Knock-Down of GFRα4 Expression by RNA Interference Affects the Development of Retinal Cell Types in Three-Dimensional Histiotypic Retinal Spheres

Abstract

purpose. To determine the role of glial cell line-derived neurotropic factor family receptor alpha 4 (GFRα4) during retinogenesis in a three-dimensional histiotypic in vitro model of the embryonic chicken retina.

methods. Retinal spheres were cultured from dissociated 6-day-old chicken retina under permanent rotation and transfected with GFRα4 siRNA at culture day 2. Alterations on proliferation, apoptosis, and differentiation were determined by semiquantitative RT-PCR, in situ hybridization, and immunohistochemistry after 24, 48, and 72 hours.

results. In contrast to control cultures, retinal spheres transfected with GFRα4 siRNA showed reduced GFRα4 mRNA expression of only 38% after 24 hours, 3% after 48 hours, and 5% after 72 hours. Based on the suppression of GFRα4, a decline in proliferating cells from 10% to 4.8% even after 24 hours and a reduction of sphere size by up to 25% at later culture stages were observed. Moreover, the number of Pax 6-positive amacrine, ganglion, and horizontal cells was significantly decreased from 36% to 16% in GFRα4 siRNA-transfected retinal spheres 72 hours after transfection. Additionally, GFRα4 gene silencing affected the development of different types of photoreceptors, as revealed by a significant decrease of blue opsin mRNA expression from 29% to 2%, whereas green opsin mRNA and the number rho4D2-positive photoreceptors were significantly increased.

conclusions. These data showed for the first time that GFRα4 plays an essential role in regulating, at least in vitro, the development and differentiation of various cell types during retinogenesis.

Glial cell line-derived neurotrophic factor (GDNF) family alpha receptors (GFRαs) include GFRα1, GFRα2, GFRα3, and GFRα4. The GFRαs are activated through binding of specific GDNF family ligands (GFLs). For example, GDNF preferentially binds to GFRα1, neurturin to GFRα2, artemin to GFRα3, and persephin to GFRα4. ^{1 2}  

Based on their different temporal and spatial expression patterns during the development of the nervous system, it has been postulated that GFRαs might be involved in the diversification of different neuronal cell types. ^{2 3} Several recent studies using GFL- or GFRα-deficient mice showed dramatic changes in the development and survival of sympathetic, parasympathetic, and sensory neurons. ^{2 4 5}  

Because GFRαs have been detected in many neuronal tissues, it was not surprising that GFRα1 and GFRα2 have also been found in chick, mouse, and rat retina. ^{6 7 8 9} A number of studies could demonstrate that GFRα1, GFRα2, and their specific ligands are involved in the proliferation and survival of retinal neurons in vivo and in vitro. ^{8 9 10 11} More recently, a third member of GFRαs has been identified in the developing chicken retina. Here, GFRα4 has been expressed in temporal and spatial patterns, reflecting the vitreal-to-scleral and central-to-peripheral gradient of retinal development. ¹² The first GFRα4 expression was detected in ganglion cells of the central retina at embryonic day 6 (E6) and was gradually extended to cells of the outermost portion of the retina and toward the ora serrata during further development. In the mature retina, GFRα4 mRNA was found in ganglion, amacrine, and horizontal cells and a subpopulation of photoreceptor cells. However, based on the absence of a GFRα4 knock-out mouse, the role of GFRα4 in neuronal and retinal development is poorly understood. 

To gain a better understanding of GFRα4 function during retinogenesis, we suppressed GFRα4 by RNA interference in histiotypic retinal spheres. Retinal spheres represent a highly reproducible three-dimensional culture system that mimics retinogenesis and can be manipulated genetically by a number of applications. ^{13 14 15 16} Under permanent rotation, retinal spheres arise through the aggregation of dispersed retinal progenitor cells of E6 chicken embryos, followed by their proliferation, migration, and differentiation. Based on the correct time course of cellular differentiation and the formation of layers, retinal spheres reflect almost all aspects of retinogenesis in vivo. ^{17 18}  

Our study demonstrates for the first time that GFRα4 gene silencing in retinal spheres affects several aspects of retinogenesis in vitro. GFRα4-deficient retinal spheres showed a decrease in proliferation and growth. Moreover, we found that the numbers of Pax 6-positive ganglion, amacrine, and horizontal cells and blue-sensitive cones were decreased, whereas the mRNA of green-sensitive cones and the number of rho4D2-positive cells were significantly increased. 

