Download Decreased retinal ganglion cell number and misdirected axon

Decreased Retinal Ganglion Cell Number and Misdirected
Axon Growth Associated With Fissure Defects in Bst/ +
Mutant Mice
Dennis S. Rice,* Qing Tang,* Robert W. Williams,* Belinda S. Harris,^
Muriel T. Davisson,\ and Dan Goldowitz*
Purpose. The autosomal semidominant mutation Bst (belly spot and tail) is often associated
with small and atrophic optic nerves in adult mice and shares several important attributes
with heritable optic nerve atrophy in humans. In this article, the authors present adult and
developmental studies on the retinal phenotype in Bst/+ mice.
Methods. Retinal ganglion cells in adult Bst/ + mice were labeled retrogradely with horseradish
peroxidase injected into the right optic tract. Labeled ganglion cells were mapped in wholemounted retinas ipsilateral and contralateral to the injection site. The number of axons in
optic nerves of these and other cases were quantified using an electron microscopic method.
Eyes of neonatal, embryonic day 15 (E15), and embryonic day 12 (E12) Bst/+ mutants were
examined histologically to understand the etiology of the retinal phenotype.
Results. Approximately 60% of adult Bst/ + mice have deficient direct pupillary light responses.
This neurologic phenotype is associated with a reduction in the number of retinal ganglion
cells from the wild-type average of 67,000 to less than 20,000 in Bst/+ mutants. Ganglion
cells with crossed projections are more severely affected than those with uncrossed projections.
Histologic analysis of eyes from E12 mice reveals a delayed closure of the optic fissure. Despite
this abnormality, other ocular structures appear relatively normal. However, some E15 mutants
exhibit marked disorganization of the retinal neuroepithelium, and ganglion cell axons are
found between pigmented and neural retina. At birth, optic nerves of affected mice are
smaller than those of wild-type mice, ectopic axons are found within the eyes, and the ganglion
cell layer contains many dying cells.
Conclusions. The expression of the retinal phenotype in Bst/+ mutants is highly variable—
ranging from a complete absence of ganglion cells to numbers comparable to that in wildtype mice. The reduction in ganglion cell number in affected adult Bst/+ mice is attributable
to the failure of ganglion cell axons to reach the optic nerve head early in development.
Delayed fusion of the fissure is consistently associated with the Bst/+ genotype and probably
contributes to the failure of ganglion cell axons to grow out of the eye. Invest Ophthalmol
Vis Sci. 1997;38:2112-2124.
Xvetinal ganglion cells are among the first cells to
be generated in the vertebrate retina.1"4 Their axons
extend into the thalamus and midbrain and provide
the sole route by which visual information is relayed
to the brain.5"8 A reduction in the number of these
From the * Department, of Anatomy and Neurobiology, College of Medicine,
University of Tennessee, Memphis, Tennessee, and f The Jackson Laboratory, Bar
Harbor, Maine.
Supported by National Institutes of Health grant NS EY095S6 (DC, RWW),
neuroscience training grant NS073233 (DSR), and giant RRO1183 (MTD, BSH).
Submitted for publication November 6, 1996; revised April 11, 1997; accepted May
I, 1997.
Proprietary interest categoty: N.
Reprint requests: Dan Goldowitz, Department of Anatomy and Newobiology, College
of Medicine, University of Tennessee, 855 Monroe Avenue, Memphis, TN 38163.
2112
Downloaded From: http://iovs.arvojournals.org/ on 04/28/2017
cells is often associated with a decrease in visual acuity
and visual field loss.9 Recently, we identified a striking
but variable reduction in the size of the optic nerves in
mice heterozygous for the autosomal, semidominant
mutation, belly spot and tail (Bst). Despite the often
severe reduction in the size of the optic nerve, mice
that carry this mutation have relatively normal appearing eyes.
The Bst gene is located on mouse Chromosome
16 in a region that is conserved on human Chromosome 3.10"12 The most common type of heritable optic
nerve atrophy in humans, dominant optic atrophy or
OPA1 (OMIM No. 165,550), maps to a 2- to 8-centi-
Investigative Ophthalmology & Visual Science, September 1997, Vol. 38, No. 10
Copyright © Association for Research in Vision and Ophthalmology
Retinal Ganglion Cell Abnormalities in Bst/+ Mice
Morgan (cM) interval on chromosome 3.13 15 OPA1
occurs with a frequency of 1:50,000 and like Bst, OPA1
shows a dominant pattern of inheritance. There are
variable and often asymmetric effects on visual acuity
and the ganglion cell population1617, and as with most
Bst/+ cases, eyes of humans with OPA1 are of normal
size and appearance.18'19 Although neither Bst nor
OPA1 has yet been cloned, there is sufficient similarity
in their chromosomal positions and their phenotypes
to raise the possibility that they are mutations in the
same gene. In this article we provide the first description of the Bst retinal phenotype and consider developmental mechanisms that may account for the severe
but variable reduction in retinal ganglion cell number
in this mutant mouse.
METHODS
Animals
Mutant mice used in this study were generated by
mating C57BLKS-£^/ + males or females to C57BLKS+ / + mice. Wild-type mice were either littermates of
C57BLKS-Bst/+ or C57BLKS-+/+ mice from the production colony at The Jackson Laboratory. The direct
pupillary light response was examined by shining a
pen light into the eye of dark-adapted unanesthetized
mice. An abnormal pupillary response was defined as
one that did not constrict to a diameter comparable
to that in wild-type mice (approximately 1 mm). Embryonic mice were obtained by timed pregnancies with
the plug date recorded as embryonic day 0. All mice
were treated in accordance with the ARVO Statement
for the Use of Animals in Ophthalmic and Vision Research.
