Postnatal Development of Corneal Endothelium
Charles F. Bahn,* Ronald M. Glassmaaf Donald K. MacCallum4 John H. Lillie4 Roger F. Meyer,§
Barbara J. Robinson,§ and Norman M. RichU
Comparison specular micrographs of infant and adult corneas from cats, cows, dogs, rabbits, and humans
demonstrate that a large decrease in central endothelial cell density occurs during maturation of the
cornea. Central endothelial cell counts of developing cat, dog, and rabbit corneas decrease rapidly during
the first months of life. This rapid decline in endothelial cell density correlates with growth of the cornea
to the adult size. Central endothelial cell counts of adult cat, cow, deer, dog, pig, rabbit, and human
corneas are similar (2500 cells/mm2) despite a wide variation in corneal size. Comparison of observed
endothelial cell counts with two hypothetical situations, one of unrestricted endothelial mitosis and the
other of only endothelial hypertrophy, indicates that hypertrophy of individual cells is primarily responsible for achieving the adult cell density of 2500 cells/mm2 for these species. This observation is
true for species that have a high adult endothelial mitotic capacity (rabbit) as well as those that do not
(cat). The human cornea is a special case because the decline in central endothelial cell density indicates
that a large apparent corneal endothelial cell loss (approximately 45%) occurs early in postnatal development. Invest Ophthalmol Vis Sci 27:44-51, 1986
The corneal endothelium is a neural crest derived
cellular monolayer that lines the posterior cornea and,
by virtue of its pump and barrier capabilities, prevents
corneal swelling.1 Recent studies indicate that the
number of neonatal corneal endothelial cells per unit
area (endothelial cell density) of several mammalian
species (cat, dog, rabbit, human) are higher than previously suspected.2"4 In the cat, a high neonatal endothelial cell density may be necessary for maintaining
a confluent endothelial covering on the posterior cornea
as it grows.3'4 Also, endothelial cells early in development may be less effective at accomplishing stromal
deturgescence than endothelial cells of an adult cornea.3
To further investigate the morphological development
of the mammalian corneal endothelium, we performed
a comparative specular microscopic study of the central
endothelial cell populations of infant and adult corneas
from species that exhibit different postnatal corneal
growth characteristics, adult corneal sizes, and potential
for endothelial cell division. Specifically, we wished to
determine: (1) the relationship between corneal size
and infant and adult endothelial cell densities, and (2)
the degree to which the endothelial cell density changed
during corneal growth in the species studied.
From the Department of Surgery, Divisions of Ophthalmology*
and Vascular Surgery,U Uniformed Services University of the Health
Sciences, Bethesda, Maryland, Department of Ophthalmology,!
Mount Sinai School of Medicine, New York, New York, Departments
of Anatomy and Cell Biology:): and Ophthalmology,§ University of
Michigan, Ann Arbor, Michigan.
The opinions or assertions contained herein are the private views
of the authors and are not to be construed as official or as reflecting
the views of the Department of the Air Force, or the Department of
Defense.
Presented at the 13th Biennial Corneal Research Conference, Boston, Massachusetts, September, 1983.
Supported by fellowships and grants from The National Eye Institute (EY-05576, EYO3573), The Uniformed Services University
of the Health Sciences (R09052), and The National Institute of Dental
Research (DE 02731).
Submitted for publication: March 5, 1985.
Reprint requests: Charles F. Bahn, MD, Major, MC, USAF, Assistant Professor of Surgery (Ophthalmology), Uniformed Services
University of the Health Sciences, 4301 Jones Bridge Road, Bethesda,
MD 20814.
Materials and Methods
Two groups of corneas were evaluated:
Group I. Corneas from outbred infant and adult cats,
cows, deer, dogs, pigs, and humans were examined by
specular microscopy. Alginate impressions were made
of the anterior and posterior surfaces of selected eyes
from which plaster casts were subsequently made to
facilitate surface area calculations (see below). The
number of eyes from each species examined is summarized in Table I. The infant human corneas were
from 3, 6, and 21 days of age.