Materials and Methods

Production of Retinal Spheres

To produce retinal spheres, retinas of 6-day-old chicken embryos (White Leghorn) were used (for details, see Rothermel et al. ¹⁹ and Willbold and Layer ²⁰ ). Briefly, central parts of the retina were isolated and collected in F12 medium. Tissue was dissociated enzymatically (0.05 mg/mL trypsin; Worthington Biochemicals/Cell Systems, Remagen, Germany) and mechanically by 30 to 40 gentle strokes with a round-bored Pasteur pipette. For production of retinal spheres, 2 × 10⁶ cells/mL were cultured in 35-mm dishes containing 2 mL growth medium (Dulbecco modified Eagle medium [DMEM], 10% fetal calf serum [FCS], 1% L-glutamine, 0.15% penicillin/streptomycin; all from Gibco, Berlin, Germany) on a gyratory shaker in an incubator (37°C, 95% air, 5% CO₂; Heraeus Holding GmbH, Hanau, Germany). 

Cryosections and Immunocytochemistry

Retinal spheres were harvested at appropriate stages, fixed in phosphate-buffered saline (PBS) containing 4% formaldehyde for 30 minutes at room temperature, washed with PBS, and soaked in 25% sucrose (in PBS). Cryosections of 10-μm thickness were cut on a cryostat (Leica, Bensheim, Germany) and mounted on superfrost slides (VWR, Darmstadt, Germany). For immunostaining, sections were preincubated with a blocking solution (3% bovine serum albumin [BSA], 0.1% Triton-X-100 in PBS) for 30 minutes at room temperature. Then tissues were incubated with the primary antibodies for 1 hour at room temperature, followed by three washes in PBS. Pax 6 hybridoma supernatant (Developmental Studies Hybridoma Bank [DSHB], Iowa City, Iowa) was added undiluted, whereas the rod-specific monoclonal antibody rho4D2 (a generous gift from David Hicks) was used in a dilution of 1:2000. For the detection of primary antibodies, sections were incubated with donkey anti-mouse-conjugated Cy3 (10 μg/mL; Dianova, Hamburg, Germany) for 1 hour. Between the last two washes, cell nuclei were stained with DAPI (0.1 mg/mL 4′,6-diamidine-2-phenylindol-dihydrochloride in PBS). For bromodeoxyuridine (BrdU) incorporation experiments, retinal spheres were incubated in the presence of 50 μM BrdU (Sigma, Deisenhofen, Germany) 16 hours before harvesting. For the detection of BrdU, we used the same procedures as described except that sections were treated first with 4 N HCl for 5 minutes before the BSA-containing blocking solution was added. To analyze apoptotic cells, the TUNEL-TMR Red Kit (Roche Diagnostics, Mannheim, Germany) was used according to the manufacturer’s instructions. 

Reverse Transcriptase-Polymerase Chain Reaction

For semiquantitative RT-PCR, total RNA was isolated from retinal cultures of appropriate stages, treated with DNase I to digest any genomic DNA, and transcribed in cDNA by using the reverse transcription system from (Promega, Mannheim, Germany). PCR was carried out with 25 cycles to ensure amplification in a linear range. Primers were designed with use of the Primer3 software (http://www-genome.wi.mit.edu/cgi-bin/primer/primer3). Target cDNA was amplified by PCR with specific primers (Table 1) . PCR fragments were separated by agarose gel electrophoresis and analyzed by densitometric quantification (Bio-Rad, Munich, Germany). PCR parameters were 94°C denaturation for 1 minute, 55°C annealing for 1 minute, and 72°C elongation for 1 minute. The relative expression of mRNA was shown as a percentage of GAPDH band intensity for each culture stage. Each experiment was carried out at least three times. 

Transfection of Retinal Spheres with siRNA

HPLC-purified duplex siRNA was designed and purchased (MWG Biotech, Munich, Germany). Duplex siRNA was directed to the GFRα4 coding sequence at nucleotide positions 244 to 265 (sense, 5′-GUGCAAGCGAGGCAUGAAAUU-3′; antisense, 5′-UUUCAUGCCUCGCUUGCACUU-3′). A nonsense GFRα4 siRNA with a nucleotide exchange at position 249 (GFRα4 siRNA mismatch) was designed to determine the specificity of GFRα4 downregulation and to test an unspecific inflammatory response (sense, 5′-GUGCACGCGAGGCA UGAAAUU-3′; antisense, 5′-UUUCAUGCCUCGCGUGCACUU-3′). Transient transfection was carried out at culture day 2 (siPROT Lipid siRNA Transfection Kit; Ambion, Austin, TX). At culture day 2, growth medium was removed and replaced by 800 μL transfection medium containing DMEM, 2% FCS, and 2 μmol siRNA. Retinal spheres were treated for more than 12 hours at 37°C, 95% air, and 5% CO₂ on a gyratory shaker with either GFRα4 siRNA or GFRα4 mismatch siRNA. At day 3, transfection medium was removed and replaced with standard growth medium containing 10% FCS. Retinal spheres were harvested at day 4 (24 hours after transfection), day 5 (48 hours after transfection), and day 6 (72 hours after transfection). 