Histology of Adult, Postnatal Day 0, and
Embryonic Material
Adult and postnatal day 0 (P0) mice were anesthetized
with tribromoethanol (Avertin; 0.2 sec/10 g body
weight) and perfused with 0.9% saline followed by
2.0% paraformaldehyde and 2.5% glutaraldehyde in
0.1 M phosphate buffer (PB; pH = 7.2). Eyes and
nerves were dissected, rinsed in 0.1 M PB with 6%
sucrose, and transferred to cold 2% osmium tetroxide
in a 6% sucrose solution for 2 hours. The tissue was
then washed in 0.1 M PB-saline overnight, stained for
1 hour in 0.5% uranyl acetate, dehydrated, and embedded in Spurr's resin. Tissue was cut in semithin (1
fj,m) and thin (75 nm) transverse sections and
mounted on glass slides and formvar-coated grids, respectively. Thin sections were stained with lead citrate.
Embryonic material was obtained from females
killed by cervical dislocation on days 12 and 15 of
pregnancy. Mutants were distinguished from wild-type
embryos because they were smaller and their tails were
Downloaded From: http://iovs.arvojournals.org/ on 04/28/2017
2113
short and kinked. Embryos were immersion fixed in
3:1 95% ethanol:acetic acid overnight. Embryos were
then transferred to 70% ethanol, embedded in paraffin blocks, cut into 6-//m thick coronal or horizontal
sections, mounted on slides, deparaffinized, stained
with cresyl violet, and covered with a coverslip.
Wholemount Analysis and Quantification of
Horseradish Peroxidase-Labeled Ganglion Cells
Adult mice were anesthetized with tribromoethanol
and positioned in a stereotaxic frame. A craniotomy
was performed over the dorsal right hemisphere.
Next, three unilateral injections (0.1 fi\ each) of a
40% solution of horseradish peroxidase (HRP IV;
Sigma, St. Louis, MO) in 5% dimethylsulfoxide were
directed into the right optic tract and dorsal lateral
geniculate nucleus for 30 minutes. After 24 hours,
mice were deeply anesthetized with tribromoethanol
and perfused transcardially with 3.5% paraformaldehyde and 0.1% glutaraldehyde in 0.1 M PB (pH =
7.2). Optic nerves were processed for ultrastructural
analysis as described above and each eye was dissected
to obtain retinal wholemounts. Wholemounts were
used to study the surface area of the retina and the
topography of peroxidase-labeled ganglion cells as
previously described.20 In most animals, a cut was
made in the dorsal eye before dissection to mark orientation. In other cases, retinal vasculature was used
to determine orientation. Retinas were dissected free
from the pigment epithelium and HRP was visualized
using a modified Hanker-Yates reaction.21 Coverslips
were placed on wholemounts in a polyvinyl alcohol
and glycerol solution and camera lucida drawings of
the wholemounted retinas were made. The surface
area of the retina was determined directly from these
drawings using a graphics tablet.
Retrogradely labeled ganglion cells in wild-type
mice (n = 4) and Bst/+ mice {n = 8) were analyzed. In
every ipsilateral retina, and in two contralateral retinas
from Bst/ + mice with an associated small optic nerve,
all HRP-labeled ganglion cells were plotted with the
aid of a drawing tube attached to a microscope
equipped with DIC optics using a 40X objective. In
contralateral retinas with a normal optic nerve, ganglion cell density was estimated in 8 to 15 samples
(7500 fim2) taken at evenly spaced intervals across
the surface of the retina. The number of HRP-labeled
ganglion cells in each retina was determined by multiplying cell density by the area of the retina.20
Detennining the Number of Axons in the Optic
Nerve
The nerve area was determined by directly measuring
a low-magnification (200 to 300X) montage of the
nerve with a graphics tablet. Mean axon density was
estimated from a set of approximately 25 high-magni-
2114
Investigative Ophthalmology & Visual Science, September 1997, Vol. 38, No. 10
FIGURE l. Ventral view of adult Bst/+ brains illustrating the variability found in the optic
nerve phenotype. (A) The optic nerves (arrows) can be normal, (B) small on one side, (C)
missing on one side, or (D) completely absent.
fixation (12,000 to 15,000X) micrographs taken at
evenly spaced intervals across the nerve cross-section.
The number of axons in each nerve was determined by
multiplying axon density by nerve area. Magnifications
were calibrated with a carbon replica grid (2160 lines/
mm).22 The total number of HRP-labeled ganglion
cells (ipsilateral and contralateral) in each retina was
compared to the number of axons in the optic nerve
Downloaded From: http://iovs.arvojournals.org/ on 04/28/2017
from the same case to determine die success of the
retrograde labeling of ganglion cells.20
RESULTS
Bst/+ mice have a kinky tail, a ventral white spot,
lack of pigmentation on the back feet, and occasional
polydactyly. In addition, they have either normal ap-
Retinal Ganglion Cell Abnormalities in Bst/+ Mice
2115
I'.
•-
9 0 -•
80
••
70
••
5
60
••
= g
50
••
I -
40 ••
30
•
•:••
• •
•'
20 ••
10
••
0
••
• •
1
2
3
FIGURE 2. Adult optic nerves of wild-type mice (A) are larger when compared to Bst/+ mice with
abnormal pupillary light responses (B). (C,D) Ultrastructure of two different sites from the small
optic nerve shown in B. (C) Most axons are comparable to wild-type mice; however, there are
several necrotic axons (arrowheads). (D) Only a few fibers are present (anvtus) and some of these
are necrotic (anvwheads). Glial cell processes are abundant in sites containing few fibers. (E) The
number of ganglion cell axons is shown for diree different groups of mice: 1, wild-type; 2, Bst/
+ mice with normal pupillary light responses; and 3, Bst,/+ mice wdi deficient pupillary light
responses. Each diamond represents the number of fibers in one nerve. The asterisk in group 3
indicates 23 eyes from Bst/+ mice with no optic nerve. Note that in some Bst/+ mice with nomial
pupil responses the number of axons in die optic nerve is less than die lowest wild-type number.
