Iris-Derived Cells from Adult Rodents and Primates Adopt Photoreceptor-Specific Phenotypes

Abstract

purpose. The purpose of this study was to investigate the effects of various genes related to photoreceptor development on rodent and primate iris cells and the potential of iris cells as donor cells for retinal transplantation.

methods. Adult rat and monkey iris tissue were cultured in serum-free medium containing basic fibroblast growth factor. Gene deliveries of Crx, Nrl, NeuroD and some combinations (Crx-Nrl, Crx-NeuroD) were performed with recombinant retrovirus. Immunocytochemistry, Western blot analysis, RT-PCR, and intracellular recording were used to examine the expression of photoreceptor-specific phenotypes in the iris-derived cells after gene transfer, . Coculture of the iris-derived cells with embryonic retinal explant was conducted, to investigate the potential integration of these cells in coculture conditions.

results. Misexpression of Crx induced adult rat iris cells to express several photoreceptor-specific antigens and transcripts, such as rhodopsin, recoverin, cGMP-gated channel, arrestin, interphotoreceptor retinal-binding protein, rhodopsin kinase, and NeuroD. In primates, a combination of Crx and NeuroD was needed to induce monkey iris–derived cells to adopt photoreceptor-specific phenotypes. Furthermore, the photoreceptor-like cells derived from both rat- and primate-iris tissues showed rod photoreceptor-specific electrophysiological response to light stimuli after Crx and Crx-NeuroD gene transfer, respectively. The results further showed that iris-derived cells integrated in the developing host retina in coculture conditions.

conclusions. Adult iris-derived cultured cells of both rodents and primates expressed photoreceptor-specific phenotypes by inductions of transcription factors. These iris-derived photoreceptor-like cells have electrophysiological characteristics of rod photoreceptors. Furthermore, they can integrate in the developing retina under coculture conditions.

Retinitis pigmentosa (RP) is a heterogeneous group of hereditary disorders characterized by the progressive loss of photoreceptors. It is the most common inherited retinal dystrophy and the most prevalent cause of visual handicap among working populations in industrialized nations. ¹ RP is marked by the initial loss of rod function, followed by loss of cone function. ² Gene therapy is expected to be a good treatment strategy for recessive forms of RP. ^{3 4 5} Ribozyme therapy leading to a reduction in the amount of toxic proteins is regarded as a potential therapy for dominant negative type RP. ^{6 7} Besides these therapies that are protective of photoreceptor cells, regenerative treatment can be considered if stem cells or progenitor cells can be fated to develop into functional photoreceptors after transplantation and appropriately connect to the host retina. 

Iris pigmented epithelium (IPE) is composed of two cell monolayers: posterior and anterior. The posterior IPE cells are heavily pigmented and share a common developmental origin with the neural retina. ⁸ Although IPE cells can gain RPE properties in some culture conditions, ⁹ it has been found that iris tissue in the adult rat eye has the potential to express a photoreceptor-like phenotype after induction of Crx. ¹⁰ This observation is of potential importance for clinical application to RP, as autologous iris tissue is obtainable by peripheral iridectomy. It is unknown, however, how close the iris-derived photoreceptor-like cells with gene transfer are to natural photoreceptors, whether the results of this gene transfer strategy apply to primate iris tissue, and whether these iris-derived cells are capable of integrating in the host retina under retinal coculture conditions. 

Crx, NeuroD, and neural retina leucine zipper (Nrl) are the transcription factors known to be related to photoreceptor generation thus far. Crx is a homeobox gene specifically expressed in photoreceptors of the developing and mature retina, and it is essential for the differentiation and maturation of photoreceptors. ¹¹ NeuroD is a basic helix-loop-helix (bHLH) gene expressed in developing and mature rod photoreceptors, as well as in developing amacrine cells. This gene functions in the determination of cell fate and is essential for the preservation of rod and cone photoreceptors. ^{12 13} Nrl is a basic motif leucine zipper transcription factor preferentially expressed in rod photoreceptors, and it is indispensable for rod photoreceptors. ^{14 15}  

We have reported that cultured iris-derived cells have the potential to express neuronal markers and furthermore to express the photoreceptor-specific antigens, rhodopsin and recoverin, by Crx gene delivery. However, it remained unknown whether Crx-transfected, iris-derived cells are genuine photoreceptors, whether primate iris cells have a potential similar to that of the rodent iris cells, and whether transplanted iris-derived cells could integrate in the host retina. In the present study, we found that the Crx-transfected rat-iris–derived cells expressed not only rhodopsin and recoverin but also cGMP-gated channel, arrestin, interphotoreceptor retinal-binding protein (IRBP), rhodopsin kinase, and NeuroD, indicating that Crx evoked more photoreceptor-specific characteristics than does rhodopsin. We also demonstrated that Crx alone was not sufficient for the generation of iris-derived photoreceptor cells in primates, but rather a combination of Crx and NeuroD was needed. These iris-derived photoreceptor cells were revealed to have photoresponse to light stimuli. Furthermore, we demonstrated that iris-derived cells were competent to integrate and survive in the developing neural retina in coculture conditions. 

