note that bycomparison with similar areain a normal monkey eye(Figure IA) there is adramatic disorganisation ofthe beam. The normal tightlypacked appearance ischanged to one withsubstantial optically emptyzones among collagenfibrilsand elastin molecules(arrow). (Monkey;x12000.)

Figure 7 Disorganisedlamina cribrosa beamstructure in experimentalglaucoma monkey eye.Elastin (arrow) isfounddissociatedfrom collagenfibrils in beam with muchlower than normal density offibrillar elements. Basementmembrane material(including collagen type IV)isfound in several layers(asterisk) instead ofthenormal single layer, and isseparatedfrom its cell oforigin (either capillaryendothelium or astrocyte).(Monkey; x 16 000.)

'.~ ~~~~~1rJ

in normal tissue. This is not likely to be anartefact of tissue preservation or handling. Bothnormal and glaucomatous eyes were fixed andsectioned identically, yet only glaucomatous eyeshad wavy elastin. It was present in frozen-sectioned and plastic-embedded tissues, andwith a variety of elastin stains (though mostvividly with Luna staining). The wavy appear-ance was found in both human and experimentalmonkey glaucomatous eyes. Finally, it was morelikely to be found within the same nerve head inareas of damage compared with areas that wereless atrophic.These facts leave little doubt that the wavy

appearance is a characteristic change in thesubstructure of the glaucomatous lamina crib-rosa. We do not know whether it is a causal factorin neural damage or a simple result of therearrangement of nerve head tissues typical forglaucoma. The wavy elastin was detected as earlyas our stage 2, a group of eyes with either normalvisual field tests or more than 75% of the normalnumber of optic nerve fibres. This places thefinding relatively early in glaucoma damage.

Electron microscopically, the relationship ofelastin fibres to visible collagen was also changed-in glaucomatous eyes. In normal laminar beams;collagen fibrils or basement membranes are

immediately next to the microfibrillar network ofthe elastin complex. After glaucoma injuryelastin appeared to be separate from other ele-ments in disorganised laminar beams. Sincethere were many beams with collagen fibrilsorientated randomly rather than in typicallyorganised fashion, it is reasonable to supposethat irreversible elongation of laminar beamsoccurred.202' Normally elastin may be directlyconnected to collagens type I, III, IV, or VI. "

If the collagen bundles lose their normal struc-ture, the elastin network might be disconnectedfrom the other elements of the connective tissue.This may explain the change in elastin appear-ance seen by light microscopy. One possibility isthat the elastin is analogous to a spring that isnormally stretched in a straight line between twofixed points. When disconnected from a fixedframework, the spring assumes a curled shape.Alternatively, the wavy appearance of the elastinmight have resulted from other causes. Forexample, the internal structure of the fibresmight have been altered. However, we could notdetect ultrastructural differences between theelastin fibres in glaucomatous and normal eyes.The curled appearance of collagen within the

disorganised laminar beams is only one of anumber of changes in laminar connective tissuein experimental or human glaucoma. There isstrong evidence that changes occur in laminarcollagen. Biochemical analyses suggest a differ-ence in the amino acid composition of collagen innormal as compared with glaucomatous eyes.22Changes in the amount or distribution of col-lagen were also suggested by immunohisto-chemical labelling of collagens.2021 In a previousimmunoelectron microscopic study2' we demon-strated reduplicated basement membrane, con-sisting in part of collagen type IV, within theformer neural bundle areas of experimentalglaucoma nerve heads. This is in a differentposition and has a different appearance from thecurled elastin within affected laminar beamsreported here.The alterations in elastin and collagen in

glaucoma may change the biomechanicalbehaviour of the lamina. In this regard ourobservations appear to be consistent withmeasurements of optic nerve head compliancemade by Zeimer and Ogura that indicate anincreased stiffness with advanced glaucomatousdamage.23

This work was supported in part by PHS Research Grants EY02120 and EY 01765 (Dr Quigley and the Wilmer Institute), byfunds from National Glaucoma Research, a programme of theAmerican Health Assistance Foundation, Rockville, Maryland,and by a Senior Investigator Award from Research to PreventBlindness, Inc, New York.

1 Quigley HA, Addicks EM, Green WR, Maumenee AE. Opticnerve damage in human glaucoma. II. The site of injury andsusceptibility to damage. Arch Ophthalmol 1981; 99: 635-49.