Group II. The corneas of 6 cats, 7 dogs, and 8 rabbits
44
Downloaded From: http://iovs.arvojournals.org/ on 06/15/2017
45
DEVELOPMENT OF CORNEAL ENDOTHEUUM / Dohn er ol.
No. 1
Table 1. Corneal measurements and estimated endothelial surface areas (mm, mm2)
Species
Cat
Cow
Deer
Age
Corneal radius*
(r)
Corneal height^
(h)
Infant (16)§
(3 wk, 0.50 kg)
4.8 ± 0.5
2.4 ± 0.5
90.4 ± 12
7.8 ± 0 . 2
5.3 ± 0.7
283 ± 34
8.1 ± 1.0
Adult (10)
(3 kg)
Infant (4)
(4 wk, 55 kg)
Adult (11)
(500 kg)
12.4 ± 0 . 7
4.2 ± 0.6
6.1 ±0.7
267.5 ±22.1
613 ±55.3
Adult (20)
(63 kg)
9.9 ± 0.4
6.3 ± 0.7
438 ±31
40.1 ± 3 . 6
Infant (10)
(3 wk, 0.8 kg)
Pig
Rabbit
3.0 ± 0.5
1.9 ±0.3
Adult (4)
(12.5 kg)
7.0
4.5
218 ± 15.3
Adult (4)
(90 kg)
7.4 ± 0.2
4.0
220 ± 0 . 1
4.6 ± 0.3
3.2 ± 0.6
99.6 ± 15
6.9 ± 0.6
5.0 ± 0.5
229 ± 35
5.2 ± 0.3
2.0
5.5 ± 0.6
2.5 ± 0.6
99 ± 11.3
116.3 ±28.9
Infant (15)
(3 wk, 0.25 kg)
Adult (21)
(3.6 kg)
Infant (3)
( 3 , 6 , 2 1 days)
Human
Endothelial surface area%
Mr2 + h2)
Adult (4)
(25 yo)
* r = 1/2 of the corneal diameter at the limbus for spherical corneas (cat,
dog, rabbit). For eliptical corneas (cow, deer, human, pig), the average of the
horizontal and vertical radii was used as r.
t h = height of the cornea from the limbus to the anterior (epithelial) surface.
% Endothelial surface area extrapolated from the calculated anterior (epithelial)
surface area.
§ Number of corneas examined, ages and weights.
from 2 litters of each species were examined sequentially by in vivo specular microscopy from 3 to four
weeks of age (just after the eye lids opened) until the
corneal diameters attained a stable size equal to that
of the maternal parent.
human corneal tissues were obtained from the Michigan Eye Bank.
Anesthesia for in vivo specular microscopy was induced by administering xylazine (1 mg/kg) intramuscularly, followed 10 min later by intramuscular ketamine hydrochloride (5-8 mg/kg). Animal sacrifice
was by an intravenous overdose of pentobarbital.
Animal Care and Procurement of Eyes
Adult cats, dogs, and rabbits were purchased from
licensed distributors. Infant kittens, puppies, and rabbits were delivered by pregnant, laboratory-conditioned
animals and reared in the laboratory. Adult cow, pig,
and deer eyes were obtained from animals killed for
other experiments, and the eyes examined as soon as
possible after death. The general health, housing, and
maintenance of all animals were supervised by the
Laboratory Animal Medicine Units at the University
of Michigan and the Uniformed Services University.
The provisions of the ARVO Resolution on the Use
of Animals in Research were followed. Infant and adult
Downloaded From: http://iovs.arvojournals.org/ on 06/15/2017
Specular Microscopy
Specular microscopy of the central corneal endothelium was performed together with simultaneous
pachymetry with a clinical specular microscope on
those corneas with clear stromas in Group I and on all
corneas in Group II. Corneas in Group I with postmortem stromal swelling that precluded an adequate
view of the endothelium were excised with a thin scleral
rim and examined by in vitro specular microscopy using a laboratory specular microscope.