Scale bar = 100 /xm in A,B; = 1 (im in C,D.
pearing optic nerves, one normal and one small nerve,
only one nerve, or no nerves at all (Fig. 1). The Bst
mutation has a dominant pattern of inheritance. However, the percentage of mutant progeny produced in
crosses between Bst/+ and + / + mice is only 33%
(179 mutants in 545 progeny; 104 litters). This percentage is significantly lower than the expected 50%
(X2 = 64.16, P < 0.001). The deficit of mutants is most
likely a result of death caused by severe developmental
abnormalities. Exencephaly is observed in approximately 30% of Bst/+ mice (data not shown).
Downloaded From: http://iovs.arvojournals.org/ on 04/28/2017
Axon Number in Adult Bst/ + Mice
By simple inspection, most eyes of Bst/+ mutants are
indistinguishable from those of wild-type mice. However,
the pupillary light response in many Bst/+ mice is abnormal. The direct pupillary light response is weak or absent
in approximately 60% (93/158) of the mutants examined. Of these 93 affected Bst,/+ mice, about two thirds
(60/93) have a unilateral deficit, the remaining one third
have a bilateral deficit in the pupil response. The abnormal direct pupillary reflex is detectable as soon as the
eyelids open in the second postnatal week.
2116
Investigative Ophthalmology & Visual Science, September 1997, Vol. 38, No. 10
3. The appearance of adult retinas from normal and severely affected Bst/+ mice is
shown in this panel of 1-^m thick, toluidine blue-stained plastic sections. (A) The normal
retina is organized into three obvious nuclear layers: outer (ONL), inner (INL), and ganglion cell layer (GCL). Ganglion cells are large cells in the GCL (arrowheads). (B,C) Sections
through the central retina, near the presumed optic disc area. Note the obvious rosette
(thin arroio) in the ONL in (B). The neural retina is buckled in (C). In both of these cases
there is an ectopia within the sclera of the eye (thick airoius in B,Q. (D) In the periphery,
the organization of the Bst/+ retina appears more normal. Note the severe depletion of
cells in the GCL in Bst/+ cases B to D. (E,F). Comparison of layers in each retina from a
single Bst/-\- mouse with a normal (E) and a unilaterally small optic nerve (F) indicates the
variability in the phenotype observed in Bst/+ mutants. Scale bar = 100 /xm in A to D; =
50 fxm in E,F.
FIGURE
Optic nerves of Bst/+ mice with an abnormal pupillary light response are smaller (Fig. 2B), contain
degenerating fibers (Figs. 2C, 2D), and have a reduction in the total number of ganglion cell axons (Fig.
2E). In 28 Bst/+ cases with an abnormal pupillary
light response, 23 had no optic nerve. The average
number of axons in the five remaining optic nerves
from eyes with an abnormal pupillary response is
10,200 ± 2,800 axons (±SEM). In C57BLKS-+/ +
mice the average is 67,100 ± 2,600 (n = 15). The
axon population in Bst/+ mice with normal pupillary
light responses is also reduced. Although these optic
nerves appear normal, the average number of axons
is only 51,300 ± 5,300 (Fig. 2E, n = 12). This is an
approximately 24% reduction in the axon population
compared with wild-type mice. Furthermore, there is
more variability in ganglion cell number in Bst/ +
Downloaded From: http://iovs.arvojournals.org/ on 04/28/2017
mice with normal pupillary light responses compared
with wild-type mice. The coefficients of variation are
36% and 14%, respectively.
Features of Retinas From Adult Bst/+ Mice
The expression of the retinal phenotype in adult
Bst/+ mice is highly variable. Some Bst/+ mice have
retinas comparable to those in wild-type mice,
whereas, others have a highly abnormal retinal appearance (Fig. 3). In severely affected Bst/+ mice, the
organization of the neural retina is often distorted,
particularly around the optic disc area (Figs. 3B, 3C).
The outer and inner nuclear layers are thrown into
folds, rosettes, or colobomas."^2'1 There are a few remaining cells in the ganglion cell layer, but based on
morphologic criteria these cells do not appear to be
ganglion cells. Consistent with a reduction in retinal
Retinal Ganglion Cell Abnormalities in Bst/+ Mice
2117
TABLE l. Ganglion Cell Number (Ipsilateral and Contralateral) and Axon Number in + / +
(Cases 1-4) and Bst/+ Mice (Cases 5-12)
Ipsilateral
Contralateral
Cells
Case
01-04
(+/+ mice; n = 4)
05
06
07
08
09
10
11
12
Normal
Ectopic
Total
Axons
Cells
Axons*
1482 ± 55
41 ± 4
(n=3)
262
214
117
105
ND
66
19
—
1513 ± 46
71,700 ± 3900
48,800 ± 3200
2900 ±
12,300 ±
13,100 ±
18,900 ±
30,000 ±
63,300 ±
66,600 ±
—
45,200 ± 5100
49,700 ± 4200
52,200 ± 5300
30,100 ± 2800
34,800 ± 4000
1570
41,400 ± 4700
454
68,100 ± 2800
(n = 3)
66,000 ± 4500
59,100 ± 3700
52,800 ± 4300
61,400 ± 4400
69,200 ± 2900
1900 ± 600
64,600 ± 4000
197
566
577
206
804
1465
1145
-t
459
780
694
311
804
1531
1164
—
300
1200
1700
1200
1500
3400
4200
NDJ
ND = not determined.
Values are mean ± SEM.
* The success of the injection can be determined by comparing the ipsilaterally and contralaterally projecting ganglion cell populations
with the total number of axons in the optic nerve.
f Because of spillage of horseradish peroxidase, this case was excluded from the analysis of the ipsilateral population.
X This optic nerve was extremely small and was lost during embedding.
ganglion cells, the inner plexiform layer is also reduced in thickness. In more severely affected eyes, the
thickness of the inner nuclear layer is also reduced
compared with normal retinas (Figs. 3E, 3F). Photoreceptors in the outer nuclear layer are the cell type
least affected in retinas from Bst/+ mice. Inner and
outer segments appear normal except where the retina is detached from the pigment epithelium or
thrown into folds.
