Recently,
Rs1h mRNA expression has been detected in
photoreceptors, but not in other retinal cells by in situ hybridization
and Northern blot analysis of wild-type and
rd/
rd mouse retinal RNA.
14 These investigators also reported
that an anti-RS1 antibody labeled not only photoreceptor cells, but
also Müller cells,
29 ganglion cells, and other
unidentified cells of the inner retina.
30 On the basis of
these results, they concluded that RS1 is synthesized only in
photoreceptors for export into the extraretinal space. It then migrates
into the inner retina where it associates with the various cells and
structures.
In the present study, we have confirmed that RS1 is abundantly
expressed in rod and cone photoreceptor cells. However, our results
differ from the earlier studies cited in two important respects. First,
our studies indicate that RS1 is expressed in bipolar cells of the
inner retina as well as photoreceptors, although at lower levels.
Second, double-labeling experiments using a variety of cell surface
markers reveal that RS1 is specifically associated with rod and cone
photoreceptors of the outer retina and bipolar cells of the inner
retina, but not Müller cells, ganglion cells, or the inner
limiting membrane.
Several lines of evidence indicate that RS1 is expressed in retinal
bipolar cells. By immunofluorescence microscopy, we detected RS1 on
bipolar cells of all mammalian retinas examined including 1-year-old
rd/
rd mice without rods and most cones, and
rd/
rd cl mice devoid of both rod and cone
photoreceptors. We also observed RS1 expression on bipolar cells
dissociated from retinal tissue and maintained in culture for 4 weeks.
Because RS1 was not detected in these cells after 1 week in culture, we
conclude that RS1 expression in bipolar cells maintained in culture for
4 weeks arises from newly synthesized protein. Finally, although we did
not detect
Rs1h mRNA expression in
rd/
rd mouse retina by Northern blot analysis, in
agreement with the results of Reid et al.,
14 we detected
Rs1 expression using the more sensitive technique of RT-PCR.
Together, these results provide strong evidence that RS1 is expressed
in bipolar cells as well as in photoreceptors.
It is becoming increasingly clear that in situ hybridization and
Northern blot analysis are useful techniques to detect abundantly
expressed mRNA but often do not have the sensitivity or reliability to
measure low-level mRNA expression. For example, the transcription
factor
Crx was initially thought to be expressed only in
photoreceptors of the retina on the basis of a strong in situ
hybridization signal in the photoreceptor layer and no
signal
18 31 or an extremely weak signal
32 in
the inner retina. In the present study, we also were unable to detect
Crx expression in
rd/
rd mice by
Northern blot analysis. However, recent immunocytochemical labeling
studies have established that
Crx is also expressed at low
levels in retinal bipolar cells.
33 The photoreceptor ABC
transporter
ABCA4 (formerly
ABCR) mRNA expression
has been observed in rod, but not cone, cells by in situ
hybridization.
34 Recently, however,
ABCA4 has
been detected in human cone as well as rod cells by immunofluorescence
microscopy and Western blot analysis techniques, a result that has
important implications in the pathogenesis of Stargardt
disease.
35 These and other studies indicate that caution
must be exercised in interpreting cell-specific expression when
relatively insensitive techniques are used. The inability to detect RS1
expression in nonphotoreceptor cells by conventional in situ
hybridization and Northern blot analysis is most likely due to the
limited sensitivity of these techniques.
Immunofluorescence labeling studies using highly specific RS1
antibodies and a variety of cell-specific markers identified the
RS1-immunopositive inner retinal neurons as rod and cone bipolar cells
in all mammalian species examined. RS1 labeling was not detected on
Müller glial cells, ganglion cells, or along the inner limiting
membrane of either peripheral or central retina. These results are in
marked contrast to previous reports showing more widespread retinal
tissue labeling.
29 30 It is unclear why the antibodies
used in our study showed selective staining of photoreceptors and
bipolar cells, whereas an N-terminal antibody used by another group of
investigators
29 30 labeled photoreceptors and most other
cells of the inner retina. However, it is interesting to note that the
antibody used in later studies labeled multiple bands in Western blot
analysis of retinal extracts, leading one to question the specificity
of this antibody.
