Abstract
purpose. Inherited retinal degenerations such as retinitis pigmentosa (RP) are
characterized by progressive death of the photoreceptors due to
apoptosis. To identify changes in gene expression associated with the
degenerative state in RP retinas, expression profiling of
apoptosis-related genes was performed using a gridded array
technique.
methods. Total RNAs from RP and control retinas were used to generate
radiolabeled cDNA probes to screen gridded membrane arrays of 205
apoptosis-related genes. Reverse transcription–polymerase chain
reaction was used to generate probes corresponding to differentially
expressed genes for Northern blot analysis and for mRNA in situ
hybridization studies of retinal cryosections. Fluorescence
immunocytochemistry was performed on retinal sections using available
antibodies.
results. By expression profiling, we identified upregulated expression of the
mRNA for secreted Frizzled-related protein-2 (SFRP2) in RP retina in
comparison with control. By Northern blot analysis, SFRP2 mRNA levels
were 2- to 20-fold higher in RP samples than in controls. The
localization of SFRP2 mRNA by in situ hybridization varied according to
the degree of degeneration, from stratified in relatively
well-preserved retinas to diffuse in the highly degenerative state. By
immunofluorescence, SFRP2 protein in RP retinas was found mainly to
colocalize with the cell adhesion and signal transducing proteinβ
-catenin.
conclusions. SFRPs can regulate apoptosis in vitro and appear to interact with the
Wnt/Frizzled signaling pathway, which includes routes to apoptotic
activation. Increased SFRP2 expression in RP retinas suggests that an
altered pattern of Wnt signal transduction may be a step in the
degenerative process linking causal mutations with eventual
photoreceptor demise.
Accumulating evidence indicates a common mechanism responsible
for photoreceptor cell death in many forms of inherited retinal
degeneration, including those classed as retinitis pigmentosa
(RP).
1 2 We obtained first indirect evidence that
photoreceptor death in RP is apoptotic in nature, through detection of
increased mRNA levels of the marker gene clusterin in affected
retinas.
3 Subsequently, several groups directly
demonstrated activation of apoptosis in photoreceptors during retinal
degeneration in different rodent models of RP.
4 5 6 Apoptosis is a term covering a multitude of cellular pathways, which,
after activation, converge upon an apparently rather uniform collapse
of cell integrity. Although transgenic studies have attempted to
determine the roles of specific genes in rodent photoreceptor
apoptosis,
7 8 9 10 as yet no clear picture has emerged to
indicate which potential pathways of apoptotic activation occur in
different instances of photoreceptor cell death.
To gain insight into mechanisms of human retinal degeneration, we
undertook expression profiling to identify further differentially
expressed apoptotic genes in RP, using screenings of gridded
gene arrays. We report patterns of retinal expression of one such gene,
SFRP2, and discuss some implications in relation to possible apoptotic
mechanisms.
Control eyes from donors with no history of ocular disease were
obtained from the UK Transplant Support Service, Bristol, Moorfields
Eye Hospital, London, or St. Thomas’ Hospital, London. Eyes from RP
donors were obtained through the British RP Society’s Eye Donor
Scheme. In all cases, written informed consent was obtained before
enucleation. The postmortem interval (pmi) between death and freezing
of dissected retinas in liquid nitrogen or fixation of whole eyes in
4% paraformaldehyde was recorded where possible. Sectors of fixed
tissue were embedded, frozen in liquid nitrogen–cooled isopentane and
stored in liquid nitrogen until use. Sections (10-μm-thick) were cut
on a cryostat, thaw-mounted on gelatin-coated slides, and stored at−
70°C until used.
For control eyes, the donor details were as follows: CON1, male, age 53
years, pmi ∼34 hours; CON2, female, 76 years, pmi = 11 hours;
CON3, female, 57 years, pmi = 29 hours; CON4, female, 72 years,
pmi = 28 hours; and CON5, male, 79 years, pmi = 21.5 hours.
For RP eyes, the donor details were as follows: RP1, male, 60 years,
pmi = 7.5 hours, advanced sporadic RP; RP2, male, 84 years,
pmi = 15 hours, probable recessive RP (six brothers in family, of
whom two were affected); RP3, male, 51 years, pmi ∼10 hours,
probable X-linked RP (two brothers, of whom one was affected); and RP4,
male, 79 years, pmi not greater than 24 hours, simplex RP.
Total RNA was extracted from frozen tissues using the RNeasy
system (Qiagen GmbH, Hilden, Germany) or according to a protocol
associated with the differential array screening procedure (Clontech
Laboratories, Palo Alto, CA). For Northern blot analysis, 3 μg RNA
samples were processed as previously described.