Materials and Methods

Animals

The use of animals in these experiments was in accordance with the ARVO Statement for Use of Animals in Ophthalmic and Vision Research and the Guideline for Animal Experiments of Kyoto University. All animal experiments were conducted with the approval of the Animal Research Committee, Graduate School of Medicine, Kyoto University. Three-week-old female DA rats and pregnant Fischer rats were obtained from Shimizu Laboratory Supplies (Kyoto, Japan), and 3-year- and 3- to 4-month-old cynomolgus monkeys were obtained from Shin Nippon Biomedical Laboratories (Kagoshima, Japan). 

Preparation of Tissue and Cell Culture

Rat and monkey iris tissues were prepared from 3- to 4-week-old female DA rats and 3-year- and 5- to 6-month-old cynomolgus monkeys, respectively. The procedure for cell culture was as described previously. ¹⁰ Briefly, iris tissue was treated with 1000 protein unit (PU)/mL dispase and 0.05% EDTA, and then it was plated on a laminin-coated chambered slide. The iris tissue was maintained in DMEM/F12/B27 (Invitrogen-Gibco, Rockville, MD) with 40 ng/mL basic fibroblast growth factor (bFGF; Genzyme/Techne, Minneapolis, MN) at 37°C. For cell differentiation, the cells were cultured in DMEM/F12/B27 with 1% fetal bovine serum and 10 ng/mL bFGF for 14 days. 

Preparation of Recombinant Retrovirus

We used a replication-incompetent retrovirus, CLIG, which directs the expression of green fluorescent protein (GFP) as a marker from the upstream long terminal repeat (LTR) promoter. ^{16 17} For construction of CLIG-NeuroD, CLIG-Nrl, and CLIG-Crx, mouse cDNAs for these factors were cloned into the EcoRI site of pCLIG, which directs expression of the cloned genes together with enhanced GFP from the upstream LTR promoter with cytomegalovirus enhancer. ¹⁸ Crx, NeuroD, or Nrl cDNA was inserted upstream of the internal ribosomal entry site (IRES) so that these genes were expressed bicistronically with GFP. For construction of CLIG-Crx-NeuroD and CLIG-Crx-Nrl, mouse NeuroD and Nrl were cloned into the BsrGI and ClaI sites of pCLIG-Crx, which are located in the 3′ region of the GFP gene, so that the GFP gene is fused in frame with NeuroD or Nrl. Retroviral DNAs were transfected (Lipofectamine; Invitrogen, Paisley, UK) into psi2mp34, an ecotropic packaging cell line, for ecotropic retrovirus, or both retroviral DNAs and pcDNA-VSV-G Env expression vectors, were transfected into 293GP cells (BD-Clontech, Palo Alto, CA), packaging cells expressing Gag and Pol, for pseudotype retrovirus. ¹⁹ The supernatant was collected as described previously. ^{18 20 21 22}  

Immunocytochemistry

Immunocytochemistry was performed as previously described. ^{21 22} We applied the following antibodies: rabbit anti-GFP (Molecular Probes, Eugene, OR) and mouse anti-rod-opsin RET-P1 (Sigma-Aldrich, St. Louis, MO). Cell nuclei were counterstained with 4′,6′-diamidino-2phenylindole (DAPI; Molecular Probes). Cultured cells were quantified by counting the number of marker-positive cells as a percentage of immunopositive cells per ocular grid area at ×20 magnification (four random areas per sample). 

Western Blot Analysis

After cultured iris-derived cells and neural retina were washed three times with cold PBS, they were solubilized in 100 to 300 μL lysis buffer (50 mM Tris [pH 7.5], 0.5 M NaCl, 1% NP-40, 1% sodium deoxycholate (DOC), 2 mM EDTA, and 0.1% SDS). After centrifugation at 10,000 rpm for 10 minutes, protein extracts were diluted with sample buffer (126 mM Tris HCl [pH 6.8], containing 20% glycerol, 4% SDS, 0.005% bromophenol blue, and 5% 2-mercaptoethanol) at a 1:1 ratio and boiled for 3 minutes. Protein extracts were then subjected to 4% to 20% Tris-glycine gel electrophoresis and transferred electrically to a polyvinylidene difluoride (PVDF) membrane (GE Healthcare, Buckinghamshire, UK). The membrane was then soaked in the blocking buffer (TBS containing 0.2% Tween-20 and 5% skim milk) for 1 hour at room temperature and incubated with primary antibody overnight at 4°C. We applied the following antibodies: rabbit anti-recoverin (the kind gift of James F. McGinnis and Rajesh J. Elias) at a dilution of 1:8000; or rabbit anti-G∂t1 (Santa Cruz Biotechnology, Santa Cruz, CA) at a dilution of 1:500. Horseradish-peroxidase–linked anti-mouse or rabbit Ig antibodies (1:10,000; GE Healthcare) were then added for 1 hour at room temperature. The membranes were washed with TBS containing 0.2% Tween-20, and the signals detected by a Western blot analysis system (ECL; GE Healthcare). 