The
RS1 gene codes for a polypeptide that has primary
structural features characteristic of an extracellular cell adhesion
protein.
9 It has an N-terminal 23-amino-acid hydrophobic
signal sequence with a signal peptidase cleavage site, a signature
characteristic of proteins destined for secretion from cells. It also
contains a discoidin domain found in discoidin I and II, hemocytin,
neuropilin, and other proteins implicated in cell adhesion
processes.
9 10 11 12 13 14 36 37
Biochemical and immunocytochemical experiments performed in this study
support the view that RS1 is a secreted protein complex that interacts
with cell surfaces. RS1 from retinal and COS-1 cell membranes was found
to migrate as a single 24-kDa polypeptide under reducing conditions and
a large (>95-kDa) complex under nonreducing conditions. This indicates
that RS1 subunits assemble into a large, disulfide-linked complex
within the oxidizing environment of the endoplasmic reticulum lumen.
The intense peripheral RS1 staining of bipolar cells and cone inner
segments and RS1 labeling of nonpermeabilized cultured bipolar cells is
consistent with the localization of a significant fraction of RS1 on
the surface of these cells. Together, these results support the model
in which RS1 is synthesized in photoreceptor and bipolar cells as a
disulfide-linked oligomeric complex. This secreted complex associates
with the surface of photoreceptor and bipolar cells. The molecular
composition of the RS1 complex and identity of the interacting cell
surface components are currently under investigation.
Many missense mutations associated with X-linked juvenile retinoschisis
involve the introduction or substitution of cysteine
residues.
9 15 It is likely that these cysteine mutations
cause misfolding and defective subunit assembly of RS1, as recently
reported for the photoreceptor-specific protein,
peripherin/rds.
38
The role of RS1 in the pathogenesis of X-linked juvenile retinoschisis
remains to be determined at a molecular level. Earlier histopathologic
and electrophysiological studies led to the conclusion that defective
Müller cells were directly responsible for the disease
state.
5 8 However, expression and localization of RS1 on
photoreceptors and bipolar cells suggest that these cells are directly
involved in the pathogenesis of X-linked juvenile retinoschisis. The
discoidin-like domain of RS1 probably participates in cellular
adhesion, possibly by stabilizing the association of the extracellular
matrix to the surface of photoreceptors and bipolar cells. Mutations in
RS1 that compromise the putative role of RS1 as an adhesion protein
could destabilize the retinal tissue. The foveal region may be
especially compromised because of its unique cellular organization,
constituting a weak spot. This fragility is seen in another human
retinal disorder, that of the macular hole in which vitreal shrinkage
pulling at the retina leads to detachment at the level of the
macula.
39 It is unclear why microcysts form between the
GCL and the inner limiting membrane, a region that appears to contain
little, if any, RS1. Perhaps, the extracellular matrix required to
maintain the integrity of the retinal tissue between the GCL and the
inner limiting membrane extends into the INL where it is anchored to
retinal bipolar cell surfaces through the RS1 complex. Loss of the RS1
adhesion function in persons with X-linked retinoschisis could lead to
destabilization of the matrix and splitting of the inner retina cell
layer.
Most persons with X-linked retinoschisis produce ERGs with normal
a-waves originating from photoreceptors but reduced or absent
b-waves.
3 4 Although the b-wave was initially thought to
be derived from Müller cell depolarization, there is growing
evidence to suggest that the b-wave originates directly from
depolarizing bipolar cells.
3 40 41 Immunostaining showing
that RS1 is associated with bipolar cells is consistent with the direct
involvement of bipolar cells in b-wave activity and in X-linked
juvenile retinoschisis.
In summary, the results of this study indicate that RS1 is expressed in
both photoreceptors and bipolar cells as a secreted, disulfide-linked
oligomeric protein. On export to the extraretinal space this complex
binds to the surface of photoreceptors and bipolar cells where it may
function as a cellular adhesion protein to maintain the integrity of
the retina tissue.