3 Isolated
cDNA inserts or polymerase chain reaction (PCR) products (∼20–60 ng)
in low melting point agarose (GIBCO–BRL Life Technologies, Paisley,
UK) were labeled with α-[
32P]dCTP using the
Rediprime kit (Amersham Pharmacia Biotech, Little Chalfont, UK)
according to the manufacturer’s instructions, and hybridizations were
performed in the presence of 50% formamide at 42°C overnight,
followed by stringent washing and autoradiography. The probes were
stripped off between hybridizations. Quantification of autoradiographic
signals was performed using a laser densitometer (LKB, Sweden).
Gridded arrays of 205 apoptosis-related partial cDNAs were
obtained from Clontech Laboratories and processed in accordance with
the manufacturer’s protocols. Total RNA samples (2 μg) from CON1 and
RP3 retinas were reverse-transcribed in the presence ofα
-[32P]dATP to produce complex probes. After
hybridization to the membranes at 68°C overnight, stringent washings,
and autoradiography, the patterns of spots were compared by direct
visual inspection of the films and after computer-based image
enhancement. Genes were then identified by reference to a key and to
the online database at<http://www.clontech.com/archive/JUL98UPD/JUL98Atlas.html>.
Approximately 1 μg of retinal total RNA (RP3) was
reverse-transcribed in 20 μl reactions with 1.25 U AMV reverse
transcriptase (GIBCO–BRL Life Technologies) and random hexamer primers
(Pharmacia, St. Albans, UK) for 60 to 90 minutes at 42°C, followed by
heat inactivation of the enzyme. PCR amplification of an SFRP2 partial
cDNA was then performed using primers HSARP1.1,
5′-GGGTCGCGCCCACGATGCTG-3′, and HSARP1.2, 5′-CTTCCTCGGTGGCTGGCAGG-3′
(derived from the human sequence, GenBank Accession Number AF017986),
in 50 μl reactions containing 10 pmol of each primer, 200 mM dNTPs,
and 1.0 U
Taq DNA polymerase (GIBCO–BRL Life Technologies),
with the following conditions: 94°C for 3 minutes, then 30 cycles of
94°C for 45 seconds, 55°C for 45 seconds, and 72°C for 60
seconds, and a final step of 72°C for 5 minutes. A PCR product of the
expected size (510 bp) was cloned into the pCR2.1-TOPO vector
(Invitrogen, Leek, The Netherlands) according to the manufacturer’s
protocol to give clone pHSARP1.TA27. For the generation of riboprobes,
the SFRP2 insert was subcloned into the vector pBluescript SK
(Stratagene, La Jolla, CA) to give clone pBS.SARP1. Cloning and
orientation of the expected SFRP2 cDNAs were confirmed by partial
sequence analysis (MWG-Biotech GmbH, Ebersberg, Germany). The control
probe was a PCR product of 835 bp amplified from human retinal cDNA
using primers specific for glyceraldehyde-3-phosphate dehydrogenase
(GAPD): GAPDH.1, 5′-CCTTCATTGACCTCAACTACAT-3′, and GAPDH.2,
5′-TACCAGGAAATGAGCTTGACAA-3′, based on the published
sequence,
11 GenBank Accession No. M33197.
Labeling of sense and antisense riboprobes generated from
linearized restriction endonuclease enzyme digestions of the
pBluescript clone of SFRP2, and subsequent hybridization of the probes
to cryostat sections of human retinas, was performed using the color in
situ kit (Amersham Pharmacia Biotech) in accordance with the
manufacturer’s recommendations. After hybridization overnight at
55°C in the presence of 50% formamide, stringent washing, and RNaseA
treatment, the hybridized probes were detected by incubation with
alkaline phosphatase–conjugated anti-fluorescein antibody and
development in the presence of NBT/BCIP substrates. A deep purple-brown
reaction product indicates hybridization.
Cryostat sections (10 μm) of fixed frozen tissues were processed
using the classic immunofluorescence technique.
12 For
detection of SFRP2, the primary antiserum was a goat polyclonal
anti–FRP-2 (sc-7426; Santa Cruz Biotechnology, Santa Cruz, CA), used
at a dilution of 1:300. The secondary antibody, a fluorescein
isothiocyanate–coupled anti-goat IgG (Sigma, St. Louis, MO), was used
at 1:200. For detection of β-catenin, the primary antiserum was a
goat polyclonal antibody (sc-1496; Santa Cruz Biotechnology) used at
1:300, with the secondary used as above.
Hybridization of single-stranded retinal cDNA probes from donors
CON1 and RP3 to gridded arrays of human apoptosis-related genes
produced differing autoradiographic patterns (data not shown). Three
genes showing increased expression in RP were identified as those
encoding c-Jun, PIG7 (manuscript in preparation) and secreted
apoptosis-related protein-1 (SARP1),
13 also known as
secreted Frizzled-related protein-2 (sFRP-2),
14 and
SDF5.