The average surface area of the retina in wild-type
mice is 16.7 ± 0.23 mm2 (16 retinas from 8 cases). In
comparison, retinas taken from Bst/+ mice with abnormal pupillary light responses are approximately 10%
smaller—14.8 mm2 ± 0.41 (protected t = 3.5; P < 0.01;
n = 25, including 11 retinas from unilaterally affected
mice and 14 retinas from 7 bilaterally affected cases).
Retinas taken from Bst/+ mice that had normal light
responses have an average area of 16.1 ± 0.32 mm2 (n
= 29). This area is slightly lower than those of wild-type
mice. This observation is consistent with the 24% reduction in retinal ganglion cell number in Bst/+ mice with
normal pupillary light responses.
Retinofugal Pathway in Bst/+ Mice
The ipsilaterally projecting population of ganglion cells
represents 2.1% of the total population of ganglion cells
in the C57BLKS-+/+ retina (Table 1). These cells are
located in the ventrotemporal part of the retina (Fig. 4A).
As expected, retinas from Bst/+ mice with a small optic
nerve ipsilateral to the injection have fewer ganglion cells
than those from wild-type mice (Table 1, cases 05 to 09;
Figs. 4B, 4C, 4D, 4E) or Bst/+ mice with normal, ipsilateral optic nerves (Table 1, cases 10 and 11; Fig. 4F).
However in four of five Bst/+ cases with a small ipsilateral
Downloaded From: http://iovs.arvojournals.org/ on 04/28/2017
optic nerve, the uncrossed population of ganglion cells
represents a larger proportion of the total ganglion cell
population compared with wild-type mice (Fig. 5). For
example, in one of these cases (case 05 in Table 1; Fig.
4B) there are approximately 3000 axons remaining—
merely 5% of the wild-type average. However, in this case
there are 459 ipsilaterally projecting ganglion cells—approximately 30% of the wild-type average. Despite an
overall reduction of ganglion cells, the ipsilateral population is overrepresented in retinas from affected Bst/+
mice.
In retinas from Bst/+ mice with an associated
small optic nerve, there is an increase in ectopic, ipsilaterally projecting ganglion cells compared with wildtype and normal Bst/+ mice (cases 05 to 08 in Table
1; Figs. 4B, 4C, 4D, 4E). These cells are located outside
of the ventrotemporal part of the retina. In wild-type
mice, there are no more than 60 of these ectopic,
ipsilateral cells—less than 5% of the ipsilateral population. In a Bst/+ mouse with a normal ipsilateral but
no contralateral optic nerve, the number of ectopic
ipsilaterally projecting cells is relatively small (about
66 cells—see case 10 in Table 1; Fig. 4F). In Bst/ +
cases with a reduction in the total number of ganglion
cells in the ipsilateral retina, there is a twofold to fivefold increase in the absolute number of ectopic ipsilaterally projecting cells when compared with the wildtype population of cells (Table 1). In one case (06),
these cells represent approximately 27% (214/780) of
the total remaining population of ipsilateral ganglion
cells (Fig. 4C). The morphology of these ectopic, ipsilaterally projecting ganglion cells is similar to ipsilateral cells in the ventrotemporal part of the retina.
Approximately 98% of the ganglion cell popula-
2118
Investigative Ophthalmology 8c Visual Science, September 1997, Vol. 38, No. 10
Features of Postnatal Day 0 Optic Nerves and
Retinas in Bst/+ Mice
Optic nerves of P0 Bst/+ mice have been examined
in cross-section to determine whether the reduction
in ganglion cells seen in adult mutants is related to
their postnatal degeneration or to an embryonic abnormality. At P0 the wild-type nerve is divided into
fascicles of unmyelinated ganglion cell axons by glial
or neuroepithelial cell processes (Fig. 7A). In contrast,
optic nerves of affected P0 Bsl/+ mice are smaller and
contain fewer axons (Fig. 7B). Similar to the variability
observed in the adult Bst/+ retinal phenotype, at P0
the size of the nerve and axon population are variable.
The overall appearance of eyes from affected P0
Bst/+ mice is comparable to that in wild-type mice. However, the morphology of the neural retina is abnormal.
In some cases, the neural retina is buckled, resulting in
its detachment from die pigment epithelium (Figs. 7D,
7E, 7F). At higher magnification (Fig. 8), the ganglion
cell layer is thinner in affected retinas compared with
normal, and diere is a marked increase in the number
of dying cells (Figs. 8B, 8C). Bundles of fibers, presumably
those of ganglion cell axons, are seen within the neuroepithelium and between neural and pigmented retina (asterisks in Figs. 7F, 8D). At the ultrastructural level, these
fibers have a size and appearance that are similar to ganglion cell axons located within the retinal neuroepithelial
endfeet (Figs. 8E, 8F).
FIGURE 4. The topography of ipsilaterally projecting ganglion
cells is shown in camera lucida plots for six retinas. Each dot
represents a single, HRP-labeled cell. Case numbers are indicated to the right of each retina. (A) An example of the ipsilaterally projecting ganglion cell population in wild-type retinas is
shown. Approximately 97% of these cells are located in the
ventral (v) and temporal (t) retina. (B to E) The ipsilaterally
projecting population in four different Bst/+ retinas with a
small optic nerve is shown. Ganglion cell number in these four
cases is reduced by 70% or greater. All four cases have fewer
ipsilaterally projecting ganglion cells compared with wild-type
mice. Many of these cells are located outside of the ventral
and temporal retina in ectopic positions. (F) The ipsiiateral
population in a Bst/+ retina with a normal optic nerve is shown.
Scale bar = 1 mm.
tion projects contralaterally at the optic chiasm in
C57BLKS-+/+ mice. These cells are distributed across
the entire surface of the retina and the density of
contralaterally projecting ganglion cells is higher in
central versus peripheral retina (data not shown). In
retinas of Bst/+ mice with an associated small optic
nerve contralateral to the injection, the number of
ganglion cells is severely reduced compared with normal (Table 1, cases 10 and 12). The surviving contralaterally projecting ganglion cells are located mostly
in the dorsal hemiretina (Fig. 6).