Generation of DIG-Labeled RNA Probes and In Situ Hybridization

Digoxigenin (DIG)-labeled GFRα4 RNA probes were generated as previously reported. ¹² For in situ hybridization, retinal cryosections were treated for 15 minutes with proteinase K (10 μg/mL PBS). Sections were rinsed in Tris-glycine (100 mM Tris/HCl pH 7.0, 100 mM glycine), postfixed for 20 minutes in 4% phosphate-buffered paraformaldehyde, and rinsed twice for 5 minutes in PBST (PBS, 0.1% Tween). For hybridization, sections were incubated with prehybridization solution (50% formamide, 20% dextran sulfate, 10 mg/mL herring sperm DNA, 500 μg/mL yeast tRNA, 50 μg/mL heparin) for 3 hours at 65°C. After removing the prehybridization solution, sections were incubated overnight at 65°C in hybridization solution containing 300 ng/mL DIG-labeled RNA probes. Sections were rinsed twice in 2 × SSC (50% formamide) for 5 minutes at 60°C, in 0.2 × SSC (50% formamide) for 30 minutes at 60°C, in 0.1 × SSC for 10 minutes at room temperature, and finally in PBST. Sections were incubated with blocking solution (2% BSA, 10% sheep serum in PBST) for 3 hours at room temperature followed by incubation with an anti-DIG antibody conjugated with alkaline phosphatase (diluted 1:2000; Roche Diagnostics, Mannheim, Germany) for 2 hours. The antibody solution was removed, and sections were rinsed twice in alkaline phosphate buffer (100 mM Tris/HCl pH 9.5, 10 mM NaCl, 260 μg/mL levamisole). Slides were then stained with NBT/BCIP (Roche Diagnostics). Staining reaction was stopped by rinsing in PBST and distilled water. Slides were dried at 37°C and mounted with Kaiser glycerin-gelatin for documentation. 

Growth Analysis of Retinal Spheres

To examine alterations in growth during the culture period, retinal spheres were documented directly in the culture dish using an inverted microscope equipped with a digital monochrome camera (Nikon, Cologne, Germany). The size of 40 to 60 retinal spheres of each stage and experiment was determined using documentation and analysis software (Nikon-Lucia G). The obtained data were transferred and processed (Excel 7.0; Microsoft, Redmond, WA). 

Cell Counting and Statistical Analysis

To determine the number of immunopositive cells, frozen sections (each containing 30 to 40 spheroids) were stained with DAPI and the corresponding antibodies. The percentage of immunolabeled cells per cryosection of a single spheroid was calculated in relation to DAPI-positive cells of the same spheroid section. At least nine cryosections of different spheroids derived from three individual experiments were analyzed (n ≥ 9 investigated spheroids), corresponding to 5000 cells per stage and staining. To compare alterations in cell number of immunopositive cells in spheres with different sizes at each culture stage, the same number of cells (counted by DAPI staining) of control, mismatch GFRα4 siRNA–, and GFRα4 siRNA-treated spheres was used for accurate quantification. Data were presented as mean ± SD and were compared using two-tailed, paired Student’s t-test. Note that SD represented the differences between the calculations of individual spheres. 

Microscopy and Photography

Photomicrographs of sections were taken either with an inverted microscope (Nikon TE2000) or with an upright microscope (Nikon E600) combined with a digital camera and processed with documentation and analysis software (Nikon Lucia G; Adobe Photoshop 7.0 [Adobe, San Jose, CA]; and Excel 7.0). 

Results

siRNA-Mediated Gene Silencing of Endogenous GFRα4 mRNA

Temporal expression of GFRα4 in developing retinal spheres was analyzed at different stages by semiquantitative RT-PCR (Figs. 1A 1B) . Expression of GFRα4 mRNA transcripts was found first at culture day 2. Afterward expression was gradually increased over the complete culture period. As expected, transfection with GFRα4 mismatch siRNA at culture day 2 showed no significant effects on GFRα4 expression when compared with nontransfected control cultures (Figs. 1C 1D) . However, when compared with control cultures, transfection with GFRα4 siRNA revealed that expression of GFRα4 mRNA was decreased from 81% to 45% at culture day 4 (24 hours after transfection; Figs. 1C 1D ) and reached a basal level of 3% at day 5 (48 hours after transfection) and 5% at culture day 6 (72 hours after transfection). 