Central endothelial cell counts were determined by
46
INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / January 1986
between two skilled observers, when 50 randomly selected frames were counted in masked fashion was 3%
of the mean cell density for both adult and infant corneas. The magnifications of the clinical and laboratory
specular microscopes (X76 and X200 respectively)
were verified by photographing a micrometer scale.
Comparison specular micrographs from corneas photographed before and after death by the two instruments gave similar cell counts (not shown).
INFANT
ADULT
Vol. 27
COW
Calculation of Cornea! Endothelial Surface Area and
Endothelial Cell Populations
DEER
PIG
•
\
DOG
Corneal surface areas were calculated from surface
measurements made on living animals and from plaster
models of the anterior eye made from alginate impression material (Caulk Jeltrate; L. D. Caulk Co, Milford
Delaware) using measuring calipers. Similar measurements were made from casts of the posterior corneas
in Group I. Corneal surface areas were calculated using
the formula Area = w(r2 + h2), where r is half of the
corneal diameter at the limbus and h is the height of
the cornea from the limbus to the anterior (epithelial)
surface (5). For elliptical corneas (cow, deer, pig and
human), the average of the horizontal and vertical radii
was used as r.* The number of endothelial cells per
cornea was estimated by multiplying the calculated
comeal endothelial surface area by the central endothelial cell density (determined by specular microscopy).
Results
RABBIT
HUMAN
Fig. 1. Casts of eyes in this study that emphasize differences in
sizes and shapes among mammalian corneas. The corneas are indicated by the blackened areas, all photographed at the same magnification.
counting at least three frames per specimen. Cell counts
were made by an experienced observer who did not
know the source of the photographs. The difference
Downloaded From: http://iovs.arvojournals.org/ on 06/15/2017
Corneal measurements and surface area calculations
used in this study are summarized in Table I. No differences were observed that could be attributed to
measurements taken from fresh corneas or casts. Measurements taken from the anterior corneal surface or
posterior corneal surface also provided similar results,
* A simple spherical expression (Area = ir([r2 + h2]) is probably
accurate for nearly spherical corneas (cats, dogs, and rabbits). This
method may yield relatively large errors (up to 23%) when applied
to more ovoid corneas (cattle, deer, pigs, and human), and the formula
that considers both the horizontal and the vertical radii is more accurate. Corneal surface areas of the most elliptical corneas in this
study (bovine) were calculated using several formulas for ellipsoidal
surfaces. These calculations indicated that the simple method reported
in our methods overestimates the bovine anterior corneal surface
area by 8% (550 mm2 compared to 613 mm2). Also, although anterior
and posterior corneal surface areas are similar, determination of the
(human) posterior corneal surface area from anterior corneal surface
measurements results in an underestimation that is also about 8%.6
Because these two inaccuracies are small, and offset one another, we
believe that our calculations provide values that are reasonably accurate for the purposes intended in our study.
DEVELOPMENT OF CORNEAL ENDOTHEUUM / Dohn er ol.
No. 1
10.000
47
oDOG
9500
9000
8500
8000
7500
7000
6500
6000
5500
"
-
0 INFANT
• ADULT
HUMAN
5000
o
4500
COW
4000
3500
« RABBIT
3000
DEER
CAT
2500
HUMAN
DOG
COW
PIG
2000
1500
1000
500
100
200
300
400
500
600
70
2
ENDOTHELIAL SURFACE AREA (mm )
10
12
14
16
AGE (weeks)
18
20
22
ADULT
(parent)
Fig. 2. Mean central endothelial cell densities of infant and adult corneas in this study (left), and sequential endothelial cell densities (right)
of cats (N = 6), dogs (N = 7), and rabbits (N = 8) studied. Error bars represent ± SD.
a finding consistent with the similarity of anterior and
posterior corneal surface areas.6 Our calculations for
anterior and posterior corneal surface areas are consistent withfiguresreported by others for rabbit7"9 and
human610 corneas. Typical casts of some of the eyes
used in this study that emphasize differences in corneal
size and shape among the species examined are shown
in Figure 1.