Downloaded From: http://iovs.arvojournals.org/ on 04/28/2017
Features of Eyes From Embryonic Days 15 and
12 Bst/+ Mice
The normal retina at E15 has three prominent layers—a nerve fiber layer, a ganglion cell layer, and a
thick neuroblast layer (Fig. 9A). In eyes from affected
Bst/+ mice these layers are present but are often disrupted. For example, bundles of axons can be seen
to penetrate the retina and collect between the retina
and pigment epithelium (Fig. 9B). Disruption of retinal morphology is variable and severe cases have
marked disorganization, particularly in the ventral retina (Fig. 9C). The overall size of the eye, however, is
comparable to that in wild-type mice.
At E12, the overall appearance of the mutant eye
is similar to that of wild-type mice. Both genotypes
have well-formed lenses and the hyaloid vasculature
appears normal (Figs. 9D, 9E, 9F). In wild-type mice
the optic fissure is closed and the neuroepithelium is
continuous around the cup (Fig. 9D). In contrast, in
the 10 Bst/+ cases that we examined the development
of the optic fissure is delayed. For example, in the eye
from a Bst/+ mouse shown in Fig. 9E, the margins of
the ventral retina are in contact, but fusion of the
margins has not yet occurred. The delay appears more
severe in the eye from a Bsl/+ mouse shown in Fig.
9F. In this case, the margins of the ventral retina are
not yet in contact with each other.
Retinal Ganglion Cell Abnormalities in Bst/+ Mice
16
T •
14
..
12
••
10
••
2119
to the optic disc. Lastly, it is possible that this variation
may simply be the result of developmental noise,22
which consists of small random differences in levels
of gene expression, patterns of cell genesis and migration, or cell death. Stochastic variations in these events
may influence Bst expression and may account for the
often dramatic differences between right and left sides
of the same individual, or affected individuals within
the same litter.
percent
ipsilateral „
6
••
4
•
2
••
10
20
30
60
axon number (x1,000)
70
80
Retinofugal Pathway in Bst/+ Mice With Small
Optic Nerves
Bst/+ mice with abnormally small optic nerves have
fewer retinal ganglion cells compared with wild-type
FIGURE 5. Plot comparing the percentage of ganglion cells
with ipsilateral projections versus the total number of ganglion cells. Each diamond {list/' + ) or circle (wild-type) represents a single animal. In wild-type mice, the average percentage of cells projecting ipsilaterally is 2.1% ± 0.05%
(SEM, n = 4). In contrast, in four of five Bst/+ mice with
a reduced number of ganglion cells, the proportion of ipsilateral cells is higher, 6.4% ± 2.5%, compared with wildtype mice. The lone exception to the trend may be caused
by an incomplete labeling of all ganglion cells.
DISCUSSION
Ganglion Cell Number in Bst/+ Mice
Many adult Bsl/+ mice have a marked reduction in retinal ganglion cell number. In these mice, the direct pupillary light response is weak or absent and one or both
optic nerves are much smaller than normal. In Bst/+
mice with normal pupillary light responses, the average
number of ganglion cells is also less than that in wild-type
mice. Furthermore, variation in ganglion cell number
in Bst/+ mice is much greater than that in coisogenic
nonmutants or in 16 other inbred strains of mice.22 Thus,
ganglion cell number in Bst/+ mice is highly variable,
ranging from a severe reduction in this population of
cells to a more subtle quantitative decrease.
The marked phenotypic variability among Bst/ -V
mutants suggests that additional factors influence the
severity of the mutation. One possible factor that we
can rule out is the segregation of modifier genes. Bst
is maintained on the genetically standardized strain
C57BLKS, therefore, all mutants have an identical genetic background. In addition, all mice are reared in
a relatively uniform laboratory colony, thereby minimizing environmental factors. A more likely explanation for the observed variation in the Bst phenotype
is that it is a product of inherently variable morphogenetic events, such as the temporal sequence of optic
fissure closure along the central-peripheral axis in
the mouse.25 The opportunity for ganglion cell axons
to exit the eye and survive can be differentially affected
by where abnormal fissure fusion occurs in relation
Downloaded From: http://iovs.arvojournals.org/ on 04/28/2017
6. The topography and number of contralaterally
projecting ganglion cells are shown in camera lucida plots
for two Bst/ + mice with a small optic nerve contralateral to
the peroxidase injection. Each dot represents a single, HRPlabeled contralaterally projecting cell. Case numbers are indicated to the right of each retina. In both cases, the total
number of ganglion cells is severely reduced (Table 1, cases
10 and 12). There is a clear decrease in the number of
contralaterally projecting cells, particularly in the ventral (v)
and temporal (t) parts of the retina. Scale bar = 1 mm.
FIGURE
2120
Investigative Ophthalmology & Visual Science, September 1997, Vol. 38, No. 10
FIGURE 7. Light microscopic appearance of optic nerves and eyes from postnatal day 0 (PO)
wild-type and Bst/+ mice in l-/xm thick sections stained with toluidine blue. The optic nerve
{arrow, B) in the Bsl/+ mouse has far fewer ganglion cell axons than the wild-type optic
nerve (A). The other fascicle in B {arrowhead, does not originate at the optic disc. In the
eyes from PO Bst/+ mice (D to F) there is a well-formed neuroepithelium comparable to
that in the normal eye (C) and an obvious ganglion cell layer {arrowheads, D to F; compare
with arrowheads in C). However, the neural retina in the mutant is detached from the
pigment epithelium {curved arrows, D,E). Note that the pigment epithelium {straight arrows,
C) is not involved in the involution in the mutant eye {straight airows, E,F). The ganglion
cell layer appears thicker in E {arrowhead), but this is an artifact resulting from the level of
the section. Scale bar for A,B = 50 jum; C to E = 350 //m; and F - 75 yum.