To confirm the data obtained by semiquantitative RT-PCR, we carried out in situ hybridization using cryosections of retinal spheres (Fig. 2) . When compared with control cultures, application of GFRα4 siRNA showed only a small but distinctive decrease in the number of GFRα4-expressing cells at day 4 (Figs. 2A 2B) . At day 6 in culture, GFRα4 expression of nontransfected (Fig. 2C)and GFRα4 mismatch siRNA-transfected cultures (Fig. 2E)was detected in cells of the inner nuclear layer (INL; Figs. 2C 2E , arrowheads), whereas in GFRα4 siRNA transfected-cultures only background staining occurred and nearly no GFRα4-positive cells were detectable (Fig. 2D) . 

Growth and Proliferation in GFRα4-Deficient Retinal Spheres

Growth analysis (Fig. 3A)of nontransfected retinal spheres showed a gradual increase of size during the first 6 days in vitro (132 ± 16 μm at day 1; 201 ± 17 μm at day 2; 375 ± 17 μm at day 4; 403 ± 23 μm at day 6). Thereafter, the average size of retinal spheres remained constant (394 ± 18 μm at day 8; 401 ± 18 μm at day 10). However, when retinal spheres were transfected with GFRα4 siRNA at day 2, we observed an average decrease in diameter of 31 μm even beyond 24 hours after transfection (GFRα4 siRNA, 344 ± 18 μm; control, 375 ± 17 μm). Between culture days 6 and 10, GFRα4 siRNA-transfected spheres were 25% smaller on average than nontreated spheres (day 6: GFRα4 siRNA, 301 ± 18 μm; control, 403 ± 23 μm; day 8: GFRα4 siRNA, 299 ± 16 μm; control, 394 ± 18 μm; day 10: GFRα4 siRNA, 302 ± 14 μm; control, 401 ± 18 μm). Next, we analyzed whether the reduction in size was caused by a decrease of proliferating cells or an increase of apoptotic cells. As shown in Figures 3B and 4 , proliferation declined from 10.5% ± 3% in mismatch to 4.8% ± 2% in GFRα4 siRNA-treated retinal spheres 24 hours after transfection, whereas after 48 hours no significant alteration was observed (mismatch siRNA, 4.1% ± 2%; GFRα4 siRNA, 4.0% ± 2%). Quantification of TUNEL-stained sections in relation to DAPI-positive cells showed no significant alterations in the number of apoptotic cells, neither in control nor in transfected cultures (Figs. 3C 5) . 

Knock-Down of GFRα4 Decreased the Number of Pax 6–Positive Cells

In the developing retina, Pax 6 is initially expressed in nearly all retinal precursor cells, ²¹ whereas its expression becomes restricted to amacrine, horizontal, and ganglion cells during further maturation of the retina. A similar expression pattern was found in mature retinal spheres after 6 days in culture. Here, most Pax 6-positive cells were arranged in the INL surrounding circular inner plexiform layer (IPL; stars in Fig. 6 ). Within these IPL-like areas, single cells—probably Pax 6-positive ganglion cells—were found. A small number of Pax 6-positive cells were also detected in nonorganized areas (Fig. 6 , arrows). Quantification of Pax 6-positive cells in relation to the total number of DAPI-positive cells (Fig. 7)showed that 37% ± 4% Pax 6-positive cells were detectable in control and 36% ± 3% in siRNA mismatch-treated cultures, whereas in GFRα4 siRNA-transfected cultures only 16% ± 2% Pax 6-positive cells were detected. This means that 2.3 times fewer Pax 6-positive cells arose in GFRα4-deficient retinal spheres. 

GFRα4 Knock-Down Affects Photoreceptor Development

As shown by a previous study, GFRα4 is also expressed in a subpopulation of photoreceptors in the developing and mature chicken retina. ¹² Based on these findings, we were interested in whether GFRα4 regulates the development of different photoreceptor subtypes. Because GFRα4 mRNA suppression was most effective between culture days 5 and 6 (48 to 72 hours after transfection; Figs. 2C 2D ), retinal spheres of 6-day-old cultures were used for RT-PCR analysis of opsin mRNA expression (Figs. 8A 8B) . With the exception of blue and green opsin mRNA, GFRα4 siRNA transfection did not significantly affect the expression of other opsin mRNAs. When compared with mismatch siRNA-treated cultures (24% ± 2%), blue opsin mRNA expression was decreased to 2% ± 1% in GFRα4 siRNA-transfected cultures. Expression of green opsin mRNA was significantly increased from 71% ± 6% in mismatch to 78% ± 4% in GFRα4 siRNA-transfected retinal spheres. Note that rhodopsin mRNA expression was not significantly changed. Immunohistochemical staining of photoreceptors by the subtype-specific antibody rho4D2, which stains rods and green cones in the chicken retina, ²² revealed that the number of rho4D2-positive cells increase from 15% ± 3% in mismatch cultures to 27% ± 2% in GFRα4 siRNA-treated cultures (Figs. 8C 9) . 