RT-PCR for Photoreceptor-Specific Genes

Total RNA was extracted from cultured rat iris–derived cells, with or without gene transfer and adult rat neural retina. The Crx- and Crx-NeuroD-transfected rat-iris–derived cells were prepared with approximately 15% to 30% transfection efficiency in the area of monolayer culture. Complementary DNA was synthesized in vitro from 1.0 μg total RNA extracted from each sample. The reaction was conducted under the following conditions: denaturation at 96°C for 30 seconds, annealing for 30 seconds, and extension at 72°C for 1 minute. Primer sequences (forward and reverse), the length of the amplified products, the annealing temperatures, and the number of performed PCR cycles are shown in Table 1 . These primers were all designed based on the amino acid sequences of the rat genes. Sets of primers except NeuroD were designed to span at least one intron, so that the mRNA amplification could be distinguished from possible genomic contamination. For each condition, at least two independent experiments were performed. β-Actin was used as a loading control for the RT-PCR reactions. 

Intracellular Recording

Crx-transfected iris cells of rat and Crx-NeuroD–transfected iris cells of monkey in culture medium were penetrated by an ultrafine glass microelectrode with a preamplifier (MEZ-8300; Nihon Kohden, Japan) with resistance at 90 to 150 MΩ. Membrane potential of a single cell cultured for two weeks after transfection was recorded under an upright microscope (BX50WI; Olympus, Tokyo, Japan) using a 40× water-immersion objective lens with a chilled charge-coupled device (CCD) camera (C5985; Hamamatsu Photonix, Hamamatsu City, Japan) under infrared (>850 nm) illumination. Light stimulus was given by a red LED (broadband light-emitting diode, λ_max = 650 nm; DHR6610, Iwasaki Electronic, Japan). The intensity of the light stimulus was 54 or 74 μW/cm². 

Coculture with Fetal Retina

We prepared the rat-iris–derived cells infected with CLIG or CLIG-Crx in advance, and we washed them heavily with cold PBS. These cells were treated with 0.05% EDTA in PBS for 10 minutes at room temperature, and a cell suspension was prepared by gentle trituration with a micropipette in culture medium. The cell suspension containing the rat-iris–derived cells was first placed on the cell culture insert membranes (diameter 30 mm, pore size 0.4 μm; Millicell-CM; Millipore, Bedford, MA), and then the retinal explant culture of embryonic day (E)18.5 Fischer rat was performed for host retinas, as described previously. ^{18 22 23 24 25} Briefly, the neural retinas without pigment epithelium were isolated from other eye parts and placed on the membranes on which the cell suspension was paved with the ganglion cell layer upward; thus, cocultured cells would be located at the subretinal area as shown in Figure 7A . The inserts with neural retina were placed in six-well plates containing approximately 1 mL/well of medium. 

Results

Monolayer Culture of Iris Cells

We first prepared adult rat and monkey iris tissues, which did not include ciliary tissues. When iris tissue of adult rat and monkey was plated and maintained in serum-free medium containing bFGF, many pigmented cells, which were regarded as iris pigmented epithelial cells, first migrated out of the iris tissue and proliferated as a monolayer of cells (Figs. 1A 1C) . The iris tissue of adult monkey were plated and maintained. The cells gradually lost pigmentation after being cultured for a few days (Figs. 1B 1D) . As previously reported, some of these iris-derived cells differentiated to the cells expressing neurofilament, a neuronal antigen. ¹⁰  