mice. There are two interesting features of the retinofugal pathway in these mice. First, there is an increase
in the percentage of ganglion cells with ipsilateral projections at the optic chiasm relative to wild-type mice
or Bst/ + mice with normal optic nerves. This increase
is caused, in part, by a substantial population of ipsilaterally projecting ganglion cells located outside of the
ventrotemporal crescent. In wild-type mice the number of these ectopic, ipsilaterally projecting cells does
not exceed 60, but in affected Bst/+ mice there are
often several hundred of these cells. The reason for
the presence of these ipsilaterally projecting cells is
unknown, but there are several possibilities. These ectopic ipsilaterally projecting ganglion cells may be
cells that should have projected contralaterally but
that projected incorrectly into the ipsilateral optic
Downloaded From: http://iovs.arvojournals.org/ on 04/28/2017
tract. It is also possible that these cells have bilateral
projections at the chiasm. Another explanation is that
these ectopic cells may be remnants of a normally
transient population of ipsilaterally projecting cells
that originates from central retina during development.8 The second interesting feature of the retinofugal pathway in Bst/ + mice with small optic nerves is
the topography of contralaterally projecting ganglion
cells. The majority of these cells are located in dorsal
retina. At present we do not yet know why some ganglion cells survive the Bst mutation. It is possible that
the time at which ganglion cells are generated influences the likelihood of survival. This idea is supported
by the fact that during development of the mouse eye,
the ipsilaterally projecting ganglion cells and ganglion
cells located in dorsal retina are among the first to
Retinal Ganglion Cell Abnormalities in Bst/+ Mice
2121
FIGURE 8. Cross-sections through the eye of PO wild-type (A) and three different Bst/+ mice
(B to D) are shown in toluidine blue-stained, l-/xm thick plastic sections. (A) In wild-type
retinas, the GCL is populated by large oval ganglion cells (arrowheads). Fascicles of ganglion
cell axons among the endfeet of neuroepithelial cells can be seen at this magnification {open
arrow). (B) In the eye from this Bst/+ neonate, the neuroepithelium is thinner compared
with wild-type and the GCL is not as obvious. Cells that appear to be ganglion cells are
present (arroiuheads), but many cells in this layer appear to be dead or dying (arrows). (C)
The GCL in the Bst/+ mutant shows a high incidence of dying cells (arrows). (D) The
neuroepithelium in this Bst/+ eye is highly disorganized. Axons (arrowheads) are found in
the middle of the section and pinned between neuroepithelium and pigment epitlielium
(PE; asterisk). The ultrastructural appearance of the axons near the PE (E) is similar to
those among the endfeet (F). Pigment granules are black, dense structures located within
the pigment epithelial cells (asterisk, E). The basal lamina of the retinal epithelium is prominent at the top of the micrograph F (arrows). Scale bar = 50 /xm in A,B>D; 100 //m in C;
and 1 //m in E,F.
be generated.1'3 However, conclusive results can only
come from birth-dating studies.
Developmental Basis of Reduced Ganglion Cell
Numbers in Bst/+ Mice
The normal external appearance of the eye in adult
Bst/+ mutants indicates that early inductive processes,
Downloaded From: http://iovs.arvojournals.org/ on 04/28/2017
those responsible for the formation of the optic vesicle
and lens,26'27 are intact in Bst heterozygotes. Furthermore, the overall appearance of the eye from El2
Bst/+ mice is relatively normal except for the delay
in closure of the optic fissure. The extent of delay is
variable and the fissure is either completely open or
the margins of the optic cup lining the fissure are in
2122
Investigative Ophthalmology & Visual Science, September 1997, Vol. 38, No. 10
FIGURE 9. Paraffin-embedded, cresyl violet-stained sections through eyes of embryonic day
15 (El5; A to C) and embryonic day 12 (E12; D to F) wild-type (A,D) and Bst/+ mice
(B,C,£, and F). (A) Three layers in the neural retina can be seen at E15 in the wild-type
retina—the nerve fiber layer (arroiohead), the GCL layer (asterisk), and the neuroepithelium.
Dividing cells are opposite the pigment epithelium in the neuroepithelium (arrows, A). (B)
Two collections of fibers that originate in the nerve fiber layer are seen between pigmented
and neural retina (arrowheads) in this Bst/+ mutant with a small optic nerve. (C) In this
Bst/+ eye, the neuroepithelium is divided by the optic fissure. Mitotic cells line the fissure
(arrows). Fibers (arrowheads) are located at the crest of the fissure, but they do not exit the
eye. (D) The optic fissure is closed and appears fused (arrowhead) in this E12 wild-type eye.
In contrast, the optic fissure (arrowheads, E,F) is not fused in the Bst/ + eye in E and remains
open in F. Scale bar = approximately 50 /iin.
contact but not yet fused. It should be stressed that
although delayed, fusion of the fissure margins in
Bsl/+ mutants does eventually occur.28
During normal development, the first axons to
grow out of the eye do so in a series of channels
formed by the processes of neuroepithelial cells oriented in the direction of the fissure.3 These channels
are rapidly filled by additional axons, resulting in the
formation of the nerve fiber layer. The cues used by
axons to navigate toward the fissure are located within
the embryonic nerve fiber layer.29"33 A delay in the
closure of the optic fissure may disrupt the temporal
presentation of these cues and result in aberrant
growth of ganglion cell axons in the eye. Alternatively,
the delayed closure of the optic fissure may be secondary to an abnormal interaction between ganglion cells
and their retinal environment. Distinguishing between these possibilities requires the identification of
the cellular target of the Bst gene.28 Despite this caveat,
it is clear that in affected E15 and neonatal Bst/ +
mice, many ganglion cell axons are found outside of
the nerve fiber layer. Thus, the decrease in the size of
the optic nerve and the number of axons in adult
mutants is largely because of the failure of axons to
Downloaded From: http://iovs.arvojournals.org/ on 04/28/2017
enter the nerve during development and consequently fail to reach their target in the brain. This
observation may well explain the high incidence of
cell death observed in the ganglion cell layer of these
mutants. Alternatively, the program of ganglion cell
development may be directly perturbed by the Bst mutation.