Discussion

In this study we have investigated the role of GFRα4 during retinogenesis in histiotypic retinal spheres by RNA interference. Four main conclusions can be drawn from these experiments. First, retinal spheres show a temporal and spatial expression pattern of GFRα4 that is comparable to the in vivo situation and therefore represent a suitable in vitro model system for GFRα4 gene silencing experiments. Second, GFRα4 is involved in the regulation of proliferation because GFRα4 mRNA suppression causes a dramatic decrease in the number of proliferating retinal progenitor cells. Third, GFRα4 decreases the development of Pax 6-positive cells. Fourth, suppression of GFRα4 affects photoreceptor development, as shown by the decrease of blue opsin mRNA expression and the increase of green-sensitive cones. 

Given that not all designed siRNAs have the same potential to knock down the gene of interest, ²³ we first determined the efficiency of the used GFRα4 siRNA. RT-PCR and in situ hybridization of retinal spheres showed that the used GFRα4 siRNA was highly efficient and suppressed endogenous GFRα4 mRNA to nearly background levels. As reported in previous studies, ^{24 25 26} we could also demonstrate the impressive specificity of siRNA-mediated gene silencing because mismatched GFRα4 siRNA, with only one nucleotide exchange regarding the target RNA, is insufficient for silencing GFRα4 mRNA. 

Expression of GFRα4 in retinal spheres was found in amacrine cells, ganglion cells (localized in IPL-like areas), and cells of nonorganized areas. The latter probably represent horizontal cells or a subpopulation of cones. The spatial expression pattern indicates that GFRα4 is expressed predominantly in cells belonging to the group of neurons that are born early during retinal histogenesis, such as ganglion cells, most amacrine cells, horizontal cells, and cone photoreceptors. ^{27 28} Because the onset and spatial expression of GFRα4 were remarkably similar to those of Pax 6, it was not surprising that the siRNA-mediated knock-down of GFRα4 strongly affected Pax 6 expression. As we have shown, GFRα4 gene silencing is accompanied by a strong decline in Pax 6-positive cells. Although almost all GFRα4-positive cells disappeared in the GFRα4 siRNA-transfected cultures, the number of Pax 6-positive cells was reduced to only 50%. One possible explanation is that different subpopulations of Pax 6-positive cells do not express GFRα4 yet still give rise to different subtypes of amacrine, horizontal, and ganglion cells. 

Nevertheless, we assume that the prominent reduction of Pax 6-positive cells directly correlates with the decreased number of proliferating cells, which, in turn, could explain the reduction in size of retinal spheres. However, the observed alteration in size could also arise by losing cells through programed cell death. Although PSPN and GFRα4 have been shown to prevent programed cell death of motor neurons and dopamine neurons in vitro and in vivo, ^{29 30 31} we found no significant effect of GFRα4 siRNA on apoptosis. Therefore, we conclude that the formation of smaller spheroids was not caused by the loss of cells by naturally occurring cell death but that it was obviously caused by decreased proliferation. 

In accordance with the concept of cell fate determination by extrinsic and intrinsic factors, ^{28 32 33} we postulate that GFRα4 signaling may increase the proliferation of retinal progenitors by regulating different proneural genes. These proneural genes, primarily belonging to the family of homeobox or basic helix-loop-helix transcription factors, are thought to determine retinal cell fate by their stage-specific distribution during retinogenesis. ³⁴ Just as Notch/Delta signaling regulates neural differentiation and promotes the maintenance of progenitor state by Hes1 and Hes5, ^{27 35} it is thought that persephin/GFRα4 signaling can act in a similar manner. Thus, activation of the GFRα4 pathway may regulate the differentiation of a subpopulation of progenitors, such as Pax 6-positive amacrine, ganglion, or horizontal retinal cells. For example, progenitor-specific regulation has already been shown for Prox1, which specifically regulates the proliferation and differentiation of horizontal cell progenitors by preventing early progenitors from leaving the cell cycle. ³⁶ Moreover, a large number of genes code for regulatory proteins that can affect the proliferation and differentiation of retinal progenitors (e.g., Pax 6, Chx10, Ath3, Math3, NeuroD), ^{28 34} representing potential targets of GFRα4 signaling. Unfortunately, persephin, the primary ligand of GFRα4, has not yet been cloned for the chicken, which makes it difficult to investigate signaling in detail through GFRα4 during retinal development. 