Photoreceptor-Specific Phenotypes in Rat Iris Tissue

The rat-iris–derived cells without any gene transfer did not differentiate into cells expressing photoreceptor-specific antigens, in our culture method. Although it has been reported that Crx-transfected iris cells possibly express rhodopsin and recoverin, photoreceptor-specific antigens, ¹⁰ it is unknown whether other genes related to photoreceptor development have a function similar to that of iris cells. To determine whether rat-iris–derived cells have the potential to differentiate into photoreceptor-like cells after gene transfer, we first induced expression of Crx, NeuroD, and Nrl, genes known to be essential for photoreceptor generation (Fig. 2A) . Replication-incompetent retrovirus was applied to the iris-derived cells approximately 4 days after culture in serum-free medium, and cells were transferred into medium containing 1% FBS and 10 ng/mL bFGF, an environment that promotes retinal cell differentiation. ^{10 26} Ratios of rod-opsin immunoreactivity in the retroviral infected rat iris-derived cells were estimated in at least three independent samples. The average percentages of rod-opsin–positive cells per GFP-positive cells ± SEM are shown. Crx could induce the differentiation of iris-derived cells into a photoreceptor-specific phenotype (CLIG-Crx, 97.9% ± 1.5%, n = 3620), but none of the cells expressing NeuroD or Nrl were immunopositive for rod-opsin (NeuroD, 0%, n = 2062; Nrl, 0%, n = 2026; Figs. 2E 2F 2G 2H 2I 2J ). The cells infected with CLIG-Crx were small and round, characteristic of rod photoreceptors in monolayer culture, whereas cells infected with CLIG showed an elongated morphology and no rod-opsin immunoreactivity (CLIG, 0%, n = 2640; Figs. 2B 2C 2D ). To ascertain whether Crx responsiveness was relatively specific for iris-derived cells, we examined whether the cells derived from rat E17.5 telencephalon tissue, a population containing neural stem cells, would also respond to Crx-induction to express photoreceptor-specific phenotypes. Although we transferred the Crx gene into these telencephalon-derived cells, few of these cells acquired rod-opsin immunoreactivity (CLIG-Crx; brain; 1.9% ± 0.8%, n = 1140; Figs. 2N 2O 2P ), which may suggest that the embryonic neural stem cells existing in the embryonic telencephalon are intrinsically restricted in their response to Crx. 

We next examined the effect of coexpression of Crx and NeuroD or Crx and Nrl. For the coexpression of Crx with NeuroD or Nrl, NeuroD, and Nrl were fused with GFP, and each was coexpressed with Crx, so that GFP expression was restricted in the nuclei of infected cells. The iris-derived cells infected with CLIG-Crx-NeuroD or CLIG-Crx-Nrl were also immunopositive for rod-opsin and had small and round shapes (CLIG-Crx-NeuroD, 97.6% ± 1.6%, n = 712; CLIG-Crx-Nrl, 94.8% ± 1.8%, n = 548; Figs. 2Q 2R 2S 2T 2U 2V ). There were, however, no significant morphologic and immunocytochemical differences between the iris-derived cells infected with CLIG-Crx and those infected with CLIG-Crx-NeuroD or CLIG-Crx-Nrl. 

To investigate which percentage of photoreceptor-like cells were prepared for different culture periods (1 and 3 days and 1, 2, and 3 weeks) in serum-free medium and infected with CLIG-Crx before transfer. We found more than 90% of Crx-transfected cells were immunoreactive for rod-opsin regardless of the culture period (data not shown). This result suggests that we could expand the iris-derived cells for at least 3 weeks before transfer of the Crx gene into these cells. 

Immunoblotting of Photoreceptor-Specific Antigens in the Iris-Derived Cells after Gene Transfer

Immunocytochemical analysis revealed that the Crx-transfected rat iris–derived cells were labeled with monoclonal antibodies against rod-opsin RETP1. To determine further whether these cells express other phototransduction components, we tested anti-recoverin and anti-α subunits of transducin (G∂t1) tested by Western blot. For comparison, we estimated that from the retina 1.0 μg of protein/lane of the fractions were sufficient to immunodetect recoverin and G∂t1 (Figs. 3A 3B) . Recoverin is a calcium-binding protein that is involved in the inactivation of phototransduction, and G∂t1 is a photoreceptor-specific G protein molecule that activates cGMP-specific phosphodiesterase (PDE). ²⁷ We recognized similar patterns in the extracts (40 μg/lane) of the Crx-transfected iris-derived cells expressed recoverin and G∂t1, although we did not detect the same sized bands with equal amount of proteins prepared from the iris-derived cells infected with CLIG, CLIG-Nrl, and CLIG-NeuroD as the corresponding retinal extracts (Figs. 3A 3B) . These results indicate that at least two additional key components of the phototransduction cascade, recoverin and G∂t1, are expressed by the Crx-transfected rat-iris–derived cells. The reason for the lower level in the CLIG-Crx lane than in the neural retina was possibly contamination by non–virus-infected cells in the extracts. 

Rod Photoreceptor–Specific Gene Expression

To examine whether the iris-derived cells infected with CLIG-Crx express other photoreceptor-specific genes, we performed RT-PCR analysis. We prepared four different samples: rat-iris–derived cells expressing Crx or Crx-NeuroD, nontransfected cells, and adult rat retina as a positive control. The mRNA-expression of the genes rhodopsin, recoverin, cGMP-gated channel, S-antigen (arrestin), IRBP, and rhodopsin kinase, which are known as photoreceptor-specific genes, ²⁸ and NeuroD were tested by RT-PCR analysis. The iris-derived cells that had undergone Crx and Crx-NeuroD gene transfer expressed all transcripts examined, but the iris-derived cells without gene transfer expressed none (Fig. 4) . Crx gene transfer alone induced NeuroD expression, indicating that Crx induction is sufficient for these cells to express photoreceptor-specific transcripts. 