Central endothelial cell densities from infant and
adult corneas of different sizes are summarized in Figure 2. The endothelial cell densities of adult mammalian corneas clustered about 2500 cells/mm2 regardless of corneal size. Infant corneas exhibited densities that varied inversely with corneal size.
Sequential central endothelial cell counts from developing cat, dog, and rabbit corneas are summarized
in Figure 2. For these species, a large decrease in endothelial cell density was observed during the first
weeks of life. Figure 3 shows that this decline correlated
with corneal growth.
Specular micrographs of infant corneas exhibited
endothelial cells that were smaller and subjectively
more pleomorphic than endothelial cells from adult
Downloaded From: http://iovs.arvojournals.org/ on 06/15/2017
corneas. Sequential examinations of cat, dog, and rabbit corneas during development demonstrated a subjective increase in the regularity of the mosaic. Image
analysis to quantify changes in the regularity (pleomorphism) of the endothelial mosaic during development is in progress in our laboratory. Representative
specular micrographs that demonstrate the decreasing
endothelial cell density of the cell monolayer during
development of one dog cornea are shown in Figure 4.
The relative contributions of hypertrophy (enlargement) and mitosis to the establishment of an adult endothelial cell population can be estimated by comparing the observed endothelial cell density in infants and
adults to two hypothetical situations where: (1) mitosis
continues after birth so that the final endothelial cell
density remains equal to that of the infant cornea, and
(2) no further division takes place after birth and the
final endothelial cell density results exclusively from
cellular hypertrophy or spreading.4 This comparison
can also be made using the number of cells/cornea, a
value calculated from cell density and corneal surface
area measurements (see Figs. 5, 6)." The latter graphic
INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / January 1986
Vol. 27
representation best describes the human cornea where
a decline in endothelial cell density is out of proportion
to the corneal growth that occurs. Expression of the
data in this fashion permits comparison of the slopes
of the observed data, to the slopes derived from the
hypothetical models of mitosis and hypertrophy among
corneas of different species (Figs. 5, 6). Also, the contribution of hypertrophy for the non-human corneal
endothelium in this study can be estimated by expressing the angle formed by the hypothetical mitosis model
RABBIT
2
4
6
8
8
10 12 14 16
AGE (weeks)
18
20 22
ADULT
<Pafen1>
10 12 14 16 18 20 22 ADULT
AGE (weeks)
(P arent >
13-
DOG!
Fig. 4. Sequential specular micrographs from one developing dog
eye that demonstrate the decreasing endothelial density with increasing age. (A) Three weeks of age, mean cell count = 9176 cells/mm2,
corneal thickness = 0.48 mm (original magnification X76). (B) Four
weeks of age, mean cell count = 7812 cells/mm2, corneal thickness
= 0.46 mm (original magnification X76). (C) Eight weeks of age,
mean cell count = 5828 cells/mm2, corneal thickness = 0.47 mm
(original magnification X76). (D) Twelve weeks of age, mean cell
count = 4092 cells/mm2 corneal thickness = 0.53 mm (original
magnification X76). (E) Twenty weeks of age, mean cell count = 3348
cells/mm2, corneal thickness = 0.56 mm (original magnification X76).
12 11
and the observed data (/3) as a percentage of the angle
formed by the mitosis and hypertrophy model (8) (see
Figure 5, Table 2).* The contribution of mitosis for
109
8
7
RABBIT
2
6
B
10 12 14 16 18 20 22 ADULT
AGE (weeks)
(P arent >
Fig. 3. Sequential average animal weights, corneal diameters, and
cornea] thicknesses of cats, dogs, and rabbits studied (same animals
as Fig. 2, right). Error bars represent ±SD.