However, in addition to the decreased number of Pax 6-positive cells, we observed an increase of rho4D2-positive cells in GFRα4 siRNA-transfected cultures. As shown in a previous study, ²² the antibody rho4D2 specifically stains rods and green cones in the chicken retina. Rhodopsin mRNA expression was not affected by GFRa4 knock-down, but green opsin mRNA was significantly increased; hence, we conclude that the increased number of rho4D2-positive cells resulted exclusively from an elevation of green cones in retinal spheres. On the assumption that blue opsin mRNA expression correlates with the number of blue cones, the actual function of GFRα4 in vivo may consist of blue and green cone differentiation during retinal development. In other words, GFRα4 signaling promotes the development of blue cones and prevents the development of green cones. In a previous study, ²² an opposite effect on the development of green cones has been shown by ciliary neurotrophic factor (CNTF) treatment of retinal monolayer cultures derived from 8-day-old embryonic chicken retina. Here, activation of the CNTF receptor (CNTFR) by CNTF increased the number of green cones, whereas rod photoreceptors were not affected. ²² An effect on blue cones has not been reported in this study, probably because blue opsin mRNA was not detectable in these low-density monolayer cultures. However, in retinal spheres, blue and violet opsins were expressed that allowed analysis of the role of GFRα4 on their expression patterns. These data indicate that both pathways, persephin/GFRα4 and CNTF/CNTFR, are probably linked downstream by the activation or suppression of the same transcription factors. Moreover, based on the fact that other photoreceptors (violet-sensitive, red-sensitive, and rod photoreceptor) were not affected by GFRα4 gene silencing, persephin/GFRα4 signaling represented a novel molecular switch able to regulate the differentiation of blue and green cones. To our knowledge, specific regulation of these two photoreceptors subtypes has not yet been reported. 

Interestingly, the effects of GFRα4 on photoreceptors in retinal spheres reflects some but not all characteristics of human enhanced S-cone syndrome. Patients with this syndrome show an increased number of S-cones (also referred to as blue cones), whereas middle- and long-wavelength cone and rod function are dramatically decreased. ^{37 38} An increase of S-cones has also been shown in mice lacking the neural retina leucin (Nrl) zipper. ³⁹ Additionally, defects in other transcriptional regulators, such as Crx and Nr2e3, caused similar effects on photoreceptor differentiation, especially on the determination of blue cones. ³⁷ Although it has been shown that several transcription factors, such as Crx, NeuroD, Nrl, and Nr2e3, are involved in the regulation of rod cell fate, little is known about factors necessary for the differentiation of S- and M-cones. ⁴⁰ Interestingly, in a recent study, ⁴¹ it has been shown that in the retinal degeneration (rd7) mutant mouse, a model for the human enhanced S-cone syndrome, nearly all photoreceptors expressed cone- and rod-specific genes, representing a hybrid photoreceptor type. These results indicate that the mutation in Nr2e3 was able to regulate the expression of rod- and cone-specific genes. Moreover, especially in the context of green and blue cone cell fate determination, it is known that at least for mice and rats, blue opsins are synthesized in prospective green cones before the expression of green opsins. ⁴² Because the developing immature cones of mice are able to express blue cone opsins but not green cone opsins, it has been suggested that the expression of blue opsins probably represents a default pathway of opsin expression. ^{43 44 45} Interestingly, it has been suggested that the blue-to-green switch in rodents probably correlates with the synthesis of rhodopsin expression in immature rods and that the regulation of cone- and rod-specific genes through the photoreceptor-specific nuclear receptor (PNR) or the orphan nuclear receptor transcription factor (Nr2e3) are somehow linked. In this context, it is thought that GFRα4 signaling is one possible upstream component of the PNR signaling pathway or that GFRα4 probably is involved in the regulation of the PNR and Nr2e3 pathways. In summary, we have demonstrated by siRNA-mediated gene silencing that GFRα4 is involved in the regulation of proliferation and differentiation of cells born early in retinogenesis. In addition to the reduction of Pax 6-positive cells, it is likely that GFRα4 signaling is essential for the determination of blue- and green-sensitive photoreceptors in the chicken retina. 

²  These authors contributed equally to the work presented here and should therefore be regarded as equivalent authors.

 Supported by EU Project Grant SENS-PESTI-QLK4-CT-2002–02264, the Deutsche Forschungsgemeinschaft (Interneuro Graduiertenkolleg Leipzig), Hochschulwissenschaftsprogramm Grant 1207/54769-8La 379/12, an Adolf-Messer grant (AAR) and SFB610/25.

 Submitted for publication November 17, 2005; revised January 6, 2006; accepted April 19, 2006.