Photoreceptor-Specific Phenotypes in Primate Iris Tissue

We have shown that rat-iris–derived cells can differentiate into photoreceptor-like cells in vitro, but it is not clear whether cells derived from adult primate eyes have a similar ability. We examined iris tissue of adult cynomolgus monkey in a manner similar to the method used for rat iris tissue. 

To determine whether primate-iris–derived cells have the potential to differentiate into photoreceptor-like cells by gene transfer, we induced expression of Crx, NeuroD, and Nrl. Statistical data are shown as the mean ± SEM. Expression of rod-opsin was not induced in primate-iris–derived cells after transduction of any of these genes (CLIG, 0%, n = 2002; CLIG-NeuroD, 0%, n = 2134; CLIG-Nrl, 0%, n = 1170; CLIG-Crx, 0%, n = 2608; Figs. 5A 5B 5C 5D 5E 5F 5G 5H 5I 5J 5K 5L ). We next examined the coexpression of NeuroD or Nrl with Crx and we found that primate-iris–derived cells became rod-opsin immunoreactive after Crx-NeuroD gene transfer (CLIG-Crx-NeuroD, 49.8% ± 6.2%, n = 1302; Figs. 5M 5N 5O ), whereas none of the Crx-Nrl–transfected cells had rod-opsin immunopositivity (CLIG-Crx-Nrl, 0%, n = 548; Figs. 5P 5Q 5R ). Only the Crx-NeuroD–transfected cells had the small, round cell bodies characteristic of the rod photoreceptors in culture (Figs. 5M 5N 5O) . At least three independent samples of CLIG-Crx, CLIG-Crx-NeuroD, and CLIG-Crx-Nrl were tested, respectively. The data from the others were obtained with two samples in independent experiments. Thus, we observed photoreceptor-specific phenotypes in adult primate-iris–derived cells only by a combination of homeobox and bHLH expression. 

Hyperpolarization to Light Stimulus

We next examined whether the rat Crx-transfected iris-derived cells and the monkey Crx-NeuroD-transfected iris-derived cells have electrophysiological characteristics of rod photoreceptors. Intracellular recordings were made from the transfected iris cells of both animals, and light-induced membrane hyperpolarization was observed to light stimulus from an LED (Figs. 6B 6C) . The peak amplitude of the photoresponse increased depending on the light intensity. The resting potential of the rat CLIG-Crx-infected, iris-derived cells was −18.5 ± 4.0 mV (n = 25) in the dark, which was significantly different from that of CLIG-infected iris cells of rats with a deeper resting potential at −44.8 ± 6.7 mV (n = 17). The resting potential of the monkey CLIG-Crx-NeuroD-infected iris cells (−27.4 ± 2.8 mV, n = 26) was also different from that of the CLIG-infected iris cells (−65.6 ± 5.8 mV, n = 7). The photoresponse amplitude of the rat was approximately 1 mV, and that of the monkey was smaller (approximately 0.5 mV). The waveform of the photoresponse of rat sagged at brighter stimulus, a typical waveform observed in rod photoresponses, not like a cone response, which has a rectangular waveform. The light-sensitive cells of the rat were found in the cluster of cells with a small, round, or oval cell body rich in granular vesicles (Fig. 6A) . Those of the monkey had small, round, densely clustered cell bodies (data not shown). The intracellular recordings made from CLIG-infected, iris-derived cells of rat and monkey did not show any obvious photoresponse (Fig. 6D) . 

Potential of Iris Cells to Integrate and Survive in the Developing Neural Retina

To examine whether iris-derived cells could integrate in the host retina and exhibit a photoreceptor-specific phenotype under in vitro coculture conditions, we cocultured rat-iris–derived cells with or without Crx transfection into embryonic explanted retinas. The explanted neural retina initially consists of two layers: the ganglion cell layer and the ventricular zone (Figs. 7B 7C 7D 7E 7F 7G) . Within 2 weeks, however, almost all cells finished differentiation and formed three layers (Figs. 7H 7I 7J 7K 7L 7M) . ^{18 23} Photoreceptor cells differentiated without outer segments in this explant culture system. 

The donor cells were first infected with CLIG or CLIG-Crx and then placed in the subretinal areas (between the explanted neural retina and cell culture membrane; Millicell; Millipore) of E18.5 rat explanted retinas (Fig. 7A) , so that cocultured donor cells could be identified by GFP-immunostaining. Some of both the CLIG- and CLIG-Crx-infected cells integrated in the ventricular zone 5 days after the beginning of coculture, when the developing photoreceptor cells had not yet expressed rod-opsin (Figs. 7B 7C 7D 7E 7F 7G) . In an interesting observation, some of the Crx-transfected iris-derived cells had round immature morphologies and were divided into two cells after the beginning of coculture (Figs. 7E 7F 7G) . 

