Eye (1990) 4, 389-424

The Corneal Endothelium S. J. TUFT and D. J. COSTER London and Adelaide

Summary The endothelium is a monolayer of cells on the posterior corneal surface that trans­ ports water from the stroma into the anterior chamber. This movement of water counters a natural tendency for the stroma to swell and is necessary to maintain a transparent cornea. Embryologic studies, in particular the demonstration of the derivation of the endothelium from the neural crest, have provided insight into the factors that govern the response of this tissue to disease. In some species the endo­ thelium can regenerate after injury, but in man cellular enlargement is the main mechanism of repair after cell loss. A clinical estimate of endothelial cell density and function is provided by specular microscopy, fluorophotometry and pachymetry. In this paper we review the development, structure and function of the corneal endo­ thelium, and then consider the pathological processes that can affect this tissue.

There

have

been

considerable

advances

towards an understanding of the physiology of the endothelium since its role in maintaining

The Normal Endothelium

1. Embryology

The eye and periocular structures are formed

corneal clarity was first recognised. Clinical

from cells that are derived from four distinct

interest has been spurred by the refinement of

embryologic

complex

mesoderm, neural tube and the neural crest

intraocular

surgical

techniques,

appreciation of the susceptability of the endo­ thelium to surgical trauma and the develop­ ment of in

( Fig. 1).1 Many of the structural roles of meso­

vivo methods of evaluation that can which

follow

seemingly

minor

injuries. In addition, a reassessment of the embryogenesis of the corneal endothelium

Ectoderm (Epiblast) Neurecloderm Neural Crest Non-Ocular

A

has influenced our interpretation of endothe­

structure and function of the corneal endo­

/\

Ocular

I

lial pathology. We are at a point where the current understanding of the development,

ectoderm,

derm in the rest of the body are assumed by

demonstrate the dramatic changes in mor­ phology

tissues-surface

/ ....... Epidermal Ectoderm

Mesoderm

Neural Tube

I

Corneal Epithelium

E�lr.-ocular My.cle.

V ••cular Endothelium

Denial Papillae M""anophoru

thelium is providing an appreciation of the surprisingly

diverse

pathology

which

can

affect this apparently simply monolayer of cells.

Fig.1. A diagramatic representation af the derivation of ocular tissues from embryonic precursors. The corneal endothelium is a product of the neural crest.

From Department of Clinical Ophthalmology, Moorfields Eye Hospital, City Road, London, and Department of Ophthalmology, Flinders Medical Centre, Bedford Park, South Australia. Correspondence to: Professor D. J. Coster, Department of Ophthalmology, Flinders Medical Centre , Bedford Park, South Australia 5042.

390

S. J. TUFT A N D D. J. COSTER

the neural crest in the head and neck, and the neural crest has a crucial role in the develop­ ment of the eye. 2.3.4.5 The fundamental studies of ocular embryo­ logy have been performed in birds, with movement of the neural crest being mapped by

transplanting

labelled

cells

between

embryos and following their development in the resulting chimera. 6 Radioactively labelled

cells can be identified for a few cell divisions

during early oculogenesis,7 but, by using quail

cells that contain an intrinsic nuclear marker which can be identified after transplantation

into chick embryos, it has been possible to fol­ low subsequent stages. 8 This type of experi­ ment

has

demonstrated

the

remarkable

mobility of neural crest cells in the embryo and has shown that it is a popUlation of these cells

that

ultimately

forms

the

corneal

endothelium.

Neural crest cells can first be identified as

they separate from the neural plate, a thick­

ening of the dorsal surface ectoderm. P rior to the closure of the neural folds at the anterior neuropore, the neural crest cells begin to move peripherally as primary mesenchyme (Fig.

2).

In the primate, part of the migrating

wave of neural crest cells comes to lie in a loose matrix at the rim of the optic cup and, at

40 days post-ovulation, these cells stream cen­

trally through the loose fibrillar material of the primary cornea that has formed in the cleft between the lens vesicle and the surface ecto­ derm. 9.10.11 .12 This centripetal movement of

neural crest cells is termed the first wave of mesenchymal migration ; a second wave later contributes cells that form the keratocytes

and stromal cells of the iris. 13 Macrophages precede the migration of crest cells, which move over a layer of fibronectin into the primary corneal stroma to form a loosely

The inducers that mediate these complex cellular interactions have not been identified but it appears that the histogenesis of each migrating cell is in part determined by its locus of origin on the neural crest, with later modi­ fication by glycoproteins in the environment through which it moves. 18.1 9,20 Needless to say, such a complex pattern may be modifed at several levels. The eventual formation of a monolayer appears to be determined by con­ tact inhibition, with the lens and the primary

cornea acting as substrata. 21 ,22 Experiments show that the cells can move over any surface made available to them, to the extent that an intact junction between the lens equator and the optic cup is required to prevent backward migration of endothelial cells over the neu­ roretina. 23 Removal of the lens vesicle results in a loss of maturation control with persistent multilayering of the endothelial cells on the posterior corneal surface. 24 Intercellular junctions begin to form when the endothelial monolayer is complete. Gap junctions develop first, the apical band is in place by the sixteenth week of gestation and an adult configuration is achieved by twenty weeks. 25 The formation of intercellular junc­ tions corresponds with the beginning of aque­ ous secretion and the maturation of the barrier and transport functions of the corneal endothelium. In the avian model, the gradual increase in corneal transparency during the second half of gestation is dependent upon normal thyroid function. 26.27 Thyroid hor­ mones may control the deposition of ground substance, and thus the ability of the stroma to bind water, or they may regulate the matu­ ration of the endothelial pump function. Beginning at the eighth week of gestation, endothelial cells deposit Descemet's mem­

arranged sheet of overlapping cells. 14 This

brane. 28 This first appears as patches of lamel­

the

sixteen

layer thins to a monolayer of endothelium by eighth

week

of

human

gestation.

Developmentally, it is not a true endothelium

and should more correctly be called the pos­

terior cell layer of the cornea or the posterior

epithelium15 because its derivation is quite distinct from that of vascular endothelium.

lar material and forms a complete layer by weeks.

Collagen

is

continuously

deposited thereafter until the eighth month of gestation, by which time the membrane forms a layer approximately

3 [-lm

thick. The col­

lagen of Descemet's membrane formed

in utero has a banded appearance by electron

Cells of a similar origin cover the anterior sur­

microscopy,

ber angle migrate peripherally to form the lining of the trabecular meshwork. 1 6,1 7

and consists largely of non-banded basement

face of the iris and those in the anterior cham­

with

a

microperiodicity

of

110 [-lm. 29 In contrast, Descemet's membrane

that is secreted subsequently is finely granular

391

THE CORNEAL ENDOTHELI UM

Fig.2e.

Fig.2a.

Fig. 2.

� ,

{

� �"

:.\:�� .

Fig.2b. membrane. The granular portion of the mem­ brane accumulates throughout life and may reach a thickness of between

10

and

40 flm.

There is, however, a wide variation in the rate at which Descemet's membrane is deposited, and a measurement of its thickness does not provide an accurate gauge as to a patients age.30

(a) Diagramatic representation of the migration of neural crest cells from a position dorsal to the neural tube to the posterior surface of the cornea, showing ectoderm (A), neural crest cells (B) , neural ectoderm (C) and neural tube (D) , (b) Outpouchings of the neuralectoderm at the level of the diencephalon form the optic vesicles which grow toward the surface ectoderm and induce the surface cells to elongate and form the lens placode, The diagram shows the optic vesicle (A), surface ectoderm ( B) and diencephalon (C). By a complex process of differential growth, the optic vesicles then invaginate to form the optic cups. The lens placode also invaginates and separates from the surface ectoderm to form the lens vesicle lying within the rim of the optic cup. The inner and outer cellular layers of the optic vesicles form the neuroretina and the retinal pigment epithelium respectively, and extend forward with the advancing rim of the optic cup to form elements of the iris. The epithelium of the cornea develops (following lens invagination) from surface ectoderm. (c) Crest cells move from a position between the surface ectoderm and the neural tube to lie at the tip of the optic cup. They migrate over the posterior surface of the cornea to form the corneal endothelium. The diagram shows the cornea (A), neural crest cells (B) and the lens (C). into old age.37,38.39 It has been estimated that between the ages of 20 and

80 years the reduc­ 0.52% per year. 40

tion in cell density averages

As the mean age of a population sample

increases, there is a spread in the range of their endothelial cell density counts,31 so that

2.

the measurement of cell density is not a

Morphology

At birth, the human endothelium comprises a

reliable index of the chronological age of a

monolayer of up to

cornea. Longitudinal data on the change of

500,000

cells, with a

density as high as 7,500 cells/mm2• During life, cell

cell density with age for individual corneas are

density progressively reduces (Fig. 3).31 ,32,33 An initial rapid decline occurs

observed reduction in cell numbers, the mean

in the first year, reflecting hypertrophy of a

cell area of the surviving cells increases33,41

fixed

population

of

endothelial

cells

not

yet

available.

Concurrent

with

an

in

and there is an increased variation of cell size.

response to continued corneal growth,34,35,36

These changes in cell density and shape with

Cell density continues to fall at a reduced rate

age have been observed to occur in the mon­

until the mid-twenties due to endothelial cell

key,42 rat,43 cat,44 dog45 and rabbit, 46 but in

loss, and there follows a more gradual decline

each of these species the adult mean cell

392

S. J . TUFT AND D. J . COSTER

with Descemet's membrane and, because no junctional complexes are present, endothelial cells are readily dislodged by mechanical injury.53 The endothelial cell cytoplasm is rich in organelles, suggesting active transport and protein synthesis, with large numbers of mito­

c

-----.::._/

chondria, both rough and smooth endoplas­ mic

reticulum,

and

a

well

developed

perinuclear Golgi apparatus. Pinocytotic ves­ icles pass from the posterior cell membrane to 10

..

80

30

Age (ye a) Fig. 3.

Graphical representation of the fall in endothelial cell density that occurs with age. An initial very rapid decline in infancy (A) levels off in the teens (8) to be followed by a gradual decline into old age (C). Details of the fall in density that occurs in childhood are sketchy, but available data suggests that there is approximately a 45% reduction in cell density by the age of one year, and that this figure is further reduced by 25% by the age of twenty years, with 50% of this loss occurring by the age offive years.

be released either into the lateral cell space or through the anterior cell membrane. Their formation is temperature dependent and they are presumably concerned with the active transport of water and metabolites. Pigment granules are occasionally seen in the cyto­ plasm of corneal endothelial cells and are probably derived from phagocytosed melanin that has been dispersed from the iris stroma or pigment epithelium.54 A circumferential band of actin filaments

density (about 2,500 cells/mm2) is remarkably

localised at the apical junction may facilitate

constant.35 Interestingly, it has been noted

cell movement to cover endothelial defects

that in the rat the total reduction in endo­

after injury (Fig. 5) and there is an extracel­

thelial cell numbers is of the same order of

lular glycoprotein matrix containing fibronec­

magnitude as occurs in humans, but the cell

tin and laminin. 55.56 Plasminogen activator has

loss is compressed into the shorter life span of

been detected in the cytoplasm of cultured

the species.43

bovine endothelial cells after injury ; this may

A typical human endothelial cell measures 5!-lm in height, is between

18

function

in vivo in a fibrinolytic system to

and 20!-lm in

maintain aqueous circulation by dissolving

width and has a round nucleus 7 !-lm in dia­

fibrin clots that threaten endothelial func­

meter. The posterior surface is covered by a

tion.57 Protein macromolecules 'float' within

variable number of microvilli that project

the lipid bilayer of the cell membrane.58 These

between' 0.5 and 0.6!-lm into the anterior

proteins may act as antigenic sites, ion pumps

chamber.47 Oligocilia, projecting 2 to 7!-lm

or enzymes.

from a pair of centrioles in the posterior cyto­ plasm, are present on some cells (Fig.

4),

particularly toward the corneal periphery.48.49

3.

Physiology

The endothelial pump

Transmission electron microscopy has shown

The maintenance of a transparent cornea

that cells are separated laterally by a gap of

depends upon the endothelium producing a

about 30!-lm, which is reduced to 3 !-lm at the

state of relative stromal dehydration.59 The

site of a gap junction toward the anterior

proteoglycan matrix that surrounds each col­

chamber.50 Focal tight junctions join the cells

lagen fibre of the stroma produces an imbibi­

at the apical third of the cell, but there are no desmosomal junctions.51 When the cells are

water

viewed from the posterior surface there is an

between epithelial cells form a barrier to

tion pressure (60 mm Hg) which tends to draw into

the

cornea.

Tight

junctions

overall hexagonal pattern with fine marginal

reduce the flow of water from the tear film

folds, but surface parallel histological sections

into the stroma, but the absence of a continu­

reveal that there are extensive and irregular

ous tight junction between endothelial cells

interdigitations between cells.52 The anterior

permits the free flow of aqueous into the

cell membrane of each cell is in direct contact

stroma. If this water is allowed to accumulate

393

THE CORNEAL ENDOTHELIUM

Fig. 41>.

Fig. 4a.

• .. C� .

.

.

.........

tl¢'?o'g�

.

.

.

\

.

.

.

..

_ .

.

.

�(J��

'-a Fig.4e.

Fig. 4c. Fig. 4.

Fig.4d.

The structure of corneal endothelium. A schematic representation of a cross-section of a typical endothelial cell (centre). (a) Details of the apical junction complex showing apical complex (A) , gap junction (B), terminal web (C) and nucleus ( D). (b) The basal structure of a cilium showing cilium (A) and basal complex (B). (c) Details of the surface features by specular microscopy and (d) by scanning electron microscopy.

s.

394

J. TUFf AND D. J. COSTER

Fig. s.

Polymerized microtubules crossing the cytoplasm of migrating endothelial cells in vitro, demonstrated by immunofluorescent labelling. Nuclei appear as dark intracellular areas. it produces stromal swelling and clouding. 60

intact endothelium, intraocular pressure is

The pump-leak concept of corneal hydration

thought to compress the stroma and force

proposes that there is a dynamic equilibrium

water out.64.

between the tendency of the stroma to swell

Research has recently been directed to

and the active transport of ions by an endo­

identify the sequence of metabolic events that

thelial pump to oppose the inward movement

determine the transparency of the cornea

of water.60 Fluid

(Fig.

transport

by

the

endothelium

6).

Insight into the selective transport of

ions by the endothelium has been gained by in

depends upon aerobic metabolism which can

vivo

be reversibly inhibited by cooling the cor­

measured the trans-endothelial potential or

experimentation

which

has

either

nea.6J Mitochondria provide the energy to

the degree of corneal swelling induced by

drive the system, producing ATP from glu­

changes of the ionic gradient across isolated

cose via the tricarboxylic acid cycle and the

perfused corneas.65•66 Under these conditions

hexosemonophosphate shunt; the required

a short circuit current (a measure of the flux of z an actively transported ion) of 27 f.lA cm- is

oxygen is thought to diffuse into the endo­ thelial (pOz

film

cells

from

the

anterior

chamber

55 mm Hg) rather than from the tear (pOz 155 mm Hg).6Z Movement of

=

=

water across the endothelium is passive, and follows the flux of actively transported ions across the posterior cell membrane.60 The

metabolically

active

processes

established

across

the

membrane.