2 Sommer A. Intraocular pressure and glaucoma. AmJr Ophthal-mol 1989; 107: 186-8.

3 Hernandez MR, Igoe F, Neufeld AH. Extracellular matrix ofthe human optic nerve head. Am J Ophthalmol 1986; 102:139-47.

4 Morrison JC, L'Hernault NL, Jerdan JA, Quigley HA.Extracellular matrix composition of the monkey optic nervehead. Invest Ophthalmol Vis Sci 1988; 29: 1141-50.

5 Hernandez MR, Igoe F, Neufeld AH. Cell culture of thehuman lamina cribrosa. Invest Ophthalmol Vis Sci 1988; 29:78-89.

6 Anderson DR. Ultrastructure of human and monkey laminacribrosa and optic nerve head. Arch Ophthalmol 1969; 82:800-14.

556

on February 23, 2021 by guest. P

rotected by copyright.http://bjo.bm

j.com/

Br J O

phthalmol: first published as 10.1136/bjo.75.9.552 on 1 S

eptember 1991. D

ownloaded from

Alterations in elastin ofthe optic nerve head in human and experimental glaucoma

7 Quigley HA, Hohman RM, Addicks EM, Massof RS, GreenWR. Morphologic changes in the lamina cribrosa correlatedwith neural loss in open-angle glaucoma. Am Ophthalmol1983; 95: 673-91.

8 Repka MX, Quigley HA. The effect of age on normal humanoptic nerve fiber number and diameter. Ophthalmology 1989;96:26-31.

9 Quigley HA, Hohman RM. Laser energy levels for trabecularmeshwork damage in the primate eye. Invest Ophthalmol VisSci 1983; 24: 1305-6.

10 Kligman LH. Luna's technique. A beautiful stain for elastin.AmJDermatopathol 1981; 3: 199-205.

11 Kajikawa K, Yamaguchi T, Katsuda S, Miwa A. An improvedelectron stain for elastic fibers using tannic acid. ElectronMicrosc (Tokyo) 1975; 24: 287-90.

12 Greenlee TK, Ross R, Hartman JC. The fine structure ofelastic fibers.J Cell Biol 1966; 30: 59-71.

13 Sakai LY, Keene DR, Engvall E. Fibrillin, a new 350-kDglycoprotein, is a component of extracellular microfibrils.Cell Biol 1986; 103: 2499-509.

14 Hernandez MR, Luo XX, Igoe F, Neufeld AH. Extracellularmatrix of the human lamina cribrosa. Am Ophthalmol1987; 104: 567-76.

15 Greene PR. Mechanical considerations in myopia: relativeeffects of accommodation, convergence, intraocularpressure, and the extraocular muscles. Am Optom PhysiolOpt 1980; 57: 902-14.

16 Quigley HA, Brown AE, Morrison JC, Drance SM. The sizeand shape of the optic disc in normal human eyes. ArchOphthalmol 1990; 108: 51-7.

17 Oxlund H,. Manschot J, Viidik A. The role of elastin inthe mechanical properties of skin. J7 Biomech 1988; 21:213-8.

18 Rosenbloom J. Elastin: an overview. Methods Enzymol 1987;144:172-96.

19 Oxlund H, Andreassen TT. The roles of hyaluronic acid,collagen, and elastin in the mechanical properties of connec-tive tissues.JAnat 1980; 131: 611-20.

20 Morrison JC, Dorman-Pease ML, Dunkelberger GR, QuigleyHA. Optic nerve head extracellular matrix in primary opticatrophy and experimental glaucoma. Arch Ophthalmol 1990;108: 1020-4.

21 Hernandez MR, Andrejewska WM, Neufeld AH. Changes inthe extracellular matrix of the human optic nerve head inprimary open-angle glaucoma. Am Ophthalmol 1990; 109:180-8.

22 Tengroth B, Ammitzboll T. Changes in the contentand composition of collagen in the glaucomatous eye -

basis for a new hypothesis for the genesis of chronic openangle glaucoma. Acta Ophthalmol (Kbh) 1984; 62: 999-1008.

23 Zeimer RC, Ogura Y. The relation between glaucomatousdamage and optic nerve head mechanical compliance. ArchOphthalmol 1989; 107: 1232-4.

557

on February 23, 2021 by guest. P

rotected by copyright.http://bjo.bm

j.com/

Br J O

phthalmol: first published as 10.1136/bjo.75.9.552 on 1 S

eptember 1991. D

ownloaded from