15 We have adopted the approved nomenclature SFRP2
for this gene
16 and report further analysis of its
expression in normal and RP retinas.
We examined the retinal expression of SFRP2 mRNA in 4 cases of RP
and 3 controls (
Fig. 1 , top panel). A major transcript of approximately 2.4 kb was detectable
in all RP samples, together with a likely precursor mRNA of ∼4.4 kb
in the most highly expressing cases; by visual inspection, no
corresponding bands were detectable in control samples. For the cases
of RP, in which the retinas exhibited different stages of
histomorphologic degeneration (see below), SFRP2 mRNA levels were
variably increased compared with controls. However, there were no
obvious correlations between the SFRP2 mRNA level and the extent of
degeneration or RP type. Probing for GAPD expression confirmed that the
differences were not due to variation in RNA loading or integrity (
Fig. 1 , middle panel). Laser densitometric analysis permitted detection of
weak signals corresponding to the SFRP2 mRNAs at ∼2.4 kb in control
samples. After normalization to GAPD levels, comparisons indicated a
range of approximately 2- to 20-fold increased expression of SFRP2 mRNA
in RP retinas relative to that in controls (
Fig. 1 , bottom panel).
We investigated the cellular distribution of SFRP2 mRNA expression
in available retinal sections using a color in situ hybridization
protocol
(Fig. 2) . Positive hybridization was indicated by the presence of a purple
reaction product; careful examination permitted discrimination of this
product from pigment in the retina or retinal pigment epithelium (RPE).
In a normal retina (CON5), mRNA was detectable at the outer margins of
the inner nuclear layer (INL), with a slightly stronger reactivity at
the inner segments of the photoreceptors (PR;
Fig. 2A ). Similar
patterns were detected in sections from two additional control retinas
(data not shown). There was some evidence of nonspecific reactivity
with the sense probe in CON5
(Fig. 2B) . In sections of retina from RP1
(Fig. 2C) , which showed severe disruption of retinal architecture, loss
of photoreceptors, gliosis, and extensive accumulation of pigment
within the retina, there was evidence of diffuse labeling throughout
the retina, including in association with pigment clumps. Although this
case showed the highest level of expression of SFRP2 mRNA by Northern
blot analysis (
Fig. 1 , top, lane 1), the intensity of in situ
hybridization labeling in the sections examined was not as great as
with case RP3
(Fig. 2E) , suggesting regional variability in SFRP2 mRNA
levels within the retina. In the macular region of retina RP3
(Fig. 2E) , which showed moderate preservation of retinal morphology,
including the presence of photoreceptor rosettes, a more intense
labeling was associated with the INL and also the residual
photoreceptor nuclei, but most strikingly very intense staining was
detected at the ganglion cell layer (GCL; arrowheads). In retina RP4
(Fig. 2G) , which showed extensive regions of relatively well preserved
retinal morphology, including surviving photoreceptors, there was
labeling in the INL and outer nuclear layer (ONL) and also at the
photoreceptor inner segments, resembling that seen in normal retinas.
The sense control probings of RP retinas
(Figs. 2D 2F 2H) showed only
slight background coloration.
SFRP2.
The pattern of immunofluorescence using a polyclonal anti-SFRP2
antibody was compared in sections of control and RP retinas (
Figs. 3A 3C 3E 3G 3I ). Very little immunostaining was evident with the
control retina (
Fig. 3A ; near the inner retinal margins there is
nonspecific fluorescence from the vasculature) or in sections from two
additional controls (data not shown); occasional amacrine cells were
faintly positive. In the three cases of RP examined, there were
differences in the distribution of immunofluorescence. In retinas RP1
and RP3 (
Figs. 3C and 3E , respectively), most of the labeling was
located at the retinal margin of the inner limiting membrane (ILM). In
retina RP4
(Fig. 3G) , however, more intense labeling was noted in
specific retinal layers, notably the GCL, the inner plexiform layer,
the outer plexiform layer and amacrine cells, the photoreceptor outer
segments, and also, most probably, the RPE. Interestingly, distinct
cytoplasmic labeling was noted in numerous residual surviving
photoreceptors of the ONL.
β-Catenin.