This

current is reduced by about a third either by the removal of exogenous carbon dioxide from the perfusing solution on the aqueous side or by the prevention of the hydrolysis of

carbon dioxide with a carbonic anhydrase

of

inhibitor.67 It is further reduced by the remo­

deturgescence are assisted by passive mech­

val of exogenous bicarbonate, but it appears

anisms. Evaporation can increase the osmo­

that the contribution of metabolically derived

corneal stroma63 and, in the presence of an

Bicarbonate is required on the stromal side

lality of the tear film and draw water from the

COz is insignificant.('8

395

THE CORNEAL ENDOTHELIUM

1

Cf) I­ W

CO2+ H20=HC03- + H+ •

:E

w () Cf) w o

2

AQUEOUS

Fig. 6. Endothelial pump mechanisms. The linkage between ion flux and the mass movement of water from the stroma to the aqueous cannot be adequately explained. Although it is impossible to provide a definitive model of endothelial pumping, the following scheme summarizes aspects of our present understanding of ion transport. The pump can be inhibited by carbonic anhydrase inhibitors (l) or by ouabain (2). of the endothelium to maintain endothelial

of sodium from the aqueous humour inhibits

pump function and there is a net flux of bicar­

fluid flow across the endothelium and causes a

bonate ions from the stroma to the aque­

significant decrease in the short circuit current. 65 ,71 Secondly, ouabain, which irre­

OUS.64.65.66.67 Such data has led to the concept that the movement of bicarbonate is the sub­ strate for the endothelial pump,65.67.69.70.71 .72

versibly blocks the sodium-potassium ATPase

although it is recognised that sodium and

current.75,77.78 Thus sodium undoubtedly has

potassium could influence fluid transport (Fig. 6).73,74 A specific bicarbonate-dependent

an important role. If a cation such as sodium

ATPase pump has been identified, but it is

primary substrate, a standing potential would

associated

mem­

be established across the endothelium, with

branes/5.76 and it is thus difficult to envisage a

the aqueous side positive, but in the rabbit

with

mitochondrial

contribution by this pump to transcellular ion flux. The mechanism of the linkage of this

pump,

abolishes

the

short

circuit

were to be pumped into the aqueous as a

there is a standing potential of the order of

580 �V

of the opposite polarity across the

anionic pump to an energy source, and hence

intact endothelium. If sodium were to be

to bulk fluid flow, is not understood.

pumped into the aqueous, but could diffuse

Two observations argue against a simple

back into the stroma, it would be possible for

bicarbonate pump model. Firstly, the removal

a sodium-potassium ATPase pump to drive

396

S. 1. TUFT AND D. J. COSTER

the transport of bicarbonate.79 However, an

focal microscopes permit study of the endo­

understanding of the nature of the linkage

thelium from a high resolution photographic

between the sodium and bicarbonate systems

image.93,94 The surface detail and extra-cel­

awaits elucidation.80•81 An alternative theory

lular features may be visualised, and posterior

of enzymatically mediated sodium-bicarbo­

corneal rings, an artefact of applanation, per­

nate exchange at the anterior cell membrane

mit the relocation of individual cells. 95

is currently attracting interest,

while the

As already stated, cell density measure­

active transport of sodium at the baso-Iateral

ments

membrane by an ouabain-sensitive sodium­

patient age or the ability of the endothelium

potassium

ATPase , reappraised.82 83

pump

is

being

in

isolation

correlate

poorly

with

to maintain corneal deturgescence, but com­ puter-assisted morphometric analysis of a

The number of ionic pump sites per endo­

digitised image is a more sensitive indicator of

thelial cell may vary in certain circumstances.

endothelial damage. As part of a morpho­

In Fuchs' dystrophy, there is an increase in the

metric analysis, individual cell areas can be

pump function of the endothelium in the early

plotted against their frequency in a given

gradual

population to give a distribution curve. An

d�terioration in cell density, but pump sites

increase in the frequency of abnormally large

are markedly reduced in the end stages coinci­

cells produces a curve skewed to the right, and

stages of the disease,

despite a

ding with the onset of corneal oedema.84,85.86

Water normally enters the stroma transcel­

a 'coefficient of skewness' can be quantified.96 Other new terms have been introduced to

lularly or via the intercellular space.87 The

describe features of the cell mosaic: 'poly­

removal of calcium from the aqueous results

megathism' refers to irregularity of the mosaic

in a rapid intlux of water and stromal swelling,

due to the presence of abnormally large cells

which corresponds with the rapid disinte­

and

gration of the microfilaments of the junctional complex.88,89 The absence of glutathione, a tri­

toward

peptide, has a similar effect.90 Glutathione

'hexagonality'

refers

to

a regular hexagonal

a

tendency

pattern that

affords the greatest packing with optimum cell to membrane ratio.97,98 There is a positive

also appears to stimulate the endothelial

correlation between these two parameters

pump, possibly by preventing the reduction of

whereby larger endothelial cells also have a

intracellular ATP.

more variable shape. A departure from a homogeneous population appears to be asso­

4. The assessment of endothelial function

Light and electron microscopy can demon­

strate the morphology of endothelial cells, but other techniques are required to assess the endothelium as a living tissue.91 Tests of endo­ thelial function may be broadly divided into morphological techniques, for guaging the number and shape of endothelial cells, and functional tests that measure their metabolic capacity. Qualitative tests may be further div­ ided into tests of the pump or barrier mech­ anisms of the endothelium.

ciated with a reduced ability to withstand trauma but further studies are being con­ ducted to determine which parameter best retlects endothelial function. Specular microscopy has greatly increased our understanding of endothelial pathology by providing an estimate of cell density before and after surgery. The impact of this infor­ mation upon anterior segment surgical tech­ niques, the design of intraocular lenses, and the constitution of intraocular irrigating tluids and medications has been marked, However, specular microscopy does have limitations, most notably the degradation of the image by

Specular microscopy

In

1919, Vogt described the specular appear­

stromal oedema when an estimate of cell mor­ phology would be most useful. Serious sam­

ance of normal and diseased corneal endo­

pling errors may occur because of regional

thelium.92 Routine slit-lamp examination of

variations in cell density, particularly when

the endothelium is simple and adequate in

using the small image size of some non-con­

most clinical situations, but in more demand­

tact microscopes.99 Sampling error has been

ing circumstances, modern specular or con-

partly overcome by using wide field

(1 mm2)

397

THE CORNEAL E NDOTHELIUM

examination and by relocation of the image

rate of recovery and endothelial cell morphol­

using posterior corneal rings. A good correla­

ogy is poor. The corneal swelling response is

tion between endothelial morphology and

slightly reduced by aphakia, which may reflect

functional performance has not been identi­

an

fied. Certainly, until cell numbers become

anterior chamber following cataract extrac­

low, there is a poor correlation between

tion or reduced corneal innervation.111,112

increased

oxygen

availability

to

the

density and corneal thickness. Therefore,

The main reservation about the use of

although specular morphometry has helped

induced hypoxia to study corneal function

identify some of the caveats of intraocular sur­

concerns the fear that it may precipitate irre­

gery, it does not have a role in routine pre­ operative examinations, where an assessment of endothelial density made from visualisa­ tion of the specular image at the slit-lamp is 00 adequate.1

orescein across the endothelial layer. This is

testing

linear

relationship

between

Corneal fluorophotometry

A quantitative assessment of the endothelium may be made by measuring the passage of flu­

Endothelia function and hypoxic stress

The

versible corneal oedema in some patients with low endothelial cell counts.

stromal

hydration and thickness means that corneal

said to be a test of the endothelial 'barrier', but it is uncertain which phy siological func­ tion of the endothelium this test measures as

thickness can be used as an index of endo­

the normal endothelium does not provide a

Similarly, the phy siological function of the

fluorescein is taken orally, the rate of transit

thelial

damage

after

surgery.1 01.l02.1 03,104

endothelium can be assessed experimentally by measuring changes in corneal thickness fol­ lowing hy poxic stress. Corneal hypoxia is pro­ duced by exposing the anterior surface of the cornea to an atmosphere of nitrogen or by placing a thick contact lens onto the eye. Anaerobic metabolism by

lOS

the epithelium

leads to an accumulation of lactic acid which causes the corneal stroma to swell, either by directly inhibiting the endothelial pump or by creating an osmotic imbalance.106,107 Upon

significant barrier to the diffusion of water. If from the anterior chamber into the stroma of the cornea can be measured with a modified fluorophotometer. Alternatively, fluorescein may be drawn into the corneal stroma by ion­ tophoresis and its appearance in the anterior chamber observed.113 In normal subjects, the values for fluorescein transfer appear to remain constant with increasing age but there 70

restoration of air to the anterior surface, the cornea thins over the following one to two hours at a rate consistent with endothelial cell pumping (Fig,

7).

Evaporation augments the

deturgescence response by

increasing the

osmolarity of the tear film; ideally, the relative humidity during the experiment should be controlled. The induced changes in corneal thickness are small and require accurate pachymetry,108

but

if

careful

sequential

measurements are made the technique can provide an indication of the capacity for deturgescence. 1 09 The results of such studies have shown that,

2

where there is a low endothelial cell density, a large corneal swelling response to hypoxia and a slow recovery occur. There is also a reduced

recovery

with

increasing

age.110

Nevertheless, the correlation between the

Time (houri> Fig. 7.

Representation of the fall in central corneal thickness that occurs in the recovery phase following corneal hypoxia. Increased humidity will shift the curve to the right.

398

S. J. TUFT AND D. J. COSTER

is a correlation between an increased transfer coefficient and corneal thickness.114 Although the results of this test in the study

a process that is inhibited by cytochalasin B,131.132,133 Migrating cells have an increased number of intracellular organelles,

fewer

of disease have produced some conflicting

endocytotic vesicles and the cells do not

results, trends have emerged. A high fluores­

appear to possess a fibronectin coatY4

cein transfer has been consistently demon­

When cells meet in the centre of the wound,

strated in Fuchs' dystrophy"5 and a reduced

further movement is prevented by contact

transfer in the iridocorneal endothelial syn­

inhibition and a wave of cell rearrangement

drome.1I6 In general, the permeability of the endothelium to fluorescein is increased after an injury and, although an intact cell layer recovers over a period of days, the recovery of full endothelial function, as adjudged by this technique, may require months.117.118,119,120

spreads outward.135 Mitotic figures appear at the wound edge at

24-36

16

hours, reach a peak at

hours and reduce again to zero within

five days.129.130.136 Division is a major component of endothelial repair in the young rab­ bit137 but in species where endothelial mitotic

5. The endothelial response to injury

figures are rarely observed, such as the cat and 3 , primates, 137 1 38,1 9 cell slide continues for up to

Although endothelial cells are slowly lost

one year. This complex process of cell reor­

from the cornea throughout life, decompen­

ganisation

sation seldom occurs. If episodes of injury or

Between four and seven days after injury, the

inflammation reduce the density below a level critical to the maintenance of corneal detur­

has

been

closely

observed,

large elongated cells at the wound margin begin to regain their original size and shape by

gescence, physiological function fails and cor­

moving relative to one another, thereby 'pull­

neal oedema ensues. This critical density has

ing' the more peripheral cells inwards, This

been estimated to be

10-15% of the normal 300 and 500 cells

cell count or between

mm2.121,122 There is, therefore, a considerable physiological reserve; effective as

500

cells mm2 is as

3,000 cells mm2 in achieving cor­

neal transparency in the short term, The

sequence requires the repeated breaking and reformation of junctions between neighbour­ ing cells, and, although a cell may become separated

from

its

previous

neighbour,

macroscopic gaps do not appear in the extra­ cellular matrix (Fig. 8).140.141

mechanisms by which cell function adapts to

Mitotic figures have been demonstrated in

the increased demands that follow cell loss are

the endothelium of cultured human cor­

not understood but it may be that the number

neas,142.143,144.145 and it is claimed that mitotic

of transport sites on the membrane of each

figures may be visualised

endothelial cell increases,123

specular

Endothelial repair has been most exten­ sively studied in the rabbit but the events of

microscope,146

does not occur

in vivo with the

Mitosis,

however,

in vivo at a rate sufficient to

replace cells that are lost by ageing or injury

repair are similar in all species.124,125,126 A

and therefore, for all intents and purposes,

localised endothelial injury initially stimulates

the human endothelium must be regarded as a

repair by cells in the immediate vicinity of the

finite and irreplaceable resource. As a result

wound; these cells flatten and slide to cover

of the inability to replace cells, a uniform cell

the defect and damaged endothelial cells are

distribution across the cornea may never be

desquamated into the anterior chamber (Fig. 8).127,128 As cells more distant become

re-established and a vertical cell disparity,

involved, a density gradient is established,

with the largest cells at the site of the injury. Surrounding cells break their intercellular connections,

develop

migrate at a rate of

pseudopodia

with a reduced number of cells near the wound, has been observed to persist for up to 20 years.147,148,149 In contrast, the coefficients

of variation and hexagonality return towards

and

normal within five months, suggesting the

to

new matrix has attained stability.149 Amitotic

0.5 to 1.0 mm per day

cover the defect, becoming flattened and

cell

attenuated in the process,129,130 These cell

nucleus without subsequent cell divisionI46,150)

movements are facilitated by f-actin that poly­

and cell coalescence have been observed in

merises across the cell cytoplasm after injury,

man,

division

lSI

(division

of

the

interphase

and giant multinucleated cells can

TH E CORNEAL ENDOTHELIUM

399

Fig. Sb.

Fig. S.

Fig. Sa.

Fig. Ik:.

Endothelial repair occurs by endothelial cell slide, followed by some cell division to replace cells that have been lost. (a) Cells elongate and slide toward the wound (W). (b) In the process, they undergo intercellular re-arrangement. /40 (c) Nuclear division may be demonstrated by the uptake of tritiated thymidine (labelled cells arrowed).

S. J. TUFT AND D. J. COSTER

400

form following injury. The role of these cells in endothelial repair is uncertain.152 Although there is a rapid morphological recovery of an intact layer of endothelial cells following injury, physiological function is more slowly re-established. In the rabbit, where endothelial proliferation occurs, the permeability to fluorescein returns to normal within 14 days of transcorneal freezing, but there is a further delay of up to seven days before the stroma returns to normal thick­ ness.153,154,155 In the cat, which like the human has little endothelial regenerative capacity, functional recovery may be delayed several months.156 If Descemet's membrane is damaged at the time of surgery, the cut edges retract elas­ tically and curl toward the stroma. A fibrin clot forms on the exposed stroma, which is then covered by migrating endothelial cells. The cells covering the stromal bed secrete a new basement membrane but the damaged Descemet's membrane is not resorbed.157 Repair of the endothelial layer by cellular sliding has one major disadvantage. At sites of persistent mechanical injury caused, for example, by the leg of an intraocular lens or by a foreign body in the anterior chamber, cells migrate to the site of injury and are lost­ like lemmings streaming off a cliff-and there can be a gradual spread of corneal decompen­ sation arising from this focus. Cell loss after surgery

All surgical procedures that involve entry into the anterior segment will damage a proporTable I

Interacting mechanisms of endothelial cell loss

Pre-existing corneal disease

Concussion Previous surgery Uveitis Dystrophy

Operative factors

Direct endothelial contact with instruments Intraocular lenses Inappropriate irrigating fluids

Post-operative factors

tion of the endothelial cells. The number of cells so lost is roughly proportional to the degree of endothelial pleomorphism prior to surgery and to the amount of corneal manipu­ lation at the time of surgery.158,159,160 If cell loss is severe, bullous keratopathy may ensue (Table I), but because cell rearrangement as a result of injury is superimposed upon a slow physiological cell loss, the onset of bullous keratopathy can be insidious and delayed. This phenomenon was well demonstrated in the case of posterior keratotomy for myopia, where the full consequences were not appre­ ciated for twenty years after the surgical pro­ cedure was first introduced.163 Post-operative bullous keratopathy has become the most common indication for pen­ etrating keratoplasty in many centres. This has been associated with the widespread adoption of intraocular lens implantation (IOL) following cataract removal. Although figures for the percentage of endothelial cells that are lost following IOL implantation vary, certain trends have emerged. Early tech­ niques produced an unexpectedly high oper­ ative cell loss (up to 70%) because no effort was made to prevent the IOL touching the endothelium. 164 Awareness of this hazard led to the modification of operative technique and, initially, by the simple expedient of main­ taining the anterior chamber with air, cell loss was greatly reduced (Table II). The evolution of surgical techniques aimed at minimising cell loss has been steady. If one considers the numbers of eyes that progress to corneal decompensation, improvements in technique have reduced the figure from almost Hl% in early trials175 to between one and 2%, or even less if iris-clip lenses are excluded from the series.176 However, long­ term follow-up is still incomplete. The question of continued cell loss follow­ ing cataract extraction remains unresolved. The data suggest that a rapid fall in central cell density ceases between three months and one Table II