The molecules involved in retinal development
are rapidly being identified and characterized.34 Several genes—ChxlO, GH3, Mitf, and Pax6—were first
described as the ocular retardation, extra-toes, microphthalmia, and small eye mutations, respectively.30"'11 Thus, the study of spontaneously occurring
mutants has contributed to the identification of essential genes involved in retinal development. Many of
these genes are putative transcription factors, but their
molecular targets have not yet been identified. Comparison of retinal phenotypes in other mutant mice
may help to identify potential molecular targets of
these genes. In this respect, the retinal phenotype observed in Bst/+ mutant mice closely resembles that
in several strains of mice with mutations in the Pax2
gene.42"44 In these mice, the optic fissure closes abnormally and retinal ganglion cells are die principal cell
2123
Retinal Ganglion Cell Abnormalities in Bst/+ Mice
type affected. In addition to Pax2, two other genes,
ChxlO and Mitf, also affect closure of the optic fissure.
Mutations in these genes, however, are associated with
small and grossly malformed eyes.35'45
7.
Bst Mutation Is a Good Model for Heritable
Ganglion Cell Loss in Humans
8.
Mutations in similar genes are often associated with similar defects in different species.'16 For example, mutation
of the human PAX2gene results in optic nerve coloboma
and kidney defects comparable to that observed in mice
with mutations in the Pax2gene.42'4347 The retinal phenotype in Bst/+ mice is similar to that in autosomal dominant optic atrophy (0PA1) and optic nerve hypoplasia in
humans—both of which are characterized by an early
and variable reduction in retinal ganglion cells in an otherwise normal appearing eye.17'48 Furthermore, the Bst
gene is located on mouse Chromosome 16 in a region
that is conserved in human Chromosome 3—where
0PA1 is located.10"15 The phenotypic similarities combined with the mapping data raise the possibility that Bst
and 0PA1 are mutations in the same gene.
Hoyt and Good49 discuss the underlying developmental events that may lead to human optic nerve atrophy and hypoplasia; however, the basis of ganglion cell
loss remains obscure. In this study we provide evidence
indicating that the reduction of ganglion cells in Bst/+
retinas is related to a developmental abnormality. The
analysis of ganglion cell development in Bst/+ mutant
retinas may provide insights into the variability and asymmetry of ganglion cell loss in humans with heritable optic
neuropathies.
9.
Key Words
10.
11.
12.
13.
14.
15.
16.
17.
18.
genetic disease, mouse, optic nerve, pupil, retina development
Acknowledgments
The authors thank Kathy Troughton and Richard dishing
for expert technical help.
19.
20.
References
1. Siclman R. Hislogenesis of the Mouse Retina Studied With
Tritiated Thymidine. In Smelser GK, ed. The Structure
of the Eye. New York: Academic Press; 1961:487-505.
2. Walsh C, Polley EH. The topography of ganglion cell
production in the cat's retina, f Neurosci. 1985;5:741750.
3. Drager UC. Birth dates of retinal ganglion cells giving
rise to the uncrossed and crossed optic projections in
the mouse. Proc R Soc Lond. 1985;224:57-77.
4. LaVail MM, Rapaport DH, Rakic P. Cytogenesis in
monkey retina. / Cornp Neurol. 1991;309:86-114.
5. Silver J, Sidman RL. A mechanism for the guidance
and topographic patterning of retinal ganglion cell
axons. J Comp Neurol. 1980; 189:101 -111.
6. Shatz CJ, Sretavan DW. Interactions between retinal
Downloaded From: http://iovs.arvojournals.org/ on 04/28/2017
21.
22.
23.
24.
25.
ganglion cells during the development of the mammalian visual system. Ann Rev Neurosci. 1986;9:l7l-207.
Guillery RW, Mason CA, Taylor JSH. Developmental
determinants at the mammalian optic chiasm./Mrurosci. 1995; 15:4727-4737.
Marcus RC, Mason CA. The first retinal axon growth
in the mouse optic chiasm: Axon patterning and the
cellular environment./JVeimwci. 1995; 15:6389-6402.
Quigley HA, Dunkelberger GR, Green WR. Retinal
ganglion cell atrophy correlated with automated perimetry in humans with glaucoma. Am f Ophthalmol.
1989; 107:453-464.
Epstein R, Davisson M, Lehmann K, Akeson EC, Colin
M. Position of Igl-I, md, and Bst loci on chromosome
16 of the mouse. Immunogenetics. 1986;23:78-83.
Rice DS, Williams RW, Ward-Bailey P, et al. Mapping
the Bst mutation on mouse chromosome 16: A model
for human optic atrophy. Mamm Genome. 1995; 6:546548.
Reeves RH, Citron MP. Mouse chromosome 16.
Mamm Genome, (supplement). 1994;5:S229-S237.
Online Mendelian Inheritance in Man: 1996;
URL:http//www.ncbi. nlm.nih.gov
Eiberg H, Kjer P, Rosenberg T. Dominant optic atrophy (OPA1) mapped to chromosome 3q region. I.
Linkage analysis. Hum Mol Genet. 1994;3:977-980.
Lunkes A, Hartung U, Magarino C, et al. Refinement
of the OPA1 gene locus on chromosome 3q28-29 to
a region of 2-8 cM, in one Cuban pedigree with autosomal dominant optic auophy type Kjer. Am J Ham
Genet. 1995;57:968-970.
Eliott D, Traboulsi El, Maumenee IH. Visual prognosis in autosomal dominant optic atrophy. Amj Ophthalmol. 1993; 115:360-367.
Hoyt CS. Autosomal dominant optic atrophy. A spectrum of disabilities. Ophthalmology. 1980;87:245-250.