To determine the integration and survival of the cocultured donor cells in the differentiated retina, we examined the cocultured retinas at 14 days of explant culture. Some of both the CLIG- and CLIG-Crx-infected cells had migrated into the ONL from the subretinal space and survived for 14 days after the beginning of coculture (Figs. 7H 7I 7J 7K 7L 7M) . Most of the GFP-positive cocultured cells infected by CLIG or CLIG-Crx were located in the ONL, where rod photoreceptors exist, or in the subretinal area of the host retinas (CLIG and CLIG-Crx; 84.3% ± 22.8% and 92.7% ± 16.3%, respectively). Opsin-immunostaining was mainly located at the plasma membrane of photoreceptor cells, whereas GFP was localized in the cytoplasm of CLIG- and CLIG-Crx-infected cells. No clear evidence of any iris-derived retinal cells except photoreceptors was observed, although the number of cells that successfully integrated into the host explanted retinas was very small. 

Discussion

In this study, we showed that iris cells derived not only from rodents but also from primates could be effectively induced to adopt a photoreceptor-specific phenotype by retroviral infection in vitro. Furthermore, photoreceptor-like cells derived in this manner have a photoresponse to light stimuli and the potential to integrate and survive within the developing retina and demonstrate rod-opsin immunoreactivity. All these characteristics are essential for the possible application of these cells to the treatment of human retinal disease. 

Crx can bind to and transactivate regulatory elements in various photoreceptor-specific genes, including those encoding rhodopsin, IRBP, cone opsin, arrestin, and β-phosphodiesterase, ^{11 29} suggesting that Crx can directly control expression of these genes. We could predicted upregulation of rhodopsin, recoverin, rod cGMP-gated channel and S-antigen genes from a previous report of Crx ^{− / −} mice. ²⁸ However, we also observed upregulation of IRBP and rhodopsin kinase, the expression of which was not affected in Crx ^{− / −} mice. Furthermore, the NeuroD gene was upregulated. Considering that rat iris–derived cells having undergone Crx gene transfer expressed all the photoreceptor-specific genes we examined; and, regardless of whether a role for Crx had been demonstrated for the expression of those particular proteins, it is possible that these cells were induced to differentiate into authentic rod photoreceptors. 

There was a significant difference in the resting potentials between the transfected iris cells and nontransfected ones of the rat and monkey, possibly owing to the channel proteins, such as cGMP-gated channels or other ion channels on the differentiated cell membrane. The resting potential of non–gene-transfected iris cells of the monkey was quite low, probably reflecting a low density of the channel proteins. In fact, the resting potential of photoreceptor cells is normally very shallow at approximately −20 mV in the intracellular membrane voltage measurement. We, furthermore, observed the specific responses to light from gene-transfected, iris-derived cells of rats and monkeys although the photoresponse amplitude was at a low level. The much lower light sensitivity of the monkey gene-transfected iris cells could be due to lower density of the channel proteins expressed in monkeys. The low light sensitivity in both animals could also be due to the low concentration of chromophore molecules, such as 11-cis retinal, in the gene-transfected cells. Rod photoreceptor cells are much more sensitive to green light than to red. The red LED we used was ultrabright with a broad-band wavelength. The intensity of its light stimulus at a 500-nm wavelength was 6 nW/cm² (150 quanta/mm² per second), which should be strong enough to elicit a rod photoresponse approximately 300 times higher than that inducing a half-saturating response. Perhaps it would be possible to obtain much higher light sensitivity if the gene-transfected iris cell had enough chromophore molecules, normally produced in retinal pigmented epithelium and supplied to photoreceptor cells. 

We found that NeuroD expression was inducible by Crx gene transfer in rat-iris–derived cells. These cells differentiated into photoreceptor-like cells, whereas in primate cells, Crx did not induce NeuroD, and coexpression of these genes was needed for further differentiation into photoreceptor-like cells. Furukawa et al. ²⁸ showed that NeuroD expression is upregulated in Crx ^{− / −} mice, suggesting that Crx can work as a negative regulator of NeuroD expression. Conversely, the present findings suggest that Crx functions as a positive regulator of NeuroD in rat iris tissue but not in monkey iris tissue. Primate-iris–derived cells may be intrinsically restricted in their response to exogenous Crx. Homeobox and bHLH genes have been shown to be essential in directing the differentiation of specific retinal cell types in embryonic retinal progenitors in the retinal explant culture method. ^{21 22} Our data suggest that the combination of homeobox and bHLH genes is significant for the primate-iris–derived cells. 