Estimated cell loss following surgery

Corneo-vitreal touch Persistent inflammation

Intracapsular extraction

Elevated intraocular

Extracapsular extraction

pressure Intermittent intraocular 161.162 lens touch

ECCE+IOL Phacoextraction AC Phacoextraction PC

165.166,167.168

6 to 15%

166.169,170.171 14 to 23% 16 13 to 20% 8 172.173 27% 167.171.174 10 to 15%

THE CORNEAL ENDOTHELIUM

year after simple extraction177 but that cell loss continues at a rate that is slightly higher than normal if an IOL is implanted, particularly if an iris supported lens is used.178.179.180,181 Mechanical trauma to the endothelium and incompatible irrigating fluids are the major contributory factors to operative cell loss. Contact with instruments,172.182 crystalline lens fragments/83 or the 'biophysical' adhe­ sive effect between the endothelial cell mem­ brane and the IOLI84 are the most frequent mechanical insults. The polymethylmethacry­ late (PMMA) of an IOL was thought to adhere to the lipid of the endothelial cell membrane and cause cellular disruption on contact,t85 but this has not been confirmed.186 Intact endothelial cells have been seen on the surface of IOLs removed from cases of pseudophakic bullous keratopathy, but it is uncertain whether these represent cells dis­ lodged at the time of the original surgery or cells that have subsequently migrated onto the surface of the lens.187,188 Interestingly, a coating of protein and cells on the surface of an IOL may alter the surface characteristics of the PMMA, making it less lipophilic and pro­ viding some degree of protection should there be inadvertent contact with the endothelium during subsequent surgery.189 By coating a hydrophobic IOL with hydrophilic sub­ stances,19O such as autologous plasma,191,192 methylcellulose,184,193 chondroitin sulphate, 194 or hyaluronic acid,195,196 this adhesive effect is reduced. An alternative approach toward safer IOL insertion is to make lenses of a hydrophilic polymer, such as hydroxymethyl­ methacrylate (HMMA).185,197 Penetrating keratoplasty for aphakic bull­ ous keratopathy is generally technically suc­ cessful, with up to 95% clear grafts at two years.198,199,200 The visual results are less encouraging, with only between 15 and 40% of patients achieving a best corrected visual acuity of 6/12 because of associated cystoid macular oedema and glaucoma.198,199,200 Like­ wise, up to 80% of corneal grafts for pseudo­ phakic bullous keratopathy remain clear at two years, with between 17 and 50% of patients having a visual acuity of 6/12 or bet­ ter.176,201.202,203 The visual prognosis is worse with iris supported lenses204 but, in the absence of intraocular inflammation or cys-

401

toid macular oedema, there is no evidence that removal of a securely fixed lens will improve the final visual outcome or affect graft survival. 201,205,206 Meticulous anterior vit­ rectomy may be required at the time of graft­ ing to prevent vitreocorneal touch and continued endothelial cell loss.207 Prolonged intraocular manipulation usually requires the intracameral use of potentially harmful substances that may con­ tribute to cell loss. To avoid damaging the endothelium special care must be taken to ensure that irrigating fluids have a correct ionic balance (200 to 400 mOsm), a correct pH (6.8 to 8.2), and that calcium and glu­ tathione are present. Preferaby, an energy source should also be available.208,209,21O Because the endothelial pump is temperature dependent, cooling by irrigating fluids can slow the metabolism and produce corneal oedema which obscures the surgeon's view, but even prolonged irrigation with cold physiological fluids (4°C) produces few per­ manent changes. The preservatives for medi­ cations intended for ocular use have a dose-dependent toxicity. Sodium bisulphite, the buffer and antioxidant frequently used with adrenalin, is toxic to rabbit endothelium after intracameral injection or perfusion at concentrations of greater than 1: 5,000,210,211 while thimerosol and benzalkonium chloride may damage the endothelium at the accepted antimicrobial concentrations.212 Toxic effects have also been demonstrated after sub-con­ junctival injection.213 The toxicity of anti­ biotics varies widely, gentamicin probably being the least damaging. Other preparations whose effect upon the endothelium has been studied include phenylephrine,214 acetyl­ choline,215 carbechoe16 and urokinase.217 As a general rule, it is unsatisfactory to use any of these drugs unless in a preparation designed for intraocular use. Air introduced into the anterior chamber during surgery is toxic to the endothelium. Animal studies demonstrate a reversible effect after contact for 30 minutes, with a more marked effect if the anterior chamber is filled with air that is allowed to resorb. Agents introduced into the eye following vitrectomy can damage the endothelium; the effects of sulphur hexafluoride (SF6) gas are probably

402

S. J. TUFT AND D. J. COSTER

similar to those of air, but perfluoropentane (C Fs) is more toxic because it is only slowly 3 resorbed.218,219,220,221,222 The presence of the lens provides protection against intraocular gas during pars plana vitrectomy.223 If silicone oil passes into the anterior chamber it has a markedly toxic effect upon endothelium,224 Acute local endothelial damage undoubt­ edly occurs following 'minor' surgical proce­ dures to the anterior segment, for example peripheral iridectomy or trabeculectomy, but it is rare for there to be a reduction in central endothelial cell density, Mechanical damage to the endothelium has been recorded fol­ lowing both continuous wave (Argon and Krypton) and short-pulsed (NdIYAG or excimer) laser irradiation, but the long-term clinical significance of this is uncertain, Con­ tinuous wave lasers most frequently cause damage during peripheral iridectomy and tra­ beculoplasty,225 ,226 when heat is conducted from adjacent pigmented structures.227 Short­ pulsed lasers produce damage when gener­ ated shock waves travel through the tissue or when high velocity tissue debris is emitted from the point of fOCUS,228,229,230,231

6, The endothelium and penetrating keratoplasty

An essential objective of penetrating kera­ toplasty is the transfer of viable endothelial cells, and their subsequent preservation is the necessary preoccupation of the corneal trans­ plant surgeon, The number of endothelial cells transferred on a graft is related to the status of the donor, in particular the donor age, and whether the eye has suffered any pre­ vious trauma. Endothelial cell numbers are also affected by the way the cornea is treated at the time of removal, the time elapsing between death and storage, the nature of the storage medium, and the time in storage. Subsequent cell loss occurs as a direct con­ sequence of surgical manipulation, post-oper­ ative inflammation or allograft rejection. The assessment of excised corneas prior to corneal grafting

In the absence of an accurate method for assessing the functional capacity of the endo­ thelium of donor corneas, young donors are

usually selected. This preference hinges upon the belief that a high cell density provides a reserve against subsequent cell loss, although there is a wide variation in endothelial cell density within any age group and a young donor does not guarantee a high initial endo­ thelial cell count or an improved long-term cell density. However it has been noted that grafts from older donors usually require a longer post-operative period to attain their final thickness,232,233 an observation that may reflect the reduced ability of the ageing endo­ thelium to re-form intercellular junctions, Although advanced donor age does not pre­ clude the use of a cornea for grafting, the life span of a transplanted endothelial cell is, as yet, unknown. The accurate prediction of the functional performance of a donor cornea is a desirable but, as yet, unattainable goaL Specular micro­ scopy permits a direct measurement of pre­ operative cell density and the identification of pathology such as cornea guttata, keratic pre­ cipitates, Descemet's folds and areas of cell loss, each of which may adversely affect graft outcome, The metabolic state of tissue prior to grafting has been investigated using the nuclear magnetic resonance (NMR) of 3lp-Phosphorus to assess the abundance of organic high energy phosphate metabolites (ATP) in donor corneas.234 Unfortunately, because of the low tissue mass of the endo­ thelium and the high signal from the meta­ bolically active epithelium, the technique does not resolve the endothelial response as distinct from the response of the whole cor­ nea. The NMR of de-epithelialised corneas labelled by incubation with 13C-Glucose could provide a more sensitive alternative.235 The endothelial cell response to grafting

Endothelial cells are readily dislodged from a donor button by surgical manipulation, and longitudinal studies show an exponential fall in the central cell density as cells migrate to cover gaps in the endothelial layer. Following grafting for keratoconus, where there is a high endothelial cell count 011 the recipient cornea, a central cell loss of at least 20 to 30% occurs on the graft during the first three post­ operative months, This rises to between 50 and 60% by the end of the second

THE CORNEAL ENDOTHELIUM

year.236,237,238,239,240,241,242,243 The rate of cell loss in subsequent years is slower, but is still greater than would be expected to occur by ageing processes alone.244 This cell loss is not the result of sub-clinical rejection and it is not prevented by topical steroids.242 When the recipient endothelial cell count is low, as in bullous keratopathy, endothelial cell loss con­ tinues at an accelerated rate which is reflected in the poor long-term results for grafting of the condition.245 Whether this is due to a late re-distribution of cells from the graft to the host cornea or continued inflammation is unclear. The mechanisms of endothelial repair following the insertion of a corneal graft are not fully understood. Nuclear labelling experiments and sex chromatin studies demonstrate that the host and donor popula­ tions of endothelial cells move in both direc­ tions across the corneal wound but do not intermingle.246,247,248,249 In the rabbit there is extensive migration/50 and the available data suggests that a similar mechanism is present in humans. However, following grafting for ker­ atoconus there is usually a higher final cell density on the host than on the donor cornea and a uniform cell density is never attained.251,252,253,254 Allograft rejection is the commonest cause of corneal graft failure.255 Although grafts placed into corneas which have never been inflamed or vascularised seldom reject, unlike grafts for keratoconus, grafts placed in a reci­ pient bed which has been inflamed or vas­ cularised are often rejected.256,257 The basis of this loss of corneal privilege is related to the accumulation of immunocompetent cells, particularly interstitial dendritic cells, in the diseased recipient cornea.258 These 'passenger cells' present foreign antigens to the host immunocytes-the first step on the afferent path of the allograft response.259 Following sensitisation, the efferent response is largely directed at the endothelium. Endothelial rejection appears clinically as either a diffuse uveitis with scattered keratic precipitates or as a rejection line enclosing a specific point of sensitisation, such as an iris synechia or vas­ cular ingrowth.255 Treated promptly and aggressively with steroids or antimetabolites, the rejection may be halted, with the

403

reappearance of a clear graft, but an abnormal appearance of the specular image persists.260 Following severe rejection, a pos­ terior collagenous layer is deposited beneath the endothelium and extensive cell loss may result in persistent oedema.261,262 Despite the role of the endothelium as a tar­ get for the allograft reaction, there is doubt as to the identity of the target antigens. Although most studies have reported that class I antigens are absent from adult corneal endothelium/63,264,265,266,267 there are notable exceptions.268,269 There is agreement, how­ ever, that class I antigens can be found on the endothelium of donors younger than two years of age.263,269 Class II antigens are not expressed on healthy endothelium, but there is evidence that they can be induced under the influence of regulatory cytokines27o,271 and in rejected corneas.272 Increased expression of class I and class II antigens has previously been shown to occur in other tissues during rejection273,274,275 and class I antigens can be upregulated after surgery alone.275 It would therefore appear that the endothelium does hold the target major histocompatibility anti­ gens for rejection but that they are only expressed under certain conditions which pre­ dispose to allograft rejection. The transplantation of endothelial cells onto a patient's own cornea has been pro­ posed as a technique to reduce the antigenic load of a penetrating keratoplasty. In rabbits, cultured endothelial cells can be seeded onto thin gelatine sheets and, when a confluent monolayer has grown, the sheet affixed to the back of a donor cornea.276 Alternatively, the cultured endothelial cells may be seeded directly onto exposed Descemet's mem­ brane.277 Both homografts and xenografts of corneal endothelium have had short-term suc­ cess in experimental animals, but the clinical applications remain distant. It should be noted that endothelial cells express class I antigens after culture.263,278 Experiments to seed the posterior corneal surface of cats with autologous vascular endothelial cells have not produced grafts that remain clear,279 and there is no evidence that vascular endothelium can assume a pump capacity comparable to cor­ neal endothelium. The synthesis of peptide mitogens that

404

S. J. TUff AND D. J. COSTER

stimulate the proliferation and differentiation of endothelial cells after injury has been demonstrated in cell culture .280 Although these peptides appear to have developed as a local regulatory system, cells can also respond to added growth factors and these have been used to stimulate the low replicative capacity of human corneal endothelium. A mitogenic effect upon endothelium in culture has been demonstrated for fibroblastic and epidermal growth factors/81.282,283 and priming tissue with these polypeptides may yield a higher final endothelial cell density after grafting. 276 This has not been used clinically. General Considerations of Endothelial Pathology

Despite the structural simplicity of the cor­ neal endothelium, it is involved in a bewilder­ ing range of diverse but often interrelated pathologies. For the clinician, the precise classification of a particular case can be diffi­ cult. This is partly because any classification tends to be descriptive as so little is known of the aetiology of these conditions. Further­ more, the terminology of general pathology is not always appropriate. For the general pathologist, a dystrophy may be defined as a disorder, usually congeni­ thl, of the structure or function of an organ or tissue due to a perverted nutrition. In its widest sense it includes agenesis, atrophy, hypertrophy, hyperplasia, and metaplasia, but in practice the term is usually applied to those disorders which do not fit into any other category. In ophthalmology, the corneal dys­ trophies are primary conditions although they may not manifest until adult life. Several other terms deserve classification. Dysgenesis refers to conditions that are the result of an abnormal migration of formative elements. Dysplasia is a descriptive term that has been loosely applied. At times it is used as an alternative term for dystrophy, describing an abnormal growth of tissue, although it should only be applied to developmental dis­ orders. This convention is usually followed by ophthalmologists, an exception being its use to describe changes found in the corneal epi­ thelium that are considered to be pre-malig­ nant. Within this framework, efforts have been made to group primary endothelial dis-

eases (dystrophies) in terms of the migration (dysgenesis) or differentiation (dysplasia) of a common neural crest precursor,284,285.286,287 and thus to link the ocular and the systemic neu­ rocristopathies (Table III). Any of these disease processes may be accompanied by endothelial metaplasia. Metaplasia describes a change in which one cell type assumes the phenotype of another similarly differentiated tissue (Fig. 9). This may occur in the endothelium as part of a developmental disorder or it may occur in response to injury. Epithelioid metaplasia is accompanied by the loss of contact inhibition, an increased expression of keratin, and the proliferation of surface microvilli; posterior polymorphous dystrophy is the best example, but similar changes occur in the irido-corneal endothelial syndrome .288 In fibroblastic meta­ plasia, the normally polygonal cells become fusiform and develop prominent cytoplasmic organelles. This appears to be a non-specific manifestation of endothelial distress and attempted repair and is observed after such varied stimuli as alkali burns, transcorneal freezing, graft rejection, vitreous touch and experimental uveitis. 289.290 Both types of meta­ plasia can be accompanied by the deposition of a posterior collagenous layer formed of long-spaced collagen .29 ! Although the primary source of the abnormal collagen is the endothelial cell,292,293 stromal keratocytes may contribute following rupture of Descemet's membrane .294 A peripheral migration of endothelial cells can accompany dystrophic disease (primary proliferation) or ocular injury (secondary pro­ liferation) . Primary endothelial proliferation most typically occurs in the iridocorneal endo­ thelial syndrome,295 while secondary prolifer­ ation may follow concussive injuries. Special Pathology of the Endothelium

The time-honored convention of classifying Table III

The neurochristopathies

Axenfeld-Rieger' anomaly Peter's anomaly Abnormal proliferation ICE syndrome Abnormal differentiation CHED PPMD Fuchs' dystrophy Acquired abnormality Metaplasia

Abnormal migration

405

THE CORNEAL ENDOTHELIUM

Fig. 9. (a) Diseased endothelium may develop an irregular mosaic with numerous microvilli on each cell, an appearan�e that may superficailly resemble epithelium. (b) A fibroblastic configuration with the deposition of collagen IS a non-specific feature of endothelial distress. disorders as developmental or acquired is

The Axenfeld-Rieger syndrome is the least

appropriate for the conditions which specific­

severely affected variant. There is only mild

ally affect the normal endothelium.

pleomorphism of the size and shape of the endothelial cells, Descemet's membrane can be attenuated, and an abnormal layer of col­

Developmental Conditions

Mesenchymal dysgenesis (Anterior chamber clel!vage syndrome) Mesenchymal dysgenesis presents as bilateral malformations of the anterior segment which are transmitted with an autosomal dominant or sporadic mode of inheritance. A radially symmetric arrangement points to an anomaly of the first wave of neural crest migration as the cause for the ocular signs/96.297 but the presence of non-ocular features involving the

lagen covers the prominent Schwalbe's line. These defects could represent an abnormal retention of primordial portions of the iris and anterior chamber angle (Table IV). At the other end of the spectrum of disease, Peter's

to

80%

a more generalised abnormality of neural

Table IV

of related syndromes : the Axenfeld-Rieger syndrome ;

posterior

keratoconus ;

(congenital

porates

a

glaucoma).

degree

of

the

endothelium with an abnormal Descemet's membrane. Individuals often exhibit features that suggest an overlap between the syn­ dromes

which

aetiology.299

emphasises

their

common

and

Features af mesenchymal dysgenesil96•297•3°O

Axenfeld-Rieger syndrome

incor­ of

50

of cases develop a secondary glau-

Peter's

Each

dysplasia

most

of patients with Peter's anomaly have

anomaly ; sclerocornea ; and trabeculodysgen­ esis

the

bilateral involvement3OO and between

teeth, facial bones and melanocytes indicates

structural abnormalities produces a gradation

demonstrates

maldevelopment of the anterior segment. Up

70%

crest derivatives.298 A combination of several

anomaly

severe form of dysgenesis with widespread

Peter's Anomaly

Prominent and anteriorly displaced Schwalbe 's line High insertion of iris Corectopia Hypoplasia of iris with holes Corneal leukoma Corneal-lenticular adhesions Anterior polar cataract Shallow anterior chamber Scleralisation of cornea

406

s.