Johnston PB, Gaster RN, Smith VC, Tripathi, RC. A
clinicopathologic study of autosomal dominant optic
atrophy. Am J Ophthalmol. 1979; 88:868-875.
Kjer P, Jensen OA, Klinken L. Histopathology of eye,
optic nerve and brain in a case of dominant optic
atrophy. Ada Ophthalmol. 1983;61:300-312.
Rice DS, Williams RW, Goldowitz D. Genetic control
of retinal projections in inbred strains of albino mice.
/ Comp Neurol. 1995;354:459-469.
Perry VH, Linden R. Evidence for dendritic competition in the developing retina. Nature. 1982; 297:683685.
Williams RW, Strom RC, Rice DS, Goldowitz D. Genetic and environmental control of variation in retinal
ganglion cell number in mice. / Neurosci. 1996;
16:7193-7205.
Silver J, Puck SM, Albert DM. Development and aging
of the eye in mice with inherited optic nerve aplasia:
Histopathological studies. Exp Eye Res. 1984; 38:257266.
Apple DJ, Rabb MF. Ocular Pathology: Clinical Applications and Self Assessment. 4th ed. St. Louis: Mosby Year
Book; 1991:8-19.
Hero I. Optic fissure closure in the normal cinnamon
2124
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
Investigative Ophthalmology & Visual Science, September 1997, Vol. 38, No. 10
mouse. An ultrastructure study. Invest Ophthalvwl Vis
Sci. 1990;31:197-216.
Pei YF, Rhodin JAG. The prenatal development of the
mouse eye. Anat Rec. 1970; 168:105-126.
Mann I. Development of the Human Eye. New York: Grune
and Stratton; 1964.
Tang Q, Rice DS, Williams RW, Goldowtiz D. Retinal
development in the Bst mouse. Soc Neurosci Abs.
1996;22:1977.
Goldberg S. Unidirectional, bidirectional and random
growth of embryonic optic axons. Exp Eye Res.
1977; 25:399-404.
Halfter W, Deiss S, Schwarz U. The formation of the
axonal pattern in the embryonic avian retina. J Comp
Neurol. 1985; 232:466-480.
Brittis PA, Lemmon V, Rutishauser U, Silver J. Unique
changes of ganglion cell growth cone behavior following cell adhesion molecule perturbations: A timelapse study of the living retina. Mol Cell Neurosci.
1995; 6:433-449.
Halfter W. Intraretinal grafting reveals growth requirements and guidance cues for optic axons in the
developing avian retina. Dev Biol. 1996; 177:160-177.
Brittis PA, SilverJ. Multiple factors govern intraretinal
axon guidance: A time-lapse study. Mol Cell Neurosci.
1995;6:413-432.
Graw J. Genetic aspects of embryonic eye development in vertebrates. Dev Genet. 1996; 18:181-197.
SilverJ, Robb RM. Studies on the development of the
eye cup and optic nerve in normal mice and in mutants with congenital optic nerve aplasia. Dev Biol.
1979;68:175-190.
Burmeister M, Novak J, Liang M-Y, et al. Ocular retardation mouse caused by ChxlO homeobox null allele:
Impaired retinal progenitor proliferation and bipolar
cell differentiation. Nat Genet. 1996; 12:376-384.
Franz T, Besecke A. The development of the eye in
homozygotes of the mouse mutant Extra-toes. Anat
Embryol. 1991; 184:355-361.
Vortkamp A, Franz T, Gessler M, Grzeschik K-H. Deletion of GLJ3 supports the homology of the human
Greig cephalopolysyndactyly syndrome (GCPS) and
Downloaded From: http://iovs.arvojournals.org/ on 04/28/2017
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
the mouse mutant extra toes {Xt). Mavivi Genome.
1992; 3:461-463.
Hodgkinson CA, Moore KJ, Nakayama A, et al. Mutations at the mouse microphthalmia locus are associated with defects in a gene encoding a novel basic
helix-loop-helix-zipper protein. Cell. 1993; 74:395404.
Hogan BLM, Horsburgh G, Cohen J, et al. Small eyes
(Sey): A homozygous lethal mutation on chromosome
2 which affects differentiation of both lens and nasal
placodes in the mouse. / Embryol Exp Morph.
1986;97:95-110.
Hill RE, Favor J, Hogan BL, et al. Mouse small eye
results from mutations in a paired-like homeobox-containinggene. Nature. 1991;354:522-525.
Keller SA, Jones JM, Boyle A, et al. Kidney and retinal
defects (Krd), a transgene-induced mutation with a
deletion of mouse chromosome 19 that includes the
Pax2 locus. Genomics. 1994;23:309-320.
Favor J, Sandulache R, Neuhauser-Klaus A, et al. The
mouse Pax2iNe" mutation is identical to a human PAX2
mutation in a family with renal-coloboma syndrome
and results in developmental defects of the brain, ear,
eye, and kidney. Proc Natl Acad Sci USA. 1996;
93:13870-13875.
Torres M, Gomez-Pardo E, Gruss P. Pax2 contributes
to inner ear patterning and optic nerve trajectory.
Development. 1996; 122:3381-3391.
Hero I. The optic fissure in the normal and microphthalmic mouse. Exp Eye Res. 1989; 49:229-239.
Glaser T, Walton DS, Maas RL. Genomic structure,
evolutionary conservation and aniridia mutations in
the human Pax6gene. Nat Genet. 1992;2:232-238.
Sanyanusin P, Schimmenti LA, McNoe LA, et al. Mutation of the PAX2 gene in a family with optic nerve
colobomas, renal anomalies and vesicoureteral reflux.
Nat Genet. 1995;9:358-364.
Frisen L, Holmegaard L. Spectrum of optic nerve hypoplasia. Br J Ophthalmol. 1978;62:7-15.
Hoyt CS, Good WV. Do we really understand the difference between optic nerve hypoplasia and atrophy.
Eye. 1992:6:201-204.