Downloaded From: http://iovs.arvojournals.org/ on 06/15/2017
* The estimated contributions of hypertrophy and mitosis for the
non-human corneas in this study are based on the assumption that
no endothelial cells are lost from the posterior cornea during postnatal
development. Also, simple comparison of the angles a and 0 to 5
only estimates the contributions of hypertrophy and mitosis. As
pointed out in a review of this article prior to acceptance, the contributions of hypertrophy and mitosis are more accurately deTangent 5 - Tangent a
(100) and
scribed by the equations
Tangent 6
Tangent a
(100) respectively. Recalculation of our data using these
Tangent 5
equations indicates that the simpler method we report underestimates
the contribution by hypertrophy and overestimates the contribution
by mitosis. The values for mitosis using the latter method are as
follows: cat = 12%, cow = 16%, dog = 7%, rabbit = 13%. We are
most grateful for this review and formula.
49
DEVELOPMENT OF CORNEAL. ENDOTHEUUM / Dahn er ol.
No. 1
the non-human corneal endothelium can be similarly
estimated by expressing the angle a as a percentage of
angle 5 (see Fig. 5). The same relationships between
corneal size and endothelial density were observed
when values obtained by other calculations, or estimates derived by other investigators6"10 were used.
Observed endothelial cell counts, increments of endothelial surface area growth, and hypothetical mitosis
and hypertrophy models are summarized in Figures 5
and 6. Wide variations in postnatal corneal growth were
observed that ranged from a 15% increase for the human cornea to a 543% increase for dogs. Except for
the human, species that exhibited the greatest corneal
enlargement also exhibited the largest decline in cell
numbers/area. Comparison of observed endothelial cell
populations to hypothetical models of mitosis and hy-
Table 2. Estimated contributions of endothelial
mitosis and hypertrophy during non-human
corneal development*
Species
Percentage of
hypertrophy^
Percentage of
mitosisX
Cat
Cow
Dog
Rabbit
82
80
86
84
18
20
14
16
* Mitosis that might occur in the human corneal endothelium during development is mathematically hidden in this model by the large net endothelial
cell loss that occurs. For the non-human corneas shown, it is assumed that no
cell loss occurs from the monolayer that would falsely elevate the contribution
by hypertrophy.
t — — — (See Fig. 5) and footnote on p. 48.
(See Fig. 5).
COW
3000
3000
2800
2800
/
/
2200
2000
2000
PI
1800
1800
£ 1600
1600
I
1400
1400
ju
1200
120(
1000
1000
800
600
400
/
2400
2200
Fig. 5. Endothelial cell
counts are compared to hypothetical models of hypertrophy and mitosis. For the
species indicated, hypertrophy appears to contribute
more to endothelial development than mitosis. In the
mitosis model, cell division
continues after birth so that
the final endothelial cell density remains equal to the infant cornea. In the hypertrophy model, no further cell
division occurs after birth,
and the final endothelial cell
density results from cellular
hypertrophy. Linear slopes
are given in parenthesis.
Comparison of the angles /3
and a to 8 can also be used
to estimate the relative contributions of hypertrophy and
mitosis during development
(see Table 2 and note on
p. 48).
/
/
2400
*
/
2600
2600
HYPERTROPHY (0)
-
800
•
600
•
400
HYPERTROPHY (0)
•
•
200
200
I
100
200
300
400
500
100
600
2
DOG
3000
3000
2800
2800
2600
2600
2400
2400
/*
2000
u
1200
1000
BOO
600
1800
1600
1400
£/i
1200
1000
i
800
600
600
400
500
2000
/
x 1800
a.