1. TUFf AND D . 1. COSTER

coma.296 A more detailed review of the syn­ dromes within this group is presented elsewhere.301 .302,303,304 Histological examination of the posterior cornea in Peter's anomaly demonstrates an attenuation of the endothelial layer centrally, while the peripheral endothelium may be normal. D�scemet's membrane is absent or represented by mUltiple laminations of base­ ment membrane interspersed with collagen fibrils.305 ,306,307,308 Giant collagen fibrils, up to 60 [.tm in diameter, have been noted, 309 and similar fibrils have been demonstrated in both CHED3!O and following experimental lensec­ tomy in the chick embryo.311 Treatment is required in only a minority of patients. Keratoplasty and lensectomy may be required to clear the visual axis if there are corneolenticular adhesions and cataract, but the visual prognosis is poor. Endothelial Dystrophies

1. Congenital hereditary endothelial dystrophy (CHED)

This is a rare disorder that presents in children or young adults as a bilaterally symmetric dif­ fuse corneal oedema varying from a mild stro­ mal haze to gross opacification of the cornea (Fig. 10).312,313 ludisch and Maumenee recog­ nised differences in expression between the autosomal recessive and the autosomal domi­ nant forms of the disease.31 4 The recessive variant tends to present earlier, have more severe corneal opacification, and visual depri­ vation can produce a pendular nystagmus. The cornea in the dominant variant may appear to be normal in infancy but oedema usually begins to develop in the first two years, although visual deterioration may be delayed. Oedema is not always progressive, but epithelial breakdown can result in photo­ phobia and tearing, and secondary changes can produce band keratopathy and spheroidal degeneration. Associated abnormalities of the anterior segment are not a prominent fea­ ture of CHED and the primary differential diagnosis is between congenital glaucoma and a mucopolysaccharidosis. The corneal diame­ ter in congenital glaucoma is usually larger than normal and there is a raised intraocular pressure with Descemet's breaks ; the opacity in a mucopolysaccharidosis tends to be an infiltration rather than oedema.

Histological examination reveals a range of endothelial morphology. The endothelial cells demonstrate fibroblast-like changes and may be absent from the central cornea. Desce­ met's membrane is often represented by only the anterior banded zone and a disorganised posterior collagenous layer can be grossly thickened and beaded.31 5 The peripheral endothelium and Descemet's membrane are relatively unaffected. This distribution sug­ gests that the disease is due to an incomplete central migration of endothelial precursors, with secondary endothelial metaplasia and hypersecretion of collagen. The diameter of the stromal collagen fibres is increased and they show non-specific corneal long-standing changes of oedema.316,317 Histological similarities between CHED al'ld posterior polymorphous dystrophy suggest that there is an overlap between the two conditions, indeed it has been proposed that the two syndromes repre­ sent the clinical spectrum of a single congeni­ tal endothelial dystrophy.318,319 The decision to perform penetrating ker­ atoplasty depends upon the degree of stromal opacity., If there is marked opacity from birth, the operation should be performed early and an appropriate optical correction provided to reduce the period of visual depri­ vation and the risk of amblyopia. This type of surgery can be difficult and requires careful follow-up as graft rejection may be insidious and develop rapidly in the infant. Even with the most meticulous care, the prognosis for graft survival in young children is poor, with only about 40% of long-term grafts remaining

Fig. 10.

Congenital hereditary endothelial dystrophy, Central corneal clouding presented at birth; the anterior segment was otherwise normal.

THE CORNEAL ENDOTHELIUM

clear.32o If the degree of opacity permits a delay, the prospects for successful grafting improve markedly.321

2. Posterior polymorphous dystrophy (PPMD)

This is a slowly progressive disorder that is most often innerited as an autosomal domi­ nant trait, although in some instances the inheritance appears to be autosomal reces­ sive.322,323 Involvement is bilateral, but the expression of the disease can be highly vari­ able and asymmetric, such that early cases may appear to be unilateraL Vision is usually well preserved and only rarely does stromal and epithelial oedema develop. In some family studies vesicles and corneal oedema have been'noted at birth. Corneal changes are best visualised with a slit-lamp using a broad tangential beam or retro-illumination (Fig. 11), At the level of Descemet's membrane, clusters of vesicular lesions are arranged in groups, scalloped geo­ graphic areas, or bands that may superficially resemble traumatic breaks in Descemet's membrane.324 A faint grey stromal haze sur­ rounds these lesions. Two separate popula­ tions of endothelial cells can be seen at specular microscopy: large pleomorphic cells and small cells that have indistinct boundaries.325 The benign non-familial occurrence of small clusters of posterior corneal vesicles has been reported.326 These do not affect vision and are non-progressive , The occurrence of features of mesenchymal dysgenesis, the iridocorneal endothelial syn­ drome (ICE), and CHED in some families with PPMD suggests a related aetiol­ ogy.319,324,327 The basic anomaly of both PPMD and ICE seems to be a proliferation of meta­ plastic epithelioid cells, although prolifer­ ation onto the anterior iris is not common in PPMD and the two conditions vary in their morphology, transmlSSlOn and natural history, 328 Broad peripheral iridocorneal adhesions, a feature of some cases of PPMD, are more substantial than the finer adhesions that occur in the Axenfeld-Rieger syndrome and they are not associated with a prominent Schwalbe's line. Glaucoma rarely occurs in patients with PPMD, but when it appears it is

407

thought to result from a proliferation of endo­ thelium over the open angle or an anterior insertion of the iris into the trabecular mesh­ work-an abnormality that is more typically seen in primary congenital glaucoma (trabeculodysgenesis).329 Histopathological studies of PPMD are lim­ ited to the few cases that progress to corneal oedema, The presence of endothelial cells with epithelioid characteristics is remark­ able.330 Affected cells form multiple layers and overgrow the normal endothelial cells,331 have desmosomal junctions and numerous microvilli facing the anterior chamber. Endothelial cultures from affected corneas produce two distinct populations: cells of a normal appearance and cells having multiple microvilli.332 An increased expression of the keratin cytoskeleton has been demonstrated by immunohistochemistry in PPMD,333,334 indicating that the epithelioid cells are prob­ ably not derived from ectopic surface ecto­ derm but are metaplastic endothelial cells.335 The presence of only a thin posterior non­ banded Descemet's membrane suggests an onset of disease in the perinatal period.331,335,336,337,338,339,340 A thick posterior collagenous layer is deposited later and may be heaped into fusiform swellings that project into the anterior chamber or form loose fibro­ cellular material which fills infoldings of Des­ cemet's membrane. Both changes can pro­ duce the clinical appearance of vesicles.341 3,

Fuchs' endothelial dystrophy

This is the most common of the endothelial dystrophies, It is characterised in its earliest stages by guttata which appear as non-reflect­ ing points in the specular image of the endothelium or as refractile particles on retro­ illumination342 (Fig , 12), Primary -guttata appear in two patterns: isolated certral gut­ tata as part of the normal ageing process or confluent guttata accompanied by corneal oedema and reduced vision which constitutes Fuchs' endothelial dystrophy. There is a con­ tinuum of disease between these two extremes. Secondary corneal guttata may appear after corneal inflammatory disease, such as syphilitic interstitial keratitis.343 Fuchs' endothelial dystrophy is typically inherited as an autosomal dominant trait with

408

S. J. TUFT AND D. J. COSTER

Fig. 1 1 . Posterior polymorphous dystrophy. (a) Numerous 'vesicles' may be seen on the red reflex. (b) On specular microscopy, these appear as localised groups of abnormal cells (arrows) amidst a mosaic that has moderate polymegathism. Fig. 1 1a.

Fig. lIb. incomplete penetrance, although in some

endothelial dystrophy with chronic simple

series up to

glaucoma ,

50%

of cases are sporadic . 344 In

70%

ocular hypertension and acute

of

closed angle glaucoma would imply a general­

confluent in only 3.7 % .345 The overall female

but other authors report an incidence of

to male ratio in this series was

chronic simple glaucoma of only

one large series, guttata were seen in individuals over the age of

40 years, but were 2. 5 to one but,

when only the patients progressing to surgery

ised abnormality of the anterior segment,348

normal outflow facility.

1.4% 344 and a

were considered, the ratio increased to six to

Abnormalities of the endothelium are evi­

one . 344 Thus the disease is more common in

dent many years before symptoms manifest.

females and the phenotypic expression is

Symptoms typically develop slowly through

more severe . 346.347 An association of Fuchs'

three stages over a period of up to twenty

THE CORNEAL ENDOTHELIUM

Fig. 1 23 .

Fig. 1 2 b . Fig. 12. Fuchs' endothelial dystrophy. (a) Numerous guttata and localised stromal oedema may be seen on retroillumination. (b) Scanning electron microscopy graphically demonstrates the protruberent nature of the guttata and the way the remaining endothelial cells are at first stretched to cover them and are eventually lost.

years. The eye is initially asymptomatic as guttata appear centrally, the thickened Desce­ met's membrane becomes visible on tangen­ tial illumination, and pigment that has been phagocytosed'by endothelial cells forms geo­ graphic patterns. During a second stage of the disease, stromal and epithelial oedema pro­ duces glare and visual loss due to irregular astigmatism and the rupture of sub-epithelial vesicles can be painful. In the final stage, cen­ tral sub-epithelial scarring results in an

409

opaque cornea that may become secondarily vascularised. Histological examination of Descemet's membrane shows a normal 110 nm banded layer anteriorly with a thin amorphous layer posteriorly. Excrescences formed of long­ spaced collagen protrude into the anterior chamber, although these may be subsequently buried by the deposition of a posterior col­ lagenous layer. 349 Endothelial cell loss results in an increase in the surface area of the remaining cells which become thinned. 350 The intercellular spaces are widened with the loss of occluding junctions, and the development of desmosomes could represent fibroblastic metaplasia. 350.351 .352 Despite the loss of junctional complexes and a decrease in the endo­ thelial cell numbers, endothelial pump function may be significantly increased in the early stages of disease, which would delay the onset of stromal oedema. 353.354 Treatment aims to control discomfort and blurred vision . Increasing the tonicity of the tear film with hypertonic drops or blowing with a hair drier may help to clear vision, and a temporary bandage contact lens can lessen the discomfort from ruptured bullae. Topical steroids have no proven beneficial effect. 355 Penetrating keratoplasty should be con­ sidered when visual acuity falls to an unacceptable level . Despite the high percent­ age of graft successes in the early post-oper­ ative period, the long-term prognosis for keratoplasty has been reported to be poor ;356 this is based on one study in which all uncom­ plicated grafts had become cloudy at nine years follow-up. 357 Other long-term series, however, have reported success rates of up to 85% at 8 years. 358.359 Viral particles have been demonstrated in one case of Fuchs' dystrophy,360 but the aetiol­ ogy of the majority of cases remains unknown . Fibrinogen and fibrin deposition have been demonstrated in the posterior col­ lagenous layer by immunofluorescence and it is possible that the fibrinolytic system of the aqueous humour adversely affects the endo­ thelial cells. Systemic tranexamic acid, an anti-fibrinolytic agent, has been used with some success in clinical experiments to induce stromal thinning in Fuchs' dystrophy,361 but the mechanism of its action is undetermined.362

410

S. J. TUFT AND D. J. COSTER

4. The irido-corneal endothelial (ICE) syndrome (Primary proliferative endotheliopathy)

The ICE syndrome is particularly difficult to classify as it does not completely satisfy the criteria for both dystrophy and dysplasia . The condition is essentially unilateral with mini­ mal changes in the other eye ; it is usually spor­ adic but family clusters have been described and it has a characteristic clinical appearance, with dysplastic endothelium the dominant his­ tologic feature. Whether the condition should be considered with the dystrophies or whether it represents a true dysplasia (but grossly delayed) of the neural crest remains to be decided. For our purposes it is appropriate to consider the syndrome along with the endo­ thelial dystrophies. Essential iris atrophy, Chandler's syndrome and the Cogan-Reese syndrome are probably variants of a single disease process in which secondary corneal oedema, peripheral anterior synechiae and changes in the iris architecture are . the main clinical features363 (Table V). They form a group of progressive disorders collectively called the irido-corneal endothelial (ICE) syndrome364 (Fig. 13). The endothelium in this syndrome has a character­ istically hammered-silver appearance in specular reflection. 365 Patients with the ICE syndrome are typi­ cally young white females in their third to fifth decade.369 The female to male ratio is esti­ mated to be as high as 5 : 1370 and, interest­ ingly, affected males tend to have bilateral involvement. Although most cases are spor­ adic, a familial bilaterally symmetric variant has also been described. 371 Minor abnormalTable V

The iridocomeal endothelial syndromt!66.367.3611

Essential iris atrophy

Chandler's syndrome

Cogan-Reese syndrome

Broad iris synechiae Distortion of pupil Ectropion uvea Iris holes Sphincter pupillae preserved Corneal oedema over synechiae Few iris features Diffuse corneal oedema Secondary glaucoma Pigmented iris nodules in addition to the above features

ities of the endothelium in the contralateral eye, iris transillumination defects and a reduc­ tion in the outflow facility all suggest that this is a bilateral disease with asymmetrical involvement .372 ,373,374 Specular microscopic examination has demonstrated that there are two cell popula­ tions: a unique 'ICE cell' that has a very irregular shape and reversed cell highlights and cells of normal appearance that may be at either a greater or lesser density than the cells of the unaffected eye. 373,375,376 There are marked regional variations, with cells of normal configuration and areas of gross pleo­ morphism.377,378,379,380 When first examined, corneal involvement can be incomplete, in which case there is a sharp interface between the two populations of cells. Dystrophic oells can gradually replace the normal cells, an event that is frequently followed by an increase in intraocular pressure. At histological examination, the striking feature is the wide variety of endothelial change, filopodia formation, multilayering of cells and localised cell necrosis. A collagenous layer forms posterior to normal Descemet's membrane and beneath ectopic endothelium on the iris surface-the 'glass membrane' seen by light microscopy.381 Iris nodules of modi­ fied melanocytes can protrude through gaps in this ectopic membrane. 382 Features suggestive of epithelialisation of the posterior corneal surface have been demonstrated in cases of Chandler's syndrome383,384 and, while the orig­ inal diagnosis of ICE syndrome must be care­ fully considered, the increased expression of cytokeratins (epithelioid metaplasia) suggests a link between the ICE syndrome and PPMD.385 The features of the ICE syndrome result from degeneration of the endothelial cells on the posterior cornea. There is an associated, but paradoxical, proliferation of the remain­ ing cells onto the iris,364,385 although mitotic division is rarely observed.379,380 A reduction in outflow facility is presumably the result of the growth of endothelial cells onto the tra­ becular meshwork. 386 ,387 Contraction of the collagen formed beneath the ectopic endo­ thelium then produces peripheral anterior synechiae, an ectopic pupil, and traction tears of the iris. Evidence for vascular occlusion

Fig. 13 .

ICE syndrome. (a) Clinical photograph demonstrating areas of corneal clouding, broad anterior synechiae, and an iris hole. Chandler's syndrome presenting as diffuse corneal oedema in an otherwise asymptomatic eye. (b) On specular microscopy, the endothelial mosaic is irregular and there are several ICE cells with reversed highlights (arrows). (c) A line of demarkation may be seen in cases of incomplete endothelial involvement.

Fig. 138.

Fig. 1 3b.

.

