In the present study on the rd mouse, there was as in
the RCS rat evidence of progressive axonal dystrophy in the optic fiber
layer and RGC loss. In both species, the axonal dystrophy is associated
with vessels that supply vascular formations at the interface of the
retina and RPE cell layer. There are, however, considerable differences
in the way the vascular disorders develop and in the pattern of RGC
loss in the two animal models.
In the pigmented RCS rat, photoreceptor loss occurs secondary to a
buildup of debris in the subretinal space.
14 30 31 Rods
die over a period of several months in a gradient running from center
to periphery and ventral to dorsal.
32 33 During this
period, there is also a progressive increase in threshold sensitivities
across the visual field recording from the SC and responses to focal
stimulation are largely lost by 6 months
34 (Sauvé Yves, unpublished data, May 1999). Although it is
very difficult to identify photoreceptors at advanced ages, visual
reflexes can still be elicited, even at 1 year or more of
age,
35 and these are presumed to be attributable to the
few remaining cones. The early development of vascular network of the
retina is quite normal, and it is only at approximately 3 months of age
that the first sign of loss of the deep vascular plexus is seen. The
initial changes are the most evident around sites where RPE cells
migrate among the vessels of the deep capillary plexus, creating local
complex vascular formations surrounded by areas devoid of
plexus.
2 3 11 36 37 38 These vascular complexes are
especially prevalent in the ventral retina close to the optic
disc.
9 The vessels from the inner retina cross nerve
bundles before running deep to supply complexes, and as they run deep,
they initially distort the bundles and later pull them into the retina
and ligate them. This is particularly evident close to the ventral
disc. As a result, axons supplying a whole wedge of retina are
affected, and that leads to loss of RGCs from that wedge.
Interestingly, a count made outside these wedges showed normal RGC
numbers, which argues against the presence of general dystrophy.
In the
rd mouse, rod photoreceptor loss is much more
rapid, being largely complete by 3 weeks.
39 40 Accordingly, an ordered map of the visual field on the SC is rapidly
lost
41 (Sauvé unpublished data). In contrast to the
rat, a distinct single layer of cone cell bodies remains over much of
the outer retina, although this gradually disappears with time. It is
reported that vascular complexes develop at points where the cone layer
is deficient.
42 Unlike the rat, there is no concentration
of vascular complexes immediately ventral to the optic disc, and the
overall numbers of foci are considerably less. Instead they are more
broadly dispersed over the retina, with the exception of the upper
temporal retina, where they are less prevalent. This differs from a
recent report on sectioned specimens where complexes were absent from
the whole temporal retina.
43 However, we have found
differences in the progress of change among animals and even between
the two eyes of the same mouse. It is likely also that slight
differences among strains, as well as epigenetic factors, may cause
variance in the development of the vascular anomalies.
Another feature recently reported in
rd mouse
retinas
44 and seen also in this study was a loss of RPE
cells from Bruch’s membrane over local areas of retina. How this may
relate to the migration of RPE cells into the retina or the development
of vascular complexes is not clear.
The first vascular complexes are found, as in the rat, in close
relation to RPE cells, although this is generally harder to see,
because in the rd mouse, these cells have much less
pigment. In contrast to the rat, the first complexes are seen
relatively late, at approximately 6 months. This may be because of the
continued presence of a coherent cone layer, which would delay the
development of a close association between RPE cells and vessels of the
deep plexus. This late onset is in contrast to the overall changes in
retinal vasculature. By the time vascular foci are first seen, the deep
plexus has largely disappeared, and only the deep drainage venules
remain, with short stubby branches. This is in contrast to the RCS rat,
in which the formation of the vascular complexes precedes loss of the
deep plexus and seems to precipitate its loss. This difference in
timing is also likely to affect the configuration of the vessels
contributing to the complexes. They more commonly originate from
vessels of the intermediate and superficial plexus, running
perpendicularly rather than obliquely through the retina. Deep drainage
venules with clusters of pigmented cells and local dilatations (seen in
the RCS rat) are much less common in the rd mouse.
Despite a somewhat different pattern of development of the vascular
foci in rat and mouse, the vessels serving them are involved in the
disruption of optic axons and loss of ganglion cells. The distribution
of axonal disruption correlates closely with the distribution of
vascular foci and is particularly prevalent close to the optic disc in
rats but more widely distributed in mice. In contrast to RCS rats,
which show sector loss of RGCs, but normal RGC densities outside the
sectors, rd mice show only local loss of RGCs correlated
with vessel traction, but reduced numbers of RGCs over the whole
retina.
The possibility of ganglion cell loss occurring in
rd mice was previously raised in a study showing loss of cells in the
ganglion cell layer in older mice.
45 In this case it was
attributed to transneuronal atrophy associated with the very early
developmental loss of photoreceptors in the first 3 weeks of life. The
results from the present study, when compared with those in our
previous studies on RCS rats, suggest again that some cell death can be
correlated with axonal dystrophy and retrograde degeneration caused by
vascular changes. However, because of the broader distribution of cell
loss, it is not possible to exclude a transneuronal effect, although
the late onset and slow progress argue against this. Another
possibility is that the substantial overall reduction in the vascular
plexus may be insufficient to support all the remaining cells of the
retina, leading to their degeneration. Previous works in albino RCS rat
has noted that there is loss of inner nuclear layer
cells,
46 but whether this is due to light damage effects
or to the RCS mutation alone is not clear. In further studies (Shaomel
Wang, unpublished data, July 1997), we have found that with
time there is reduction in the number of NADPH-diaphorase positive
amacrine cells in
rd mice suggesting that there may
indeed be generalized neuronal cell loss.
This poses the question of which mechanism may play a role in the loss
of RGCs over time in patients with RP
17 18 19 47 and which
animal serves as a better model. Certainly, the RCS rat shows a
histopathology that is very similar to RP including the loss of
photoreceptors, invasion of RPE cells into the inner retina, narrowing
of blood vessels,
21 the presence of vascular
formations,
22 leaky blood vessels,
20 and
progressive loss of retinal ganglion cells.
The rd mouse, although homologous to a form of RP,
differs in the very early loss of rods and of the deep vascular plexus.
Although neither transneuronal atrophy nor generalized vascular loss
appears to play a role in RGC loss in RCS rats, these cannot be
completely ruled out in rd mice, although the late
onset, as in humans, makes transneuronal atrophy unlikely. It appears
therefore that both focal and general RGC loss secondary to vascular
events should be considered to be complications occurring after the
photoreceptor loss that occurs in RP.
The vascular changes have repercussions for repair strategies in the
rd mouse. For example, the rapid loss of the deep
vascular plexus in the
rd mouse presents a potential
problem for transplantation of new photoreceptors. For such transplants
to be stable and functional, the deep plexus may have to be
re-established. Two studies have indicated that such transplants
deteriorate with time,
48 49 and the loss of the plexus
clearly may contribute to this. Furthermore, efficacy of
transplantation at too late a stage–in the
rd mouse
after 7 months of age–could be further compromised by ganglion cell
loss.
The authors thank J. Lawrence for comments and help
regarding this study; Toby Holmes for expertise in producing wholemount
pictures; Anthony S. L. Kwan for help and advice; Niyi
Ademuso for technical assistance; and the Department of Psychology,
University of Sheffield for providing some of the animals.