|
1600
1400
400
2200
2200
%
300
ENDOTHELIAL SURFACE AREA (mm2
ENDOTHELIAL SURFACE AREA (mm )
,
L_
200
±<-—
400
HYPERTROPHY (0)
200
200
100
200
300
400
500
ENDOTHELIAL SURFACE AREA (mm2
Downloaded From: http://iovs.arvojournals.org/ on 06/15/2017
600
100
200
300
400
500
ENDOTHELIAL SURFACE AREA (mm3)
600
50
INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE/January 1986
3000
2800
MAN
2600
2400
2200
2000
°
2
1800
160
°
z
o
1400
^
1200
UJ
"
1000
HYPERTROPHY (0)
800
OBSERVED
600
(-18464)
400
200
100
200
300
400
500
600
ENOOTHELIAL SURFACE AREA (mm2)
Fig. 6. Endothelial cell counts of infant and adult human corneas
are compared to the hypothetical models of hypertrophy and mitosis.
Unlike the other species examined (see Fig. 5), the observed decline
in endothelial cell density indicates that a net cell loss occurs during
growth of the cornea to the adult size. Linear slopes are given in
parenthesis.
pertrophy indicates that mitosis contributes relatively
little to the postnatal development of the coraeal endothelial cell populations of these species exhibiting a
considerable degree of corneal enlargement. In the case
of the human cornea, not only does the endothelial
cell density decline, but there is also a decrease in the
absolute number of endothelial cells/cornea (Fig. 6).
Discussion
Central endothelial cell densities of the adult mammalian corneas that we examined, regardless of size,
rate of growth, or mitotic capability, congregated
around a density of 2500 cells/mm2. This cell density
may provide an ideal balance between the metabolic
demands of an actively pumping endothelium and the
limited nutrients present in the aqueous humor.
The decline in central corneal endothelial cell density
with age is well documented for several mammalian
species including cats,3 dogs,12 rabbits,13 and humans.1114"17 This decline in endothelial cell density
occurs concomitantly with enlargement of the cornea
from the infantile to adult size, and then continues to
decline at a slower rate throughout the remainder of
life.31217 Sequential examinations of developing eyes
in this study demonstrate that the decline in endothelial
cell density occurs at a rapid rate during early postnatal
development. Differences of age that are as small as 1
Downloaded From: http://iovs.arvojournals.org/ on 06/15/2017
Vol. 27
wk can account for larger differences in endothelial
cell density than previously suspected.112
Our calculation of the number of endothelial cells/
cornea is based on the assumption that the endothelial
cell density in the central cornea is representative of
the cell density in all regions of the cornea in both
infants and adults. We did not perform peripheral
specular microscopy. Previous reports by others, however, indicate a 10% or less difference between central
and peripheral counts for young and old human corneas.1718 This difference between central and peripheral
cell densities would not change the interpretation of
our results.
Central endothelial cell densities declined with age
for all of the species we examined, but the calculated
absolute number of endothelial cells per cornea increased for all species except in the human. Thus some
mitosis must occur in the non-human corneal endothelia during maturation. The human cornea, however,
is unique among those examined in that a decrease in
the absolute endothelial cell number/cornea accompanies maturation of the cornea. If mitosis occurs in
the human corneal endothelium during development,
it is mathematically "hidden" by the large net cell loss
that occurs.
For all of the species that we examined, except the
human, a decline in endothelial cell density more or
less matched the growth of the cornea. Reports by others indicate that a similar pattern of endothelial development may occur in avian19 and rodent20 species.
In human corneas, the increase in surface area is
modest during postnatal development (about 15%), so
that a net endothelial cell loss (about 45%) must occur
to account for the observed adult endothelial cell density. This apparent early postnatal endothelial cell loss
appears to be specific to man among the mammalian
species examined so far. This apparent endothelial cell
loss early in the development of the human cornea
might result in a population of cells with an unusually
restricted regenerative (mitotic) capability.
Key words: cornea, endothelium, specular microscopy, development, cell number
Acknowledgments
The authors gratefully acknowledge the assistance of Mr.
Harold Henry who prepared the casts of the corneas, Douglas
Wilson, PhD who assisted with the corneal surface area calculations, and David Beebe, PhD who critiqued the manuscript.