412

S. J. T U FT AND D . J. COSTER

producing the iris holes in this disease is conflicting; some studies have failed to demonstrate any occlusion372 while others have demonstrated widespread vascular change.388.389 It is tempting to postulate that differences in the relative contribution of the three basic features of this condition--endo­ thelial cell loss, endothelial migration and the deposition of an ectopic Descemet's mem­ brane�ould be reflected as three clinical variants.39o The presence of normal Desce­ met's tissue beneath a posterior collagenous layer points to a post-natal onset of disease,379 interpreted as either the delayed expression of abnormal neural crest development or an acquired inflammatory lesion. Most patients with ICE syndrome maintain excellent vision and require only periodic review. The breakdown of the syndrome into three groups is of little clinical significance, particularly as patients may progress between groupS.369 However, it is important to make the distinction from other causes of iris tum­ our, for example iris melanoma, and to recog­ nise a tendency to develop a raised intraocular pressure. In the presence of glaucoma, management should initially be directed toward control of aqueous production by medical means, but filtering surgery is eventu­ ally required in about 50% of cases.39J Sub­ sequent proliferation of endothelium into the filtering bleb may cause failure of some oper­ ations, but surgery can be repeated. Symp­ toms of early corneal oedema may be controlled by hypertonic drops or a soft con­ t,act lens, but penetrating keratoplasty, fol­ lowing the control of raised intraocular pressure, may be required if corneal oedema produces significant visual loss or ocular dis­ comfort. There are no histological reports of recurrence of the primary disease in the donor corneal button.392.393 Acquired Disorders-the Endothelial Response to External Influences

The corneal endothelium has a limited reper­ toire of responses to challenge. Sub-lethal injury results in minor pleomorphic changes in the cellular array, more damaging injury may produce metaplasia, but cell loss is the most common response to trauma. Surgical intervention is the most common noxious stimulus experienced by the cornea.

Surgical Injuries

1. Epithelial downgrowth and other post­ surgical retrocorneal membranes

Cells other than endothelial cells may be seeded onto the posterior corneal surface where they may grow. Epithelial cells may spread after traumatic implantation or spread through a gap in the integrity of the globe. Study of organ cultured corneas suggests that epithelium will not overgrow healthy endo­ thelium/94,395 and progressive endothelial disease seems to be a prerequisite for clinical disease,396,397,398,399 However, in an inverted keratoplasty model of epithelial downgrowth developed in the cat, epithelial cells grow across residual endothelium onto the iris.400 Iris pigment epithelial cells, iris stromal mel­ anocytes or pigment-containing macrophages can also be seeded onto the back of the cor­ nea, where they develop into pigmented retrocorneal membranes.40J,402 Melanocytes in this location may cover a large area, but such membranes do not invariably induce cor­ neal oedema. It has been suggested that, as they are of neural crest origin, they are able to develop a pumping facility.403,404 Alter­ natively, functioning endothelial cells may survive beneath the melanocytic layer, These hypotheses require further study. 2,

Post-Surgical Endothelial Cell Loss

Loss of some endothelial cells after intra­ ocular surgery is inevitable. This subject has been discussed in detail in a previous section.

Non-Surgical Injuries to the Endothelium

1. Concussive injuries

Breaks in Descemet's membrane can produce endothelial damage after forceps injuries, in congenital glaucoma and keratoconus, and in high myopia.405 The cell loss after a blunt injury is then exacerbated by an associated hyphaema, uveitis or raised intraocular pres­ sure.406,407 A condition termed traumatic pos­ terior annular keratopathy has been observed to follow the impact of high velocity particles on the superficial stroma, most frequently after powder flash or sand blasting injuries: as particles strike the cornea they generate a shock wave that passes backward from the point of impact producing a ring of vacuolated

413

THE CORNEAL ENDOTHELI U M

endothelial cells with an adherent plaque of

pump is overwhelmed, 427,428,429 The endothe­

leukocytes and fibrin. This usually heals with­

lium can adapt to a gradual increase in

out significant endothelial cell loss.408

IOP,430,431 but rapid rises in l OP result in endo­

Secondary proliferation of the endothelium

thelial cell damage. The cell loss after an acute

can follow blunt ocular injury, especially if the

angle-closure attack is roughly proportional

blow is of sufficient force to produce an angle l recession;409,4 IO, 41 less frequently it is seen in

to the duration of stromal swelling during the

association with rubeosis iridis.412 There is a

reported, 432,433 In contrast, cell loss has not

reduction in the density of the remaining

been consistently demonstrated in primary

endothelial

cells

and

obstruction

of

the

attack; a mean cell loss of

23%

has been

open angle glaucoma or pigmentary disper­ sion glaucoma,434 although there appears to

anterior chamber angle.295

be a tendency for cell loss in pseudexfoliative 2.

glaucoma. 435,436,437 In other situations it is diffi­

Hypoxia

Transient bleb-like structures in the specular image develop after overnight lid closure, 413 or within minutes of placing a low oxygen transmission contact lens onto the unadapted eye.414

These changes,

and

a

co-existent

increase in corneal thickness, could be the direct result of endothelial hypoxia, 415 but more probably are the result of epithelial hypoxia, the retention of CO and the accu­ 2 mulation of lactic acid producing an osmotic

cult to isolate the effect of increased lOP ; for example, a

10%

cell loss has been reported

following glaucomatico-cyclitic crises,435 but this may be due in part to associated inflam­ mation. A similarly complex pattern occurs in herpes zoster ophthalmicus, where there can be a considerable cell loss but the relative con­ tributions of viral endotheliitis, inflammation, and ischaemia are not known.438

imbalance.416,417 Although the acute changes

4.

persist for only one to two hours after removal

The cornea is effectively opaque to radiation

of the contact lens, a reduced regularity of the

of less than

endothelial matrix has been documented after

to broad band

the long-term use of both daily wear and

induce endothelial swelling, desquamation

extended wear contact lenses wear­ J ers.4 8,419,420,421 , 42? ,423 Although there is no con-

and

vincing evidence of a resulting abnormality of

thelium

endothelial function, P MMA contact lenses

induced by epithelial damage is unclear.

Electromagnetic radiation

the

290 nm,

but prolonged exposure

UV-B has been documented to

development

of

microvilli.439,440

Whether this is a direct effect upon the endo­ or

subsequent

to

inflammation

may increase endothelial cells loss if worn

Other wavelengths have no demonstrated

after penetrating keratoplasty, 424

effect.

Anterior segment hypoxia could contribute directly to the endothelial cell loss and corneal oedema that can follow temporal arteritis, anterior segment necrosis, herpes zoster oph­ thalmicus and acute closed-angle glaucoma.425 This is supported by experimental obser­ vation in which ablation of the long posterior ciliary arteries in the rabbit produces a sharp drop in the anterior chamber oxygen tension and glucose concentration, which is followed by corneal oedema.426

Effects of Ocular Disease

1. Uveitis

Leukocytes that enter the anterior chamber can migrate between and beneath endothelial cells where they appear as pseudoguttata on the specular image.441 ,442 Following mild uvei­ tis, these changes are reversible, but marked cell loss can follow severe uveitis, especially beneath mutton fat keratin precipitates.443 This may be the result of mononuclear cells, which have been observed to penetrate apical

3,

Raised intraocular pressure

In the presence of an intact endothelial layer,

junctions in experimental uveitis and to dis­ lodge endothelial cells into the anterior cham­

an increase in intraocular pressure (lOP ) will

ber.444 Polymorphonuclear leukocytes in the

initially produce corneal thinning; corneal

anterior chamber can also produce toxic

oedema is produced only if the endothelial

endothelial damage and oedema.445 Similar

414

S . J. TUFT AND D. J . COSTER

changes have been produced by the hydrolytic enzyme cathepsin and by neutral proteases that digest cell surface proteins and glycoproteins.446

2. Viral infection

The herpes simplex virus can be cultured from the stroma of a percentage of patients with disciform keratitis and it is probable that the virus can invade endothelial cells directly. The resulting endotheliitis and secondary attack by immunocompetent leukocytes produces pseudoguttata and oedema of the overlying stromal and epithelium.447 A bilateral and symmetric stromal oedema that moves centrally behind a line of keratic precipitates has been termed presumed auto­ immune endothe1iopathy.448 Although it responds readily to topical steroid, the aetiol­ ogy of this rare condition remains an enigma. Similar, but unilateral, cases that do not respond to steroid have been reported.449 The anterior chamber in these cases contains macrophages and lymphocytes, but attempts to isolate viruses have been unsuccessful. Endothelial cells may also be involved by disease processes that have their primary effect upon other corneal layers. For example, there may be endothelial deposition of acid mucopolysaccharide in macular cor­ neal dystrophy.450 Systemic Metabolic Disease

Diabetes

Human diabetic corneal endothelium appears to have a normal cell density but it is more sus­ ceptible to stress than normal endothelium and stromal oedema following surgical pro­ cedures may be slow to resolve.451 The endo­ thelium of both juvenile and adult onset diabetic patients is often abnormal on mor­ phometric analysis452 and has a higher per­ meability to fluorescein.453 A failure to demonstrate an increase in the thickness of the basement membrane of the endothelium (Descemet's membrane) comparable to increases demonstrated elsewhere in dia­ betics is thought to reflect the low turn-over of these cells.454.455 Conclusion

That so much attention has been directed to

studies of the corneal endothelium by labora­ tory-based and clinical scientists is an indica­ tion of the importance and complexity of disorders affecting this seemingly simple monolayer of cells. The pathway to a com­ plete understanding of the disorders of the endothelium extends before us. Much has been learned in the last two decades, and is the subject of this review, however much remains to be learned in the years to come. With the advances in basic science, particu­ larly molecular biology, transplantation and immunology, powerful tools have been devel­ oped which are being used to increase an understanding of the biology of the corneal endothelial cell. Progress in this area of research is accelerating and the prospects of conquering the common disorders of the endothelium in the forseeable future are very real. We would like to thank Dr Emil Sherrard and Dr Ali­ son McCartney of the Institute of Ophthalmology, London , and Dr Keryn Williams of the Flinders Medical Centre , Adelaide . for reviewing various aspects of the manuscript, Dr Nick McKecknie for providing Fig. 5, Emil Sherrard for Fig . 13c and Dr Nan Carroll of the Royal Victorian Eye and Ear Hos­ pital , Melbourne , for providing scanning electron micrographs. References I

Ozanics V and Jakobiec FA : Prenatal Development of the eye and its A d n e x a . I n Duane TO , Jaeger E A e d s . Bio­

medical Foundations of Ophthalmology. Philadelph i a : Harper

&

R o w 1 9 8 2 : Vol I , Chapter 2 .

2 Weston J A : The migration a n d differentiation o f neural

crest cells. A d v Morph 1970,

8:

41-1 14.

3 Noden O M : A n analysis of t h e m i gratory behaviour o f avian cephalic neural crest � e l l s . D e v Bio! 1 975 ,

42:

1 0 6-30 . 4 Le Lievre CS and Le Douarin N M : Mesenchymal deriva­ tives of the neural crest: analysis of chimeric quail and chick e mbryo s . J Embryo! Exp Morph 1975 , 34: 1 25-5 4 .

5 Johnson M C , N o d e n O M , Hazelton R D , el a!: Origins o f

avian ocular and periocular tissues . Exp Eye R e s 1 979,

29: 27-4 3 . 6 N o d e n O M : Periocular mesenchym e : n e u r a l crest a n d mesodermal interacti o n s . I n D u a n e T O , Jaeger EA

e d s . Biomedical Foundations of Ophthalmology. Phila­ 7

delphia : Harper

&

Row 1 9 8 2 : Vol I, Chapter 3 .

Johnston M C : A radioautographic study o f the migration and fate of cranial neural crest cells in the chick embryo.

Anal Rec 1966, 156: 1 43-56 . 8 L e Douarin N M : Characteristiques ultra structurales du

noyau interphasique chez la caille et chez Ie poulet et

utilisation de cellules de caille ' marquers biologiques'

e n e m b ryologie experimentalc . A n n Embryol Morpho! 1 9 7 1 , 4: 1 25-35 .

9 Duke-Elder

S

and

Cook

C o. Normal

and

abnormal

developme n t . I n Duke-Elder S e d . System of ophthal­ mology St Loui s : CV Mosby 1 963 : Embryology.

Vol 3, Part I ,

415

THE CORNEAL ENDOTHELIUM

l U Schook P : Cell action and interaction during ocular mor­

phogenesi s . In Duane TO , Jaeger EA e d s . Biomedical

" Speedwell L, Novakovic P, Sherrard E S , Taylor O S : The

infant corneal endothelium . A rch Ophthalmol 1 98 8 ,

Foundations of Ophthalmology. Philade l p h i a : Harper II

&

Row 1 98 2 : Vol 1 , Chapter 5 .

Hay ED: Development of the vertebrate corn e a . Int Rev

1 0 6 : 771-5 . 3h B a h n CF, Glassman R M , MacCallum O K , et al: Postnatal

development of corneal endothelium . In vest Ophthal­

Cyto1 1 9 8 0 , 63: 263-322 .

'2 0zanics V, Rayborn M, Sagun 0 : Observations of the morphology of the developing primate corn e a . J Mor­ phol 1 97 7 , 153: 263-9 8 . 1 3 Mann I : T h e development of the human eye . 5th e d . New

Yor k . Grune & Stratton Inc, 1 964. 1 4 Kurkinen M, Alitalo K, Vaheri A, et al. Fibronectin i n the

development of the embryonic chick eye . Dev Bioi

1 9 7 9 , 69: 589-600 . 1 5 Meier S: Initiation of corneal differentiation prior to cor­

neal-lens association . Cell Tiss Res 1 97 7 , 184: 25 5-68 . 16 Hansson H - A and Jerndal T: Scanning electron mi cro­

scopic studies of the development of the iridocorneal angle i n human e y e s . In vest Ophthalmol Vis Sci 1 97 1 ,

1 0 : 252-65 .

1 7 Van B uskirk E M : Clinical implication of iridocorneal

angle developme n t . Oph thalmology 1 9 8 1 , 88: 3 6 1 -7 . 1 8 Noden OM: The control of avian cephalic neural crest cytodifferentiation. Dev Bioi 1 97 8 , 67: 296-3 1 2 . 1 9 Coulombre AJ : Regulation of ocular morphoge n e s i s .

8:

In vest Ophthalmol Vis S c i 1 96 9 ,

25-3 1 .

20 Weston J A : The regulation of normal and abnormal neu­ ral crest develop m e n t . A d v Neural 1 98 1 , 29: 77-95 .

2 1 Genis-Galves JM: Role of the l e n s in the morphogenesis of the iris and corn e a . Nature 1 9 6 6 , 2 1 0 : 209- 1 0 .

22 Nelson G A and Revel J-P: Scanning electron m i croscopic

study of cell movements i n the corneal endothelium of the avian embryo . Dev Bioi 1 975 , 42: 3 1 5-37 .

2 3 Bard J B L , H a y E D , Meller S M : Formation of the endo­

thelium of the avian corne a : a study of cell movement i n

vivo . Dev B i o i 1 975 ,

42 :

334-6 1 .

24 Zinn K M : Changes in corneal ultrastructure resulting from

early

lens

removal

in

the

developing

mol Vis Sci 1 986 , 27: 44-5 1 . 17Sturrock G O , She rrard E S , Rice N S C : Specular micro­ scopy of the corneal endothe lium . B r J Ophthalmol 1 978 , 62: 809- 1 4 .

1 R Hoffer KJ and Kraff MC: Normal endothelial cell count

range . Ophthalmology 1 98 0 , 87: 861-5 .

39 Wilson RS and Roper-Hall M J : Effect of age on the endo­

thelial cell count i n the normal eye . Br J Ophthalmol 1 982 , 66: 5 1 3-1 5 .

4 [)

Murphy C , Alvarado J , Juster R , e t ale Prenatal and post­ natal cellularity of the h u m a n corneal endothelium . In v

Ophthalmol Vis Sci 1 9 8 4 , 2 5 : 3 1 2-22 . 4 1 Laule A , Cable M K , Hoffman C E , et a l : Endothelial cell

population changes of human cornea during life . A rch Ophthalmol 1 97 8 , 96: 203 1-5 .

42 B aroody R A , Bito L Z , Derousseau CJ , Kaufm a n P L :

Ocular develor m e n t of ageing I . Corneal endothelial

changes i n cats and i n free-ranging and caged rhesus monkeys. Exp Eye Res 1987,

43 Fitch

KL,

45:

Nadakavukaren MJ ,

607-22. Richardson

A:

Age­

related changes i n the corneal endothelium i n the rat . Exp Gerontal 1 9 8 2 , 1 7 : 1 79-83 .

44 M acCallum O K , B a h n CF, Lillie J H , et ale Evidence for

corneal endothelial cells hypertrophy during postnatal

growth of the cat corn e a . In vest Oph thalmol Vis Sci 1983 , 24: 247-50 .

45 G w i n R M , Lerner I , Warren J K , et a l e Decrease i n canine corneal endothelial cell density and i n crease i n corneal

thickness as functions of age . In vest Ophthalmol Vis Sci 1 98 2 , 22: 267-7 1 .