Explanted retinas provide a good model for the primary study of retinal cell transplantation. ²⁵ We showed, for the first time, that cocultured iris-derived cells were competent to integrate, survive, and express a photoreceptor-specific antigen in a developing neural retinal explant; however, the number of incorporated grafted cells was very small. Transplanted adult rat-hippocampus-derived neural stem cells are known to integrate into neonatal normal retinas, ³⁰ mechanically injured adult retinas, ³¹ retinas exposed to transient ischemia, ³² genetically degenerated retina, ³³ and explanted developing retinas. ²⁵ Although this neural stem cell line is known as one of the cell lines most capable of incorporation in the host retina after transplantation, the cells have not become authentic retinal neurons and do not express any photoreceptor-specific phenotype even by Crx induction, ¹⁰ possibly because of intrinsic restrictions. A recent study of ours showed retinal progenitor cells gain migratory ability increasingly after a long culture period and many passages, which may be a result of the level of immaturity of the progenitor cells. ³⁴ Transplanted marrow stromal cells also have the potential to migrate into host retinas. ³⁵ Although the characters of transplanted cells are concerned with their migratory and survival abilities, the host condition would be another important factor for successful transplantations. For example, adult rat-hippocampus–derived neural stem cells did not migrate into the normal adult rat retina, ^{30 31} whereas they did integrate into injured or developing retinas. A recent study showed that the retinas of mice deficient in GFAP and vimentin provide a permissive environment for transplanted cells to survive and migrate. ³⁶ The integration of so few cocultured, iris-derived cells into the explant is possibly due to the conditions not only of the donor cells but also of the host retinas. Further study of appropriate host conditions for efficient transplantation should help us to understand the keys for clinical application of iris-derived cells. 

In conclusion, we produced photoreceptor-like cells from adult iris-derived cultured cells of both rodents and primates by induction of transcription factors. These iris-derived photoreceptor-like cells have electrophysiological characteristics of rod photoreceptors. Furthermore, they integrated in the developing retina in coculture conditions. 

 Supported by the Advanced and Innovational Research program in Life Sciences from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), the Japanese Government.

 Submitted for publication September 18, 2004; revised February 7 and April 21, 2005; accepted July 11, 2005.

 Disclosure: T. Akagi, None; J. Akita, None; M. Haruta, None; T. Suzuki, None; Y. Honda, None; T. Inoue, None; S. Yoshiura, None; R. Kageyama, None; T. Yatsu, None; M. Yamada, None; M. Takahashi, 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.

 

List of Primers and Their Respective Sequences used for RT-PCR

Gene	Primer Sequence	Product Size (bp)	Annealing Temp. (°C)	Cycle
β actin
Forward	5′-ATCGTGGGCCGCCCTAGGCACCAG-3′
Reverse	5′-ATGTCACGCACGATTTCCCTCTCAGC-3′	536	68	26
Rhodopsin
Forward	5′-TATGTGCCCTTCTCCAACATCACG-3′
Reverse	5′-GCCATGACCCAGGTGAAGGCCAC-3′	464	68	30
Recoverin
Forward	5′-ATGATCAAGCCTGAGGATGTGAAG-3′
Reverse	5′-AGGTTCAAACTGGATCAGTCGCAG-3′	177	60	30
Rod cGMP-gated channel
Forward	5′-ATGAACTCCAGTCTGATTACCTAG-3′
Reverse	5′-GTCTCTGTCCTCTGGAAGAACTC-3′	301	60	30
Rod arrestin
Forward	5′-GTGCGGTTACTGATCCGGAAGGTG-3′
Reverse	5′-GTTATTGGTCACAGTCACGGTCAC-3′	181	60	30
Interphotoreceptor retinal-binding protein (IRBP)
Forward	5′-ACTTTGGACACACTCGCAGCTCAC-3′
Reverse	5′-ATGGTGATATAGAGATGCGTGTCG-3′	214	60	30
Rhodopsin kinase
Forward	5′-CACATCTACAATGTGGATGAGGAG-3′
Reverse	5′-GGGGTCCCTGCATAGCCCTTGGTC-3′	236	60	30
NeuroD
Forward	5′-CTACTCTAAGACACAGAAGCTGTC-3′
Reverse	5′-GGAAGACATGGGAGCTGTCCATGG-3′	333	60	30

Figure 7.

 

Survival and integration after the rat-iris–derived cells were cocultured in the subretinal space of the retinal explant. (A) Scheme illustrating the retinal coculture system. The iris cell suspension after viral infection was placed between membrane and neural retina, regarded as the subretinal space. (B–M) The retinal sections obtained 5 (B–G) and 14 (H–S) days after beginning of coculture of the CLIG-(B–D, H–J) and CLIG-Crx-infected, iris-derived cells (E–G, K–M). (C, F, I, L) sequential combinations of confocal images of cocultured GFP-positive cells in (B), (E), (H), and (K). (D, G, I, M) Bright-field view merged with (B), (E), (H), and (K). Both the CLIG- and CLIG-Crx-infected cells integrated in the developing retinas from the subretinal area without any destruction of layer structure. The Crx-transfected iris clone divided into two cells after the beginning of coculture and displayed round progenitor-like morphology (E–G). Red: rod-opsin–positive cells; green: GFP-positive cocultured cells. GCL, ganglion cell layer. VZ, ventricular zone. INL, inner nuclear layer. ONL, outer nuclear layer. Scale bars, 50 μm.