References
1. Waring GO, Bourne WM, Edelhauser HF, and Kenyon KK:
The corneal endothelium: normal and pathologic structure and
function. Ophthalmology 89:531, 1982.
2. Bahn CF, Meyer RF, MacCallum DK, Lillie JH, Lovett EJ, Sugar
A, and Martonyi CL: Penetrating keratoplasty in the cat: a clinically applicable model. Ophthalmology 89:687, 1982.
3. Bahn CF, MacCallum DK, Pachtman MA, Meyer RF, Martonyi
No. 1
4.
5.
6.
7.
8.
9.
10.
11.
12.
DEVELOPMENT OF CORNEAL ENDOTHELIUM / Bohn er ol.
CL, Lillie JH, and Robinson BJ: Effect of age and keratoplasty
on the postnatal development of feline corneal endothelium.
Cornea 1:233, 1982.
MacCallum DK, Bahn CF, Lillie JH, Meyer RF, and Martonyi
CL: Evidence for corneal endothelial cell hypertrophy during
postnatal growth of the cat cornea. Invest Ophthalmol Vis Sci
24:247, 1983.
Diem K and Lentner K, editors: Documenta Geigy: Scientific
Tables. Basle, Ciba-Geigy Limited, 1970, p. 143.
Kwok LS: Calculation and application of the anterior surface
area of a model human cornea. J Theor Biol 108:295, 1984.
Maurice DM: The permeability of sodium ions of the living rabbit's cornea. J Physiol 122:367, 1951.
Kinoshita S, KJorpes TC, Friend J, and Thoft RA: Limbal epithelium in ocular surface wound healing. Invest Ophthalmol Vis
Sci 23:73, 1982.
Fischbarg J and Lim JJ: Determination of the impedence locus
of rabbit corneal endothelium. Biophys J 13:595, 1973.
Maurice DM: The cornea and sclera. In The Eye. 2nd ed. Vol
I. Vegetative Physiology and Biochemistry, Davson H, editor.
London, Academic Press, 1969, pp. 489-600.
Bahn CF, Falls HF, Varley GA, Meyer RF, Edelhauser HF, and
Bourne WM: Classification of corneal endothelial disorders based
on neural crest origin. Ophthalmology 91:558, 1984.
Gwin RM, Lerner I, Warren JK, and Gum G: Decrease in canine
Downloaded From: http://iovs.arvojournals.org/ on 06/15/2017
13.
14.
15.
16.
17.
18.
19.
20.
51
corneal endothelial cell density and increase in corneal thickness
as functions of age. Invest Ophthalmol Vis Sci 22:267, 1982.
von Sallmann L, Caravaggio LL, and Grimes P: Studies on the
corneal endothelium of the rabbit. Am J Ophthalmol 51:955,
1961.
Hoffer KH and Kraff MC: Normal endothelial cell count range.
Ophthalmology 87:861, 1980.
Hiles DA: Discussion of "Normal endothelial cell count range."
Ophthalmology 87:865, 1980.
Laing RA, Sandstrom MM, Berrospi AR, and Leibowitz HM:
Changes in corneal endothelium as a function of age. Exp Eye
Res 22:587, 1976.
Laule A, Cable MIC, Hoffman CE, and Hanna C: Endothelial
cell population changes of human cornea during life. Arch
Ophthalmol 96:3031, 1978.
Blackwell WL, Gravenstein N, and Kaufman HE: Comparison
of central corneal endothelial cell numbers with peripheral areas.
Am J Ophthalmol 84:473, 1977.
Masterson E, Edelhauser HF, and Van Horn DL: The role of
thyroid hormone in the development of the chick corneal endothelium and epithelium. Invest Ophthalmol Vis Sci 16:105,
1977.
Fitch KL, Nadakavukaren MJ, and Richardson A: Age-related
changes in the corneal endothelium of the rat. Exp Gerontol 17:
179, 1982.