4h Oh J O : Changes with age i n the cornea of normal rabbits. A cta Ophthalmol 1 96 3 ,

41:

5 68-7 3 .

chick

4 7 H o g a n M J , Alvarado J A , Weddell J S : I n Histology of t h e

25 Wul l e K G : E l ectron microscopy of the fetal development

" G all agher B : Primary cilia of the corneal endothelium .

11:

49 Svedbergh B and B i l l A: Scanning electron m i croscopic

embryo. In vest Ophthalmol Vis Sci 1 9 7 0 , 9 : 1 65-82 .

of the corneal endothelium a n d Descemet's membrane of the human e y e . In vest Ophthalmol Vis Sci 1 972 ,

897-904 .

26 Coulombre AJ and Coulombre JL: Corneal development

Am J A nat 1 98 0 ,

159:

475-84 .

studies of the corneal endothelium i n m a n and mon­

k e y s . Acta Ophthalmol 1 9 n , 50: 32 1-36 .

I I I . The role of the thyroid i n dehydration and the

50 Kreutziger G O : Lateral membrane morphology and gap

1 05- 1 4 .

Eye Res 1 97 6 , 23: 285-9 3 . 5 1 Hirsch M , Renard G , Faure J P, et al: Study of t h e ultra­

development of transparency. Exp Eye Res 1 964 , 3 :

2 7 Masterson E , Edelhauser H F, Van Horn D L : Role of thy­

roid hormone i n the development of the chick corneal

endothelium and epithelium . In vest Ophthalmol Vis Sci 8

human eye . Philadelphia : W . B . Saunders 1 97 1 : 102-3 .

1 9 7 7 , 1 6 : 105- 1 5 .

2 Wul l e K G , Ruprecht K W , W i n d r a t h L C : E l e c t r o n micro­

scopy of the development of the cell junctions i n the embryonic

and

fetal

human

corn e a l

endothelium .

In vest Ophthalmol Vis Sci 1 974 , 1 3 : 923-3 4 .

29 Waring G O , L a i b s o n P R , Rodrigues M M : C l i n i c a l and pathological alterations of D e scemet ' s membrane with

emphasis on endothelial metaplasi a . Surv Ophthalmol 1 974 ,

18:

325-68 .

30 Johnson D H , Bourne WM , Campbell R J : The ultrastruc­

ture o f Descemet's membrane I. A rch Ophthalmol

1 982 , 100: 1 942-7 . 3 1 Laing R A , Sandstrom M M , B e rrospi A R , et al: Changes

i n the corneal endothelium 'as a function of age . Exp Eye Res 1 97 6 , 22: 5 87-9 4 .

32 B o u r n e WM a n d Kaufman H E : Specular m icroscopy o f human corneal endothelium i n vivo . A m J Oph thalmol 1 9 7 6 , 8 1 : 3 1 9-23 .

3 3 B latt H L : Endothelial cell density i n relation t o mor­ phology. Invest Ophthalmol Vis Sci 1 979 ,

18:

856-9 .

34 Hiles D A , Biglan AW, Fetherholf E C : Central corneal

endothelial cell counts i n childre n . A m Intra- Ocular Implant Soc J 1 979 , 5 : 292-300 .

j unction structure i n rabbit corneal endothelium . Exp

structure of the rabbit corneal endothelium by the freeze-fracture technique : apical and lateral j unctions .

Exp Eye Res 1 97 7 , 25: 277-8 8 . 52 Ringvold A , D avanger M , Olsen E G : O n the spatial organisation of the corneal e n dothelium . A cta Ophthal­

mol 1 984, 62: 9 1 1- 1 8 . 53 Waring G O , Laibson PRT, Rodrigues M : Clinical and

pathological alterations of Descemet's m e m brane : with

emphasis o n endothelial metaplasi a . Surv Ophthalmol

1 974 , 18 : 325-68 . 54 Kampik A , Patrinely J R , Green RW: Morphologic and clinical features of retrocorneal melanin pigmentation

and pigmented pupill ary membrane . Surv Ophthalmol

1 9 82 , 27: 1 6 1-80 . 55 Gordon S R , Essner E,

Rothstein H :

In situ demon­

stration of action i n normal a n d injured ocular tissues

using 7-nitro-benz-2-oxa 1 , 3 diazole phallacidin . Cell

Moti1 1 982 , 4: 343-5 4 . 56 Gospodarowicz 0 , Greenburg G , Vlodavsky L , et a l : The

identification and localisation of fibronectin i n cultured corneal

endothelial

cells .

Exp Eye Res

1979,

29:

485-509 .

57 Fehrenbacher L, Gospodarowicz 0 , Shuman M A : Syn­

thesis of plasminogen activator by bovine corneal endo­ thelial c e l l s . Exp Eye Res 1979, 29: 2 1 9-28 .

416

S . J. TUFf A N D D. J. COSTER

58 Singer SJ and Nicholson G L : The fluid mosaic model of

82 Kaye G I and Tice LW: Studies o n the cornea . V. Electron

microscopic localisation of adenosine triphosphatase

the structure of cell membrane s . Science 1972, 1 7 5 :

activity i n the rabbit cornea i n relation to transport.

720--3 1 .

59 Mishima S : Corneal thickne s s . Surv Ophthalmol 1 9 6 8 , 1 3 : 57-9 6 .

60 H e d b y s B O , Mishima S , M a u r i c e D M : The inhibition

pressure of the corneal stro m a . Exp Eye Res 1 963 , 2 : 99-1 1 1 . 6 1 Hm-ris J E and Nordquist LT: The hydration of the corn e a . A m J Oph thalmol 1955 , 40 : 1 00-- 1 0 .

62 Weissman BA, Fatt I, Rasson J: D i ffusion of oxygen in

In vest Ophthalmol Vis Sci 1966, 5 : 22-3 0 .

" Lyslo A , Kv e r n e s S , G a r l i d K , et al: Ionic transport across corneal

63 Mishima S and Maurice D M : The effect of normal evap­

oration upon the eye . Exp Eye Res 1 96 1 , 1: 46-5 2 . 64 M a y e s K R a n d Hodson S : S o m e effects of hydrostatic pressure o n corneal hydration during specular micro­

scopy. Exp Eye Res 1 97 8 , 26: 1 4 1 -5 . 65 Hodson S and M i l l e r F: The bicarbonate ion pump in the

endothelium which regulates the hydration of the rabbit corn e a . J Physiol 1 976, 263: 563-7 7 .

tochemical localisation of ATPase i n human dysfunc­ tional corneal endothelium . Curr Eye Res 1 987 , 12:

1 479-8 6 . 86 S a s a k i Y, Turberville AW, W o o d TO , M c L a u g h l i n B J :

Freeze fracture s t u d y of h u m a n corneal endothelial dys­

function . In vest Oph thalmol Vis Sci 1 98 6 , 27: 480-5 .

" Mishima S and H e dbys B O : The permeability of the cor­

neal epithelium and endothelium to w a t e r . Exp Eye Res

1 9 67 , 6 : 1 0--3 2 . 88 Kaye G I , Hoefle FB , Donn A: Studies o n the corn e a .

V I I I . Reversibility of the effects o f i n v i t r o perfusion o f

the

68 Lim J J : Effects of bicarbonate on the potential difference

across the rabbit corneal endothelium . Curr Eye Res 1 9 8 2 , 2 : 529-3 5 .

20-9.

7 1 Fischbarg J a n d Lim J J : Role o f cations , anions and car­ endothelium . J Physiol 1 9 7 4 , 24 1 : 647-75 .

73 Fischbarg J , Lim JJ , Bourguet J : Adenosine stimulation

74 Neuwirth 0 and Dickstein S : T h e effect of cyclic A M P o n

t h e rabbit corneal endothelial fluid pum p . Curr Eye Res 1 9 8 2 , 8 : 565-7 .

75 Riley MV:

Anion-sensitive ATPase i n rabbit corneal

endothelium and its relation to corneal hydration . Exp Eye Res 1 977 , 25: 483-9 4 .

thiol-oxidation of glutathione with diamide o n corneal

thelium . XXXV I I I Edward Jackson memorial lecture .

Am J Ophthalmol 1 9 8 2 , 93: 1-29 .

92 Vogt A: Die sichtbarkeit des lebenden hornhautendoth­ els i m lichtbueschel der gullstrandschen spaltlampe . Klin Monatsbl A ugenheilkd 1 9 1 9 , 63 : 233-4 .

93 Laing R A , Sandstrom M M , Leibowitz H M :

endothelial cells. Invest Ophthalmol Vis Sci 1 98 1 , 2 1 : 8 1 9-25 . 77 Trenberth SM and Mishima S: The effect of ouabain on

the rabbit corneal endothelium . In vest Ophthalmol Vis Sci 1 9 6 8 , 7: 44-5 2 .

7R Geroski D H , Kies J C , Edelhauser H F : T h e effects of o u a ­ bain o n endothelial function i n h u m a n and rabbit cor­ nea . Curr Eye Res 1 9 8 4 , 3 : 33 1-8 .

7 9 Fischbarg J , Hernandez J , Liebovitch LS , et a l : The mech­ anism of fluid and electrolyte transport across corneal endothelium . Curr Eye Res 1 985 , 4: 3 5 1 -60.

80 Liebovitch LS and Fischbarg J: Effects of i n hibitors of passive N a + and HC03- fluxes on electrical potential and fluid transport across rabbit corneal endothelium . Curr Eye Res 1 982 , 2: 1 83-6 . 81 H uff JW and Green K: Characterisation of bicarbonate ,

sodium and chloride fluxes in the rabbit corneal endo­ thelium . Exp Eye Res 1 983 , 36: 607- 1 6 .

Clinical

specular microscopy. I. Optical Principl e s . A rch Oph­ thalmol 1 9 7 9 , 97: 1 7 14- 1 9 .

9 4 Sherrard ES and Buckley RJ : Clinical specular m icro­

scopy of the corneal endothelium . Trans Ophthalmol Soc UK 1 9 8 1 , 1 0 1 : 1 56--6 2 .

" Sherrard ES and B u c k l e y R J : E n d o t h e l i a l w r i n k l i n g . In

Trevor-Roper P e d . Tra n s V I Cong Europ Soc Ophthal­ mol . London : Roy Soc Med 1980: 69 .

" Waring G O , Bourne WB , Edelh auser H F, el a l : The cor­ neal endothelium . Ophthalmology 1 9 8 2 , 89: 5 3 1-90 .

97 Rao G N , Lolhman L E , Aquavella JV:

Cell size-cell

relationships i n corneal endothelium . In vest Ophthal­

mol Vis Sci 1 9 8 2 , 22: 271-4 . 98 Matsuda M , Suda T , Man abe R : Serial alterations i n endothelial cell shape a n d pattern after intraocul ar sur­

" Whikhart D R and S o p p e t D R : Activities of transport enzymes located i n the plasma membranes of corneal

Edelhauser H F, Van Horn D L , M i l l e r P, et al: Effect of

endothelium . J Cell Bioi 1976, 68: 5 67-7 8 . 9 1 M i s h i m a S : Clinical investigations of t h e corneal endo­

of fluid transport across rabbit endothe lium . J Mem Bioi 1 97 7 , 35: 95-1 1 2 .

calcium-free

1 9 6 8 , 7 : 53-66 .

�I

active transendothelial bicarbonate flux in the rabbit

corn e a . Biochim Biophys A cta 1 97 8 , 514: 286-9 3 .

with

n e a . V I I . Effects of perfusion with a C a + + free medium

bonic anhydrase in fluid transport across rabbit corneal

72 Mayes KR and Hodson S: Local osmotic coupling to the

endothelium

on the corneal endotheli u m . In vest Oph thalmol Vis Sci

7 0 D i kstein S and Maurice D M : The metabolic basis to the fluid pump i n the cornea . J PhysioI 1 97 2 , 22 1 : 29-4 1 .

corneal

8 9 Kaye G I , Mishima S , Cole J D , e t a l : Studies o n the cor­

6 9 Hodson S : Evidence for a bicarbonate-depe ndent sodium

pump i n corneal endothelium . Exp Eye Res 1 97 1 , I I :

rabbit

medium . In vest Oph thalmol Vis Sci 1 973 , 12: 98-1 1 3 .

drase inhibitors o n the permeability a n d fluxes and cor­ neal thickness . In vest Ophthalmol Vis Sci 1 9 7 7 , 1 6 :

63:

Ophthalmology 1 985 , 92: 759-63 .

bicarbonate transport and the effect of carbonic anhy­

883-92 .

1 98 5 ,

85 McCartney M D , Wood TO , McLaughlin B J : I m m unohis­

the human c o r n e a i n vitro . Exp Eye R e s 1 98 1 , 32: 67 Hull D S , Green K, Boyd M , et al: Corneal endothelium

Ophthalm o l

Pump function of the human c o r n e a l e ndothelium .

66 H o d s o n S , Wigham C , Williams L , et a l : Observations o n

35 3-60 .

A cta

" Geroski D H , Matsuda M , Yee RW, Edelhauser H F :

human corneas i n vivo . In vest Ophthalmol vis Sci 1 9 8 1 , 20: 1 23-5 .

endothe lium .

1 1 6--2 5 .

gery. Am J Oph thalmol 1 984 , 98: 3 1 3- 1 9 .

9 9 Hirst LW: Quantitative analysis o f wide-field endothelial

specular photomicrographs . · A m J Ophthalmol 1984,

97: 488-95 . 1 lJ(] Holladay JT, Bishop J E , Prager TC: Quantitative endo­ thelial

lUI

biomicroscopy.

Ophthalmic

Surg

1 983 ,

14:

33-4 0 .

H e d b y s B O and M i s h i m a S : T h e thickness-hydration

relationship of the cornea . Exp Eye Res 1 96 6 , 5 : 221-8. 1 02 Wood WJ and Maumenee A E : Corn e a l thickness after cataract surge ry. Trans A m A cad Ophthalmol Otolar­

yngo1 1 975 , 79: 63\-4. l U3 Olsen T: Corneal thickness and endothelial damage after intracapsular

cataract

extraction .

A cta

Ophthalmol

1 9 8 0 , 58: 424-3 3 . 104 Cheng H, B ates AK, Wood L , McPherson K : Positive

corre lation of e n dothelial thickness and endothelial cell

loss. A rch Oph thalmol 1 988 , 106: 920--2 . 1 05 Mandell R B , PoIse K A , Fatt I: Corneal swelling caused by contact lens wear. A rch Oph thalmol 1 97 0 , 83: 3-9 .

417

THE CORNEAL E NDOTHELIUM

1 06 Klyce S O : Stromal lactate accumulation c a n account for corneal oedema osmotically following epithelial hypo­ xia i n the rabb i t . J Physiol 1 9 8 1 , 32 1 : 49-6 4 .

1 07 Weissman BA, Fatt I, Horn B: Reversal of anoxic corneal swelling by breathing oxyge n . In vest Oph thalmol Vis 1 06

and cell shape . In vest Oph thalmol Vis Sci 1 977 , 16: 2

1 0 8 1 -4 .

! 3 Gordon SR: The localisation of action i n dividing corneal endothelial cells demonstrated with nitrobenzoxadia­

zoIc phalladi n . Cell Tissue Res 1 98 3 , 229 : 5 33-9 .

Sci 1 9 82 , 22: 1 24-7 . Azen SP, B urg K A , Smith RE et al: A comparison of three methods for the measurement of corneal thick­ n e s s . In vest Ophthalmol Vis Sci 1 979 , 1 8 : 5 3 5-8 .

1 09 O'Neal M R and Poise KA: I n vivo assessment of mech­ anisms controlling corneal hydration . In vest Ophthal­ mol Vis Sci 1 985 , 26: 849-5 6 .