 Disclosure: A. Rothermel, None; K. Volpert, None; M. Burghardt, None; C. Lantzsch, None; A.A. Robitzki, None; P.G. Layer, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Primers Used for Reverse Transcription-Polymerase Chain Reaction

Primer	Sequence	GenBank Accession No.
GFRα4	Sense 5′CGA GGC ATG AAA AAG GAG AA 3′	AF045162
	Antisense 5′AAC TCT GCA TAA CGC GAC CT 3′
Red-sensitive opsin	Sense 5′ATC AAC CAG ATC TCG GGG TA 3′	M62903
	Antisense 5′CAG AAG CAG TAG GCC ACG AT 3′
Green-sensitive opsin	Sense 5′CAT GAT GGG GAT AGC TTT CA 3′	M92038
	Antisense 5′TGG TGA TCA TGC AAT TAC GG 3′
Blue-sensitive opsin	Sense 5′CGC TCT GAG TGT GAT GAG GA 3′	M92037
	Antisense 5′AGC TCC GTT CCC ATA GCA GA 3′
Violet-sensitive opsin	Sense 5′CGT GGG CCT TCT ACC TAC AG 3′	M92039
	Antisense 5′TAC TCG CTG CGG TAT TTC GT 3′
Rod-sensitive opsin	Sense 5′TCA CCA TCC AGC ACA AGA AA 3′	D00702
	Antisense 5′GAC CAG GTT CCC ATA GCA GA 3′
GAPDH	Sense 5′GGA CAG TTC AAG GGC ACT GT 3′	V00407
	Antisense 5′CTT CTG TGT GGC TGT GAT GG 3′

Figure 1.

 

Temporal expression of endogenous GFRα4 mRNA under normal culture conditions (A, B) and after transfection with siRNA (C, D). GFRα4 and c-Ret mRNA were analyzed by semiquantitative RT-PCR at different culture stages (A). Densitometric quantification of PCR products showed a gradual increase of GFRα4 and c-Ret during the culture period (B). To test the efficiency of the used GFRα4 siRNA (C, D), retinal spheres were transfected for 12 hours with GFRα4 siRNA and with mismatch GFRα4 siRNA at culture day 2. Retinal spheres were harvested at culture day 4 (24 hours after transfection), at culture day 5 (48 hours after transfection), and at culture day 6 (72 hours after transfection). Agarose gel electrophoresis of RT-PCR products revealed that GFRα4 expression was effectively suppressed 48 and 72 hours after transfection, whereas mismatch GFRα4 siRNA with only one nucleotide exchange showed no significant effect (C). Densitometric quantification of at least three RT-PCRs of three different experiments was used to calculate the efficiency of GFRα4 siRNA (D). GFRα4 mRNA was downregulated from 86% (mismatch siRNA) to 3% after 48 hours and from 85% to 5% after 72 hours, respectively. *P < 0.01; **P < 0.001.

Figure 2.

 

In situ hybridization of cryosections of retinal spheres after transfection with GFRα4 siRNA. In contrast to nontransfected (A, C) and mismatch-transfected (E) retinal spheres, GFRα4 siRNA-treated cultures showed a dramatic decline in the number of GFRα4-expressing cells (B, D). In control cultures (A, C, E), GFRα4-positive cells were detected in the INL (arrowheads), within the IPL (star), and in nonorganized areas (arrows). In addition to the low level of GFRα4 expression, GFRα4 siRNA-treated spheres (B, D) appeared smaller than in control cultures (A, C, E). Scale bar: (A, B) 50 μm; (C, D) 100 μm.

Figure 3.

 

Analysis of growth (A), cell proliferation (B), and apoptosis (C) of retinal spheres in GFRα4 siRNA-treated and nontreated cultures. Transfection was carried out at culture day 2 for 12 hours. The growth curve of retinal spheres showed that GFRα4 siRNA-treated spheres were already smaller 24 hours after transfection than the corresponding control cultures (A). This reduction in size was more pronounced at day 6 (72 hours after transfection) and was constant for the remaining culture period. BrdU incorporation experiments revealed that the number of proliferating cells was reduced by more than 50% compared with control cultures (B). The percentage of apoptotic cells in transfected and nontransfected spheres showed no significant changes 96 hours after transfection (C; see Fig. 5 ). Comparison of mismatch siRNA– and GFRα4 siRNA-treated cultures did not show any difference with respect to proliferation. BrdU- and TUNEL-positive cells were analyzed by counting immunopositive cells in relation to DAPI-positive cells of at least nine retinal spheres of three different experiments. To compare the number of immunopositive cells in control and transfected retinal spheres, the same number of cells (DAPI-positive cells) per culture stage were used for accurate quantification. *P < 0.1; **P < 0.01; ***P < 0.001.