Figure 1.

 

Monolayer culture of rat and monkey iris–derived cells. Rat iris tissue cultured in serum-free medium for 1 (A) and 4 (B) days. Monkey iris tissue cultured in serum-free medium for 3 (C) and 7 (D) days. The iris cells derived from both rat and monkey proliferated well as a monolayer of cells, although the proliferation rate of monkey iris cells was slower than that of rat iris cells. Scale bar, 100 μm.

Figure 2.

 

The potential to adopt photoreceptor-specific phenotypes in the rat-iris–derived cells. (A) Viral constructs used to express NeuroD, Nrl, Crx, Crx-NeuroD, and Crx-Nrl. (B–J) The rat-iris–derived cells infected with the control retrovirus CLIG (B–D), CLIG-NeuroD (E–G), and CLIG-Nrl (H–J) showed no rod-opsin immunoreactivity. (K–V) The rat-iris–derived cells infected with CLIG-Crx, CLIG-Crx-NeuroD, and CLIG-Crx-Nrl expressed rod-opsin (K–M, Q–V), whereas the Crx-transfected cells derived from E17.5 rat telencephalon did not express rod-opsin (N–P). Red: rod-opsin–positive cells; green: viral-infected cells; blue: nuclei stained with DAPI. Note that the GFP-immunolabelings of the cells infected with CLIG-Crx-NeuroD and CLIG-Crx-Nrl were localized in the nuclei, because the GFP was fused with NeuroD or Nrl proteins in these cases (Q–V). Scale bar, 50 μm.

Figure 3.

 

Western blot analysis of photoreceptor-specific antigens in iris-derived cells with gene transfer. (A, B) Cell lysates (40 μg per lane) from CLIG-Crx-, CLIG-, CLIG-Nrl-, CLIG-NeuroD(ND)-infected, iris-derived cells and normal rat neural retinal protein extracts (1 μg per lane) were probed with recoverin (A) and G∂t1 antibodies (B). Arrows at right: electrophoretic mobility of recoverin (A) and G∂t1 (B). (A, B) molecular mass is shown at right in kilodaltons.

Figure 4.

 

Expression of photoreceptor-specific genes in iris-derived cells. RT-PCR analysis of total RNA isolated from the rat-iris–derived cells infected with CLIG-Crx (lane 1), CLIG-Crx-NeuroD (lane 2), and without gene transfer (lane 3), when cultured in a differentiating environment, and from adult rat retina (positive control, lane 4). Crx- and Crx-NeuroD-transfected cells both expressed rhodopsin, recoverin, rod cGMP-gated channel, rod arrestin, interphotoreceptor retinal-binding protein (IRBP), rhodopsin kinase, and NeuroD. The β-actin fragment was amplified as a control for the quality of cDNA.

Figure 5.

 

Induction of photoreceptor-specific phenotypes in primate-iris–derived cells. The cells were subjected to immunocytochemistry after culturing in a differentiating environment for 2 weeks after viral infection. The cells infected with CLIG (A–C), CLIG-NeuroD (D–F), CLIG-Nrl (G–I), CLIG-Crx (J–L), and CLIG-Crx-Nrl (P–R) had no immunoreactivity for rod-opsin. Some of the cells infected with CLIG-Crx-NeuroD expressed rod-opsin (M–O). Red: rod-opsin–positive cells; green: viral-infected cells; blue: nuclei stained with DAPI. Scale bar, 50 μm.

Figure 6.

 

Intracellular recordings of a rat Crx-transfected and a monkey Crx-NeuroD-transfected iris cell and their morphology. (A) Micrograph of the cell located at the tip of the glass microelectrode in a cluster of cells in a culture chamber. The cultured cells were penetrated by ultrafine glass microelectrodes while the cell morphology was viewed with an upright microscope with a chilled CCD camera under infrared (>850 nm) illumination. The shape of the cell somata was small and round. (B) Light-induced hyperpolarization of a rat cell to red light stimulus from an LED (light emitting diode, λ_max = 650 nm) with 500 ms duration. The intensities of the photostimuli for waveforms 1 and 2 were 54 and 74 μW/cm², respectively. The response waveforms were averaged 12 and 11 times in waveforms 1 and 2, respectively, due to its low light sensitivity. (C) Similarly obtained light-induced hyperpolarization of a primate cell. The response waveforms were averaged 12 and 14 times in waveforms 1 and 2, respectively. (D) Intracellular recordings of CLIG (control