1 1 0 O'Neal M R and Poise K A : Decreased endothelial pump function with agein g . In vest Ophthalmol Vis Sci 1 9 8 6 , III

1 3 1 Gipson IK: Cytoplasmic filaments: their role i n motility

27: 457-63 . Chaston J a n d Fatt I : Corneal oedema i n phakic and

1 33 Fuj i n o Y and Tanishima T: Actin in wound h e a l i n g of rab­ bit corneal endothelium . Jap J Oph thalmol 1 98 7 , 3 1 : 393-40 4 . 1 34 Goldminz D , V l o d a v s k y I , Johnson L K , et al: Contact inhibition and the regulation of e ndocytosis i n the cor­ neal endothelium . Exp Eye Res 1 97 9 , 29: 3 3 1 -5 1 . i35 Hirsch M , Renard G , Faure J P, e t al: Form ation o f inter­

cellular spaces and j unctions in regenerating rabbit cor­

neal endothelium . Exp Eye Res 1 97 6 , 23: 3 85-97 . 1 3 6 Yano M and Tani s h i m a T: Wound healing in rabbit cor­

aphakic e y e s . Curr Eye Res 1 9 8 1 , 1: 1 1 9-2 1 . 1 1 2 Holden B A , PoIse K A , Fonn D , et al: Effects of cataract

i37 Van Horn DL, Sendele DD, Seideman S, et al: Regener­

1 9 8 2 , 22: 343-5 0 . 1 1 3 Coakes RL and B rubaker R F : Method of measuring

cat . In vest Ophthalmol Vis Sci 1 97 7 , 1 6 : 597-6 1 3 . 1 3 6 Matsubara M and Tan ishima T: Wound-healing of the

surgery on corneal function . In vest Oph thalmol Vis Sci

neal endothelium . Jpn J Ophthalmol 1 98 0 , 24: 297-309.

ative capacity of the corneal endothelium in rabbit and

aqueous humour flow and corneal endothelial per­

corneal endothelium i n the m o n k e y : a morphometric

meability using a fluorophotometry nomogra m . Invest

study. Jpn J Ophthalmol 1 9 8 2 , 26: 264-73 . 1 39 Van Horn DL and Hyndiuk R A : Endothelial wound

Ophthalm o l Vis Sci 1 97 9 , 18: 288-302 . 1 1 4 Bourne W M , Nagataki S, Brubaker R F : The permeab­ ility of the corneal endothelium to fluorescein i n the normal human eye . Curr Eye Res 1984, 3 : 509- 1 3 . 1 1 5 Burns RR, Bourne W M , B rubaker R F : Endothelial func­ tion in patients with cornea guttata . In vest Ophthalmol Vis Sci 1 9 8 1 , 20: 77-85 . 1 16 Bourne WM and B rubaker R F : Decreased endothelial permeability i n the iridocorneal endothelial syndrom e . Ophthalmology 1 98 2 , 8 9 : 5 9 1 -5 . 1 1 7 Sawa M, Araie M, Tanishima T: Permeability of the cor­

repair in primate corn e a . Exp Eye Res 1 975 , 2 1 : 1 1 3-24 . 1 40 Honda H, Ogita Y, Higuchi S, et al: Cell movements in a living mammalian tissue : long-term observation of indi­ vidual cells i n wounded corneal endothe l i a of cats. J Morpho l 1 98 2 , 1 7 4 : 25-39. 1 4 1 Matsuda M, Sawa M, Edelhauser HF, et al: Cellular migration and morpholoy i n corneal endothelial wound repair. In vest Oph thalmol Vis Sci 1 985 , 26: 443-9 . 1 42 Kaufman HE, Capella J A , Robbins JE: The human cor­

neal endothelium to fluorescein-A follow-up of ker­

neal endothelium . A m J Ophthalmol 1 9 6 6 , 61: 835-4 1 . 1 43 Rahi A H S and Robins E : Human corneal endothelial cell

permeability i n transplanted cornea s . A m J Ophthalmol

1 9 8 1 , 1 0 1 : 30--4 . 1 44 Simonsen A H , Sorensen K E , Sperling S : Thymidine

atoplasty case s . Jpn J Ophthalmol 1 98 2 , 26: 326-- 3 7 . 1 1 8 Bourne WM and B rubaker R F : D ecreased endothelial 1 983 , 96: 362-7 . 1 1 9 Tsuru T, Araie M , Sawa M, et al: A fluorophotometric study o f the corneal graft . Graefes A rch c/in Exp Oph ­

thalmol 1 985 , 22: 263-5 . 1 20 Ota Y: Endothelial permeability to fluorescein in corneal grafts and bullous keratopathy. Jpn J Ophthalmol 1975 , 1 9 : 286-95 .

repair in health and disease . Trans Ophthalmol Soc UK

incorporation by human corneal endothelium during

organ culture . A cta Ophthalmol 1 9 8 1 , 59: 1 1 0-1 8 . 1 45 Treffers W F : Corneal endothelial wound h e a l i n g .

(Thesi s ) , J a nssen/Print Nij megan 1 9 8 2 . 1 46 Laing RA , Neubauer L , O a k SS, Kayne HL, Leibowitz M : Evidence for mitosis i n the adult corneal endo­

thelium . Ophthalmology 1 9 8 4 , 91: 1 1 29-34 .

1 2 1 Hoffer KJ : Corneal decompensation after endothelium

1 47 Hoffer KJ : Vertical endothelial cell disparity. A m J Oph ­

1 22 Waring GO, Bourne WM , Edelhauser H F, et al: The cor­

thalmol 1 9 7 9 , 87: 344-9 . 1 46 Galin MA, Lin L C , Fetherolf E, et al: Time analysis of

cell cou n t . A m J Oph thalmol 1 97 9 , 87: 252-3 .

neal endothelium . Normal and pathologic structure and

12

12

function . Ophthalmology 1 982 , 89: 5 3 1-90.

3 Geroski D H , Matsuda M, Yee RW, et al: Pump function

corneal endothelial cell density aft e r cataract extrac­

tion . A m J Ophthalmol 1 979 , 88: 93-6.

1 49 Schultz R O , Glasser D B , Matsuda M, et al: Response of

of human corneal endothelium . Ophthalmology 1 985 ,

the corneal endothelium to cataract surgery. A rch Oph ­

92: 759-63 .

thalmol 1 9 8 6 , 104: 1 1 64-9.

4 Doughman OJ, Van Horn D, Rodman WP, et a l : Human corneal endothelial layer repair during organ culture .

1 5 0 Renard G, Pouliquen Y, Hirsch M: Regeneration of the human corneal endothelium . A S E M study. Albrecht

its

Von Graefes A rch Klin Ophthalmol 1 98 1 , 2 1 5 : 3 4 1 -8 . 1 5 1 Laing RA , Neubauer L , Leibowitz HM, et a l : Coales­

lar m icroscopy. Ophthalmology 1 98 0 , 87: 849-60 . 1 2 7 Rao GN, Shaw EL, Arthur E, et a l : Morphologic appear­

1 7 1 2- 1 5 . 1 5 2 Olbert D a n d Daus W: Multinucleated endothelial c e l l s .

A rch Oph thalmol 1 97 6 , 94: 1 7 9 1 - 6 . 1 25 Sherrard ES: The corneal endothelium

in

vivo :

response to mild trauma . Exp Eye Res 1 9 7 6 , 27: 347-5 7 . 1 26 Koester CJ , Roberts CW, Donn A , et al: Wide field specu­

ance of the healing corneal endothelium . A rch Ophthal­

mol 1 978 , 96: 2027-3 0 . 1 28 Silverstein A M , Khodadoust A A , Prendergast R A : Des­ quamation of corneal endothelial c e l l s . In vest Ophthal­ mol Vis Sci 1 982 , 22: 3 5 1-8 . 1 29 Binder RF and B i n d e r H F : Regenerative processes in the endothelium of the corne a . A rch Ophthalmol 1 9 5 7 , 57:

1 1-1 3 . 1 30 Chi H H , Teng C C , Katzin H M : Healing process in the mechanical denudation of the corneal endotheliu m .

Am J Ophthalmol 1 960 , 4 9 : 693-703 .

cence of endothelial cells i n the traumatised corn e a . I I .

Clinical observati o ns . , A rch Ophthalmol 1 983 , 1 0 1 :

Klin Monatsbl A ugenheilkd 1 984 , 1 8 5 : 462-4 . 1 53 Khodadoust AA and Green K. Physiological function of

regenerating endothelium . Invest Ophthalmol Vis Sci

1 97 6 , 1 5 : 96-1 0 1 . 1 54 Tsuru T , Araie M ,

M atsubara M , e t a l : Endothelial

wound-healing of monkey cornea . Jpn J Ophthalmol 1 98 4 , 28: 1 05-25 . 1 55 Minkowski J S , B artels S P, Delori F C , et al: Corneal endo­

thelial function and structure following cryo-injury i n

t h e rabbit . In vest Oph thalmol Vis Sci 1 984 , 2 5 : 1 4 16--2 5 .

1 5 6 Huang PT, Nelson LR, Bourne WM: The morphology

418

S . J . T UFT AND D . J . COSTER

and function of healing cat corneal endothelium . In vest

1 80 Kraff MC, Sanders D R , Lieberman HL: Monitoring fer continuing endothelial cell loss with cataract extractioll

Ophthalmol Vis Sci 1 9 89 , 30: 1 794-80 1 .

1 57 Cogan D G and Kuwabara T: Growth and regenerative

potential of Desce m e t ' s membra n e . Trans Ophthalmol

Soc UK 1 97 1 , 9 1 : 875-9 4 . 1 5 8 Sugar J, Mitchelson J, Kraff M: Endothelial trauma and

and

1 81

intraocular

lens

implantation .

OphthalmolDD

1982 , 89: 30-4 .

Liesegang TJ , Bourne W M , I 1strup O M : Short and term

endothelial

cell

loss

extraction and i ntraocular lens implantation .

cell loss from intraocular lens insertion . A rch Ophthal­

10D,_

associated with cataract

Am I

mol 1 97 8 , 96: 449-5 0 . 1 59 O l s o n T and Erikson J S : Corneal thickness and endo­

Oph thalmol 1 9 8 4 , 97: 32-9 . 1 82 Polack FM and Sugar A: The phacoemulsification

Ophthalmol 1980, 5 8 : 773-86 . 1 60 Rao G N , Acquavel l a J V , Goldberg

' " B i n d e r P S , Sternberg H , Wickman M G , et al: Corneal

thelial dam age after intraocular lens implantation . A cta

Pseudophakic

bullous

keratopathy.

pro­

cedure I I I . Corn e a l complications . Invest Ophthal""" Vis Sci 1 977 , 1 6 : 9-4 6 .

SH,

Bork

SL:

Ophthalmology

1984 , 9 1 : 1 1 35-40.

'" Drews R C : Intermittent touch syndro m e . A rch Ophthal­

endothelial damage associated with phacoemulsifiQ­

tio n . A m J Ophthalmol 1 9 7 6 , 82: 48-5 4 . 1 84 Kaufman H E and Katz J I : Endothelial damage fro. intraocular lens inserti o n . In vest Ophthalmol Vis

mol 1 9 8 2 , 100: 1 440- 1 .

,.2 Miller 0 and Doane M G : High speed photographic eval­

uation of intraocular lens move m e n t s . A m J Ophthal­

damage with intraocular ' le n s e s . Science 1 977, 1911

mol 1 984 , 97: 752-9 .

,.3 Yamaguchi T, Kanai A, Tana k a M, et al: B ullous kera­ topathy after anterior-posterior radial keratotomy for myopia and myopic astigm atism . A m J Oph thalmol 1982, 93: 600-6. 164 Bourne W M and Kaufman HE: Endothelial damage asso­ ciated with intraocular lense s . A m J Oph thalmol 1 97 6 , 8 1 : 482-5 .

165 Forstot S L , Blackwell W L , Jaffe N S , et al: The effect of intraocular lens i m p l a n t ation of the corneal e n dothe­ lium . Trans Am A cad Ophthalmol and Oto1 1 977 , 83:

525-7 . 1 86 Sherrard E S : Intraocular lens dam age to the endothelilllll of the in vitro rabbit corne a . Trans (j'ph ihalmol Soc Ult 18

scopy of intraocular l e n s and entothelial cell interse:.

tion . A m J Ophthalmol 1 97 8 , 86: 1 57-6 1 . 1 88 Apple O J , Mamalis N , Loftfi e l d K, et al: Complications

of in traocular l e n s e s . Surv Ophthalmol 1 9 8 4 , 29: 1-54, 1 8. Thorburn D E and Levenson JE: Corneal e ndothelial damage from previously implanted intraocular lenses.

Am Intraocular Implant Soc J 1 98 0 , 6 : 236-8 .

,.0

ylcellulose .

" 7 Abbott RL and Forster R K : Clinical specular microscopy

1 476-9 . 1 68 Kraff M C , Sanders Dr, Lieberman H L : Specular micro­ scopy i n cataract and intraocular lens pati e n t s . A rch Ophthalmol 1 980 , 98: 1 7 82-4 . ,.9 Hirst LW, Snip R C , Stark WJ , et a l : Quantitative corneal and cataract surgery. A m J Ophthalmol 1 977 , 84:

1 70 Bourne W M , Brubaker RF, O ' Fallon WM : U s e of air to

decrease endothelial cell loss during i ntraocular lens

implantation . A rch Ophthalmol 1 97 9 , 97: 1 473-5 . 1 7 l Hoffer KJ: Effects of extracapsular implant techniques

Ophthalmol

Vis

Sci

1 984,

15;

, . , K i r k S , B u r d e R M , Waltman S R : Minimising corneal endothelial damage due to i n t raoculular lens contact.

In vest Ophthalmol Vis Sci 1 977 , 16: 1 053--Q . 1 92 Cheng H, Law A B , McPhersoll K, et al: Longitudinal study o f i n traocular lens implants after i n tracapsulal

cataract extractio n . Trans Ophthalmol Soc UK 1981, 101: 79-83 . 1 93 Fechner PU and Fechner M U : Methylcellulose and lens

implantation . Br J Ophthalm o l 1 9 8 3 , 67: 259-63 . 1 94 MacRae S M , Edelhauser H F, Hyndiuk R A , et al: The

effects of sodium hyaluro n a t e , chondroitin sulphate

and methyl cellulose on the corneal endothelium and

intraocular

on endothelial density. Am J Ophthalmol 1 98 2 , 100: 791-2.

In vest

1 3 29-3 2 .

endothelial evaluation i n intraocular lens implantation

775-80 .

Hammer ME and B u rch T G : Viscous corneal protection

by sodium hyaluronate , chondroitin sulfate , and meth­

1978 , 8 5 : 394-9 .

and intraocular surgery. A rch Ophthalm o l 1 97 9 , 97:

1983 , 103: 5 65-7 6 .

7 Sugar J, B urnett J, Forstot S L : Scanning electron micro­

195-203 . 166 Sugar A, Fetherolf EC, Lin LLK, et al: Endothelial cell

loss from i ntraocular lens inserti o n . Op hthalmology

ScI

1976 , 1 5 : 996-1000. 1 85 Kaufman HE, Katz J, Valenti J , et al: Corneal endothelial

pressure .

Am

J

Ophthalmol

1983,

95:

Polack FM and Sugar A : The phacoemulsification pro­

332-4 1 . 1 95 Packard R B S , Garner A , Arnott EJ : poly-HEMA a s a

mol Vis Sci 1 97 6 , 1 5 : 458--Q9 . 1 73 Irvine A R , Kratz RP, O ' Donnell J J : Endothelial damage

nary report . Br J Ophthalmol 1 9 8 1 , 65: 5 85-7 . 1 96 Percival P: Protective role of haelon during lens implan­

tation . A rch Opthalmol 1 978 , 96: 1 023-6 . 1 74 Patel J , App l e , H a n s e n SO , et al: Protective effect of the

" 7 B arrett G and Constable IJ: Corneal endothelial loss with

m

cedure II: corneal endothelial change s . Invest Ophthal­

with phacoemulsification and intraocular lens implan­

anterior lens capsule during extracapsular cataract

extractio n . Ophthalmology 1 9 89 , 96: 598-60 2 . 1 75 Binkhorst C D , Nygaard P , Loones LH : Specular micro­ scopy of the corneal endothelium and lens implant sur­

gery. A m J Ophthalmol 1 97 8 , 85: 597-605 . 1 7' Taylor O M , Atlas B F, Romanchuk K G , et al: Pseudo­ phakic bullous keratopathy. Ophthalmology 1 9 83 , 90:

1 9-24 . 1 77 Galin M S , Fetherolf E , Lin L, et al: Experimental catar­ act

surge ry-electron

microscopy.

material for intraocular lens impl a n t ation : a prelimi­

tation . Trans Ophthalmol Soc UK 1 98 1 , 1 0 1 : 77-8. new intraocular l e n s . A m J

Ophthalmol

1 984 ,

98:

1 5 7-65 .