Figure 4.

 

Proliferation was significantly decreased in retinal spheres by siRNA-mediated gene silencing of GFRα4. Retinal spheres were transfected with GFRα4 siRNA (A, B) or with mismatch GFRα4 siRNA at culture day 2 (C, D). For BrdU incorporation experiments, retinal spheres were incubated in the presence of 50 μM BrdU 16 hours before harvesting; 48 hours after transfection, retinal spheres were harvested and cryosections were analyzed by immunostaining with anti-BrdU antibodies (A, C) and DAPI staining (B, D). Scale bar: 100 μm.

Figure 5.

 

Apoptosis was not affected in retinal spheres by siRNA-mediated gene silencing of GFRα4. Retinal spheres were transfected with GFRα4 siRNA (A, B) or with mismatch GFRα4 siRNA at culture day 2 (C, D). After 72 hours, retinal spheres were harvested and cryosections were analyzed by TUNEL (A, C) and DAPI staining (B, D). Scale bar: 50 μm.

Figure 6.

 

The number of Pax 6-expressing cells was decreased in GFRα4-deficient retinal spheres at culture day 6. Cryosections of nontransfected (A, B), GFRα4 siRNA-treated (C, D), and mismatch-treated spheroids (E, F) were double labeled with Pax 6-specific antibodies (A, C, E) and DAPI (B, D, F). Double-staining experiments demonstrated that the most Pax 6-positive cells in all investigated spheroids were localized in the INL surrounding the IPL (stars), whereas a smaller number was found within the IPL (stars) and in nonorganized areas (arrows). Here, Pax 6-positive cells can be identified as amacrine cells localized in the INL, ganglion cells in the IPL (stars), and probably blue cones and horizontal cells in nonorganized areas (arrows). However, it is obvious that the overall number of Pax 6-positive cells in GFRα4 siRNA-treated retinal spheres (C, D) was reduced compared with nontransfected (A, B) and mismatch-treated cultures (E, F). (For quantification of Pax 6-positive cells, see Fig. 7 .) R, rosettes. Scale bar: 100 μm.

Figure 7.

 

Quantification of Pax 6-positive cells in relation to the total cell number of DAPI-positive cells at culture stage 6. For calculating the percentage of Pax 6-positive cells, the overall number of DAPI-positive cells of at least nine cryosections of three different experiments was compared with the number of Pax 6-positive cells of the same sections. Here, 37% Pax 6-positive cells were detected in controls and 35% in siRNA mismatch-treated retinal spheres. In contrast, in GFRα4 siRNA-treated cultures, only 15% Pax 6-positive cells were found. To compare the number of immunopositive cells in control and transfected retinal spheres, the same number of cells (DAPI-positive cells) per culture stage were used for accurate quantification. **P < 0.01.

Figure 8.

 

The effect of GFRα4 siRNA transfection on the development of photoreceptors was analyzed by semiquantitative RT-PCR (A, B) or by immunostaining of rho4D2-positive cells (C). To distinguish individual photoreceptors, sequence-specific primers for different opsins were used to amplify opsin mRNA (A). Densitometric analysis showed that only blue and green opsin mRNA transcripts were significantly changed (B). Quantification of immunostained rods and green-sensitive cones by rho4D2 revealed an increase of rho4D2-positive cells in GFRα4 siRNA-transfected cultures after 6 days in culture (C; see Fig. 9 ). The percentage of rho4D2-positive cells was calculated by counting immunopositive cells in relation to DAPI-positive cells of at least nine sections from individual retinal spheres of three different experiments. To compare the number of immunopositive cells in control and transfected retinal spheres, the same number of cells (DAPI-positive cells) per culture stage was used for accurate quantification. For densitometric quantification, at least three RT-PCRs of three different experiments were used to analyze opsin mRNA expression. *P > 0.1; **P < 0.01; ***P < 0.001.

Figure 9.

 

Transfection of retinal spheres by GFRα4 siRNA increased the number of rho4D2-positive cells. Retinal spheres were transfected either with GFRα4 siRNA (A, B) or with mismatch GFRα4 siRNA at culture day 2 (C, D). After 72 hours, spheres were harvested and cryosections were stained with the antibody rho4D2 (A, C) and DAPI (B, D). Although retinal spheres appeared smaller, the number of rho4D2-positive cells was increased when compared with mismatch siRNA-transfected cultures. Scale bar: 100 μm.

The authors thank Elmar Willbold and Peter Wolf for helpful discussions, Jutta Huhn-Smidek and Meike Stotz-Reimers for expert technical assistance, and Tamara Kessel for the critical reading of the manuscript.