In view of the implicated role of SFRPs in regulating the expression ofβ
-catenin via the Wnt signaling pathway
12 13 (see also
Discussion), we examined the pattern of β-catenin immunoreactivity on
adjacent retinal sections
(Figs. 3B 3D 3F 3H 3J) . In control
retina, the outer limiting membrane, outer and inner plexiform layers,
and (at a lower intensity) the GCL were immunopositive
(Fig. 3B) ;
similar results were obtained with two additional controls (data not
shown). In retinas RP1 and RP3, the immunolabeling was localized mainly
to the ILM
(Figs. 3D 3F) , although with evidence of labeling also at
Bruch’s membrane seen most clearly in RP3 (
Fig. 3F , arrow), and
diffuse labeling throughout the body of the degenerative retina. There
was some indication that this labeling was cytoplasmic in the ONL, INL,
and GCL of RP3
(Fig. 3F) . In retina RP4, the pattern of localization
(Fig. 3H) was similar to that observed with normal retina, although
with a greater intensity and with additional labeling of the ILM.
Control sections lacking primary antibodies
(Figs. 3I 3J) displayed
minimal nonspecific immunofluorescence, mainly associated with the
retinal vasculature and choroid. The RPE and some intraretinal pigment
granules showed a yellow autofluorescence.
By examination of apoptosis-related genes in human retinal
degeneration using differential screening of gridded arrays, we
identified increased expression of the mRNA for SFRP2, a member of the
family of secreted Frizzled-related proteins, in RP retinas. The SFRP
genes, of which human, rodent, bovine, and
Xenopus members
have been described,
16 encode secreted proteins with
cysteine-rich domains homologous to those of Frizzled (Fz) proteins but
lacking the membrane-spanning segments of the latter group. The
numerous Frizzled genes encode cellular membrane proteins with seven
transmembrane domains, which act as receptors for the similarly diverse
array of secreted Wnt proteins.
16 17 18 Through signaling
interactions conserved between
Drosophila and vertebrates,
Wnts have been implicated in the control of key developmental processes
via regulation of cell morphogenesis, cell mitogenesis, and cell fate.
SFRPs may modulate signaling via Fz through their binding interactions
with Wnts, or indeed via direct binding to Fz receptors
themselves.
19 A recent report described SFRP2 mRNA
localization to the INL of bovine retina and SFRP5 expression in the
RPE, suggesting possible roles for these genes in regulating the
polarity of photoreceptors or other retinal cells.
20 Our
present findings add to the evidence for a role for SFRP2 in the eye,
potentially in influencing retinal cell survival.
In the cases of RP examined, we found increased and altered
patterns of expression of SFRP2 mRNA and protein relative to controls,
although there was variability in both the levels and distribution of
transcripts and immunoreactive protein. SFRP2 protein localization
correlated most closely with the mRNA pattern of expression in the case
of RP4, where the retina was relatively well preserved. In the other RP
cases compared, with more advanced degeneration, the majority of
immunoreactive protein was concentrated at the inner margins of the
ILM, possibly in association with the terminal end feet of Müller
glial cells which proliferate during degeneration, suggesting that as
the disease progresses SFRP2 is increasingly vectorially secreted out
of the neural retina into the vitreous. The discrepancy between high
SFRP2 mRNA levels and relatively low levels of immunoreactive SFRP2
protein, notably in retinas RP1 and RP3, may further support this
possibility, although more information on the relationship between
transcription, translation, and posttranslational processing and
degradation of SFRP2 will be required to establish this.
The role of SFRPs in modulating or antagonizing the Wnt signaling
pathway, taken together with our present findings, suggests that
dysregulation of this pathway may be occurring in the apoptotic
degenerative processes in RP retinas. Of potential relevance are
reports of increased expression of SFRP homologues during apoptosis
occurring in vitro
13 21 and in involution of rat mammary
gland, ovary, and prostate.
22 Furthermore, overexpression
of transfected SFRP2 in breast adenocarcinoma cells increases their
resistance to apoptotic signals, associated with increased
intracellular levels of β-catenin,
13 a multifunctional
protein engaged in cell–cell junction interactions with E-cadherin and
transcriptional regulatory complexes with members of the
TCF/Lef-1/pangolin family.
17 Immunoreactive SFRP2 andβ
-catenin largely colocalized in RP retinas, but we found no clear
evidence of increased β-catenin levels, suggesting that if modulation
of Wnt signaling is occurring in retinal degeneration, the interplay of
regulatory interactions is more complex than in in vitro situations.
Speculatively, SFRP2 upregulation in RP retinas may reflect an
anti-apoptotic response, analogous to that proposed for the
glycoprotein clusterin, which ultimately fails to protect the
photoreceptors due to loss of regulation of the Wnt signaling pathway
and downstream collapse of apoptosis suppression mechanisms. Although
the complex network of Wnt signaling and its ramifications over a
diversity of cellular interactions ensures that progress in unraveling
the significance of altered SFRP expression in RP will not be
straightforward, potentially new therapeutic targets may emerge for the
eventual treatment of retinal degenerative diseases.
The authors thank the Rayne Management Committee for the provision
of research facilities.