1 98 Farge EJ : Results of p e n e trating keratoplasty over a four­ year period . Ophthalmology 1 978 , 85 : 650-3 . 1 99 Olson RJ , Waltman S R , Mattingley TP, Kaufman HE: Visual

results

after

penetrating

keratoplasty

for

aphakic bullous keratopathy and Fuch s ' s dystrophy.

Am J Ophthalmol 1 979 , 88: 1 000-4 . 2IK' Waring GO, Welch SN, Cavanagh HD, et al: Results of penetrating keratoplasty in 1 23 eyes with pseudophakic and aphakic corneal oede m a . Ophthalmology 1 983 , 90:

Ophthalmology

1 979 , 86: 608-20. 1 7 8 Sugar J, Mitchelson J, Kraff M : The effects of phacoe­

2

Ophthalmol 1978 , 96: 446-8. 1 7' Rao G N , Stevens R E , Harris

1980, 90: 677-8 1 . 1 2 12 Waltman S R : Penetrating keratoplasty for pseudophakic

mulsification o n corneal endothelial cell density. A rch JK,

et al:

Long-term

changes i n corneal endothelium following in traocular lens implantatioi n . Ophthalmology 1 98 1 , 88: 386-97 .

11 1

25-3 3 . Meyer R

and

Sugar

A:

Penetrating keratoplasty in

pseudophakic bullous keratopathy. Am J Ophthalmol

bullous

4 1 5-1 6 .

keratopathy.

A rch

Ophthalmol

1 98 1 ,

99:

419

THE CORNEAL ENDOTHELIUM

203 Schanzlin DJ, Robin JB, Gomez DS, et al: Results of pen­

etrating keratoplasty for aphakic and pseudophakic

bullous

keratopathy.

Am

J

Ophthalmol

1984,

98:

302- 1 2 . 204 Kozarsky AM, Stopak S, Waring GO, et al: Results of penetrating

keratoplasty

for

pseudophakic

corneal

oedema with retention of the intraocular lens . Ophthal­ mology 1 9 84 , 9 1 : 1 1 4 1 -6 . U 2 5 Sugar A, Meyer RF, H e i d e m a n D, et al: Specular micro­ scopic follow-up of corneal grafts for pseudophakic bullous keratopathy. Oph thalmology 1985 , 92 : 325-3 0 .

21)6 Waring G O , S t u l t i n g R D , S t r e e t D : P e n e t r a t i n g k e r a ­

argon laser iridotomy. A m J Oph thalmol 1 982 , 98: 1 53-6 . 226 Thoming C, Van Buskirk E M , Samples J R : The corneal

endothelium after laser therapy for glaucoma . A m J

Ophthalmol 1 987 , 103: 5 1 8-22 . 227 Hirst LW, Robin A L , Sherman S , Green WR, D ' Anna S ,

D u nkelberger G : Corneal endothelial changes after

Argon laser iridotomy a n d panretinal photocoagula­ tion . A m J Ophthalmol 1 982 , 93: 473-8 1 .

m Meyer KT, Pettit TH , Straatsma B R : Corneal endothelial

damage with Neodymium : YAG l a s e r . Ophthalmology 1 9 8 4 , 9 1 : 1 022-8 .

with

m Richardson T M , Brown S V , Thomas JV: Shock-wave

Corneal

mology 1 9 85 , 92: 1 387-95 . U 23 Kerr Muir M G and Sherrard E S : D a m age to the corneal

2'" Gonnering R , Edelhauser H F, Van Horn D L , Durant W:

Oph thalmol 1985 , 69: 77-85 . 2J \ Marshall J , Trokel S, Rothery S, Krueger R: An ultra­

toplasty

for

pseudophakic

corneal

oedema

e x c h a n g e of intraocular lens . A rch Oph thalmol 1 98 7 , 1 0 5 : 58-62 .

207 Leibowitz

HM,

Laing

RA,

Chang R , et al:

oedema secondary to vitreocorneal contact . A rch Oph ­ thalmol 1 98 1 , 99: 4 1 7-2 1 .

The p H tolerance of rabbit and human corneal endo­ thelium . In vest Ophthalmol Vis Sci 1 97 9 , 18: 373-90 .

2()9 Edelhauser HF, Hanneden A M , Pederson HJ , Van Horn DL: Osmotic tolerance of rabbit and human corneal

effect o n anterior segment structures following experi­ mental Neodymium : YAG laser iridotomi e s . Ophthal­

endothelium during Nd/YAG photodisruption . B r J

structural study of corneal incisions i nduced by an

excimer laser at 1 93 nm. 749-5 8 .

Ophthalmology 1985 , 92:

�

endothelium . Arch Oph thalmol 1 98 1 , 99: 1282-7 .

232 Ehlers N: G raft thickness a ter penetrating keratoplasty.

neal oedema and the i n traocular use of epinephrin e .

233 Kimura C and Tanishima T: Thickness of the corneal graft

2 10 Edelhauser HF, Hyndiuk RA, Zeeb A, Schultz RO: Cor­ A m J Oph thalmol 1982 , 9 3 : 327-3 3 .

2 1 1 Hull D S , Chemotti MT, Edelhauser H F, Van Horn D L , Hyndiuk R A : Effect of epinephrine o n t h e corneal

endothelium . A m J Ophthalmol 1 97 5 , 79: 245-5 0 .

2 1 2 Van Horn D L , Edelhauser HF, Prodanovich G, Eiferman R , Pederson HJ: Effect of the ophthalmic preservative

thimerosol on rabbit and human corneal endothelium . Invest Oph thalmol Vis Sci 1 9 7 7 , 1 6 : 273-80 .

2 1 3 Wei n reb RN, Wood J, Tomazzoli J, Alvarado J: Subcon­

junctival injections . Preservative-related changes i n the corneal endothelium . In vest Oph thalmol Vis Sci 1 98 6 , 27: 525-3 1 .

2 1 4 Cohen KL, Van Horn DL, Edelhauser HF, et a l : Effect of phenylephrine on normal and regenerated endothelial

cells i n cat corn e a . In vest Ophthalmol Vis Sci 1 979 , 1 8 : 242-9 .

2 1 5 Vaughn E D , Hull D S , Green K: Effect of intraocular miotics o n corneal endothelium .

A rch

Ophthalmol

1978 , 96: 1 897-901 .

2 1 6 B i rnbaum D B , Hull D S , Green K, Frey N P : Effect of car­ bachol

on

corneal

endothelium .

A rch

Oph thalmol

1 987 , 1 0 5 : 25 3-5 .

2 1 7 Hull D S and Green K : Effect of urokinase o n corneal endothelium . A rch Ophthalmol 1 9 8 2 , 98: 1 285-6 .

2 1 6 Van Horn D L , Edelhauser H F, Aaberg T M , et al: In vivo effects of air and sulfa hexafluoride gas on rabbit cor­

neal endothelium . In vest Ophthalmol Vis Sci 1 97 2 , 1 1 :

1 028-3 6 .

2 1 9 Leibowitz HM, Laing RA, Sandstrom M: Corneal endo­

theliu m : the effect of air i n the anterior chamber. A rch Ophthalmol 1974 , 92: 227-30.

220 Olson RJ : A i r and the corneal endothelium . A rch Oph­ thalmol 1 9 8 0 , 98: 1 283-4 .

22 1 Eiferman R A and Wil k i n s EL: The effect of air o n human corneal

endothelium .

A m J Ophthalmol

1 98 1 ,

92:

328-3 1 .

222 Foulks GN , de J u an E, Hatchell D L , McAdoo T, Hardin

J : The effect of perfluoropentane on the cornea i n rab­ bits and cats . A rch Ophthalmol 1 987 , 1 0 5 : 256-9 .

223 Friberg TR and Doran DL: The effect of vitreous and ret­

inal surgery on central corneal cell density. Ophthal­

mology 1 98 4 , 9 1 : 1 1 66-9 .

224 Sternberg P, Hatchell DL, Foulks GN, Landers MB: The

effect of silicone oil on the corne a . A rch Ophthalmol 1985 , 103: 90-4 .

225 Smith J and Whitted P: Corneal endothelial changes after

A cta Oph thalmol 1 974 , 5 2 : 893-903 .

after penetrating keratoplasty. Jpn J Oph thalmol 1 975 ,

1 9 : 348-5 3 . 234 Greiner J V , Lass J H , G l o n e k T : Metabolic status of fresh versus eye bank processed corne a s . A rch Ophthalmol 1 984 , 102: 1 676-7 . 235 Gottsch J D , Chen C-H , Aguayo I B , et al: G l ycolytic activity i n the human cornea monitored with nuclear

magnetic resonance spectroscopy. A rch Oph thalmol 1 986 , 104: 886-9 . 236 Bourne WM and Kaufman HE: The endothelium of clear corneal

transplants .

A rch

Ophthalmol

1 97 6 ,

94:

1 730-2 . 237 B o urne WM and O'Fallon W M : Endothelial cell loss

during penetrating keratoplasty. A m J Ophthalmol

1 978 , 8 5 : 760-6 . 23 6 Bourne WM: One

year observation o f transplanted

human corneal endothelium . Ophthalmology 1 980 , 87:

673-9 . 239 Rao G N , Stevens R E , Mandelberg A I , et al: Morpho­

logic variations i n graft endothelium . A rch Ophthalmol

1980, 98: 1 403-6 . 240 Culbertson WW, Abbott R L , Forster RK: Endothelial

cell loss in penetrating keratoplasty. Ophthalmology

1 98 2 , 89: 600-4. 24 1 B a h n C F , Meyer R F , M a c C a l l u m D K , et al: Penetrating keratoplasty

in

the

cat.

Ophthalmology

1982,

89:

687-99 .

242 Bourne WM: Morphologic and functional evaluation of the endothelium of transplanted human corn e a s . Trans

A m Ophthalmol Soc 1 98 3 , 8 1 : 403-5 0 . 243 Matsuda M , S u d a T , M a n a b e R : S e r i a l alteration of endo­ thelial cell shape and pattern after i ntra-ocular surgery. A m J Ophthalmol 1 98 4 , 98: 3 1 3- 1 9 .

244 S a t o T : Studies of t h e endothelium of the corneal graft . Jpn J Oph thalmol 1 9 7 8 , 22: 1 1 4-26 . 245 Volker-Dieben HJ . Kok-Van Alphen C C , De Lange P :

Corneal grafts for endothelial decompensation . A cta

Ophthalmol 1 9 8 4 , 62: 436-4 4 . 246 Pollack F M , Smelser GK, Rose J : Long-term survival of isotopical l y labelled stromal and endothelial cells in corneal homograft s . A m J OphthalmoI 1964, 57: 67-7 8 .

2 4 7 Basu P K , M i l l e r I , O r m s b y H L : Sex chromatin as a bio­ logical cell marker i n the study of the fate of corneal

transpl a n t s . A m J Ophthalmol 1 9 6 0 , 49: 5 1 3-1 5 . 246 Olson RJ and Levenson JE: Migration of donor endo­ thelium in keratoplasty. A m J Ophthalmol 1 977 , 84:

7 1 1-14.

420

S. J. TUFT AND D . J. COSTER

249 Bourne WM: I n vivo survival of cryopreserved endo­

class II (Ia) alloantigen expression on corneal endo­

thelial cell s . A rch Ophthalmol 1974, 92: 1 4 6-8.

250 Yan g H J , S a t o T, Matsubara M , et al: Endothelial wound h e aling i n penetrating corneal graft for experimental

bullous keratopathy in rabbit . Jpn J Ophthalmol 1 985 ,

thelium in vivo and in vitro . In vest Oph thalmol Vis Sci 1 985 , 26 : 575-80 .

272 Pepose JS, Gardner KM, Nestor MS, Foos RY, Pettit TH : Detection of H L A class I and II antigens in rejected human corneal all ograft s . Ophthalmology 1 985 , 92:

29: 378-93 .

25 1 Bron AJ and B rown N A P : Endothelium o f the corneal graft . Trans Ophthalmol Soc UK 1974, 94: 863-7 3 .

252 Ruusuvaara P : Endothelial cell densities i n donor and recipient tissue after keratoplasty. A cta Ophthalmol 1 9 8 0 , 5 8 : 267-7 7 .

1480-4 .

271 Dallman MJ and Mason DW: I nduction of l a antigens on murine epidermal cells during rej ection of skin allo­ graft s . Transplantation 1 983 , 36: 222-34 .

274 H a l l B M , Bishop G A , Duggin G G , Hprvath J S , Phillips

253 Rao GN and Aquavella JV: Peripheral recipient e n d o ­

J, Ti ller DJ: I ncreased expression of HLA-Dr antigens

Ophthal­

on renal tubular cells i n renal transpl a n t s : relevance to

254 Abbott R L , Fine M , Guillet E: Long-term changes in cor­

275 Milton A D and Fabre JW: Massive induction of donor­

t h e l i u m following corneal transplantation . mology 1 98 1 , 88: 50-5 .

the rejection respo n s e . Lancet 1 984, 2 : 247-5 1 .

type class I and class II major hi stocompatability com­

neal endothelium fol lowing penetrating keratoplasty.

plex antigens i n rej ecting a l lografts i n the rat . J Exp

Ophthalmology 1 98 3 , 90: 676-85 .

255 Khodadoust A A : The a l lograft rejection: the leading cause of late failure of corneal graft s . I n Porter R,

Med 1 985 , 1 6 1 : 98- 1 1 2 .

27 6 McCulley J P, M a urice DM, Schwarz BD: Corneal endo­

Knight J e d s . Corneal G raft Failure . Ciba Foundation

thelial

Symposium

1 94-20 1 .

15

( n e w series ) .

Amsterda m :

Elsevier

Excerpta Medica North�Holiand 1973: 1 5 1-67.

25 6 Coster DJ: Factors affecting the outcome of corneal

experimental

model .

A rch

The cat as an

Ophthalmol

1979 ,

97:

m Newsome D A , Takasugi M , Kenyon K R , et al: Human

corneal endothelial cells i n vitro : morphology and his­

25 8 Williams K A , White MA, Ash JK, Coster DJ: Leucocytes Analysis by i m m unohistol ogy and actuarial graft sur­

87:

2 1 63-9 .

Dordrecht . D W Junk Publishers 1984 : 1 - 1 7 4 .

i n the graft bed associated with corneal graft fai lure .

1980,

tation of cultured bovine endothelial cells to species with non-regenerative endothelium .

91-7.

non-im munological factors on corneal graft survival .

Ophthalmology

277 Gospodarowicz D , Greenburg G , Alvarado J : Transplan­

transplantatio n . A n n Royal Coli Surg Eng 1 98 1 , 63: 257 Volker-Dieben H J M : The effect of immunological and

transplantation .

tocompatibility ( H L- A ) antigens of pure cell popula­ t i o n s . In vest Ophthalmol Vis Sci 1 974 , 1 3 : 23-3 2 .

279 B a h n C F , Wi nd G , Hufnagel H , et al: C o r n e a l endothelial cell replacement with autologous vascular endothelium

vival . Ophthalmology 1 989 , 96: 38-4 4 .

2 5 9 W i l l i a m s K A , C o s t e r D J : The r o l e of the l i m b u s i n cor­

i n c a t s . In vest Oph thalmol Vis Sci (Supp) 1 9 8 6 , 27: 9 1 .

neal allograft rej e c t i o n . Eye 1 98 9 , 3: 1 5 8-1 6 6 .

23 Couch J M , C u l l e n P , C a s e y TA , Fabre JW: M i t o t i c activ­

ity of corneal endothelial cells i n organ culture with

recombinant human epidermal growth fact o r . Ophthal­

H L A typi n g . Transplant Proc 1 9 8 3 , 15: 1 2 6-9 .

mology 1 9 87 , 94: 1 -6 . r " Bolande R P : The neurocristopathie s . Hum Path 1 974 , 5 :

CS : Expression of HLA:A/B/C and D r antigens on epi­

2R5 Johnston M C : The neural crest i n abnormalities of the

2 65 Fuj i kawa LS , Colvin RB , B a h n AK, Fuller TC , Foster theli a l , stromal and endothelial cells on the human cor­ n e a . Cornea 1 982 , 1: 2 1 3-22 .

266 Vantrappen L, Geboes K, Missotten L, Maudgal P C ,

D e s m e t V : Lymphocytes and langerhans c e l l s in t h e normal h u m a n corn e a . In vest Ophthalmol Vis Sci 1985 , 26: 220-5 .

2 67 Williams KA , Ash JK, Coster D J : Histocompatibility

antigen and passenger cell content of normal and

409-2 9 .

face and brai n . Birth Defects 1 975 , 1 1 : 1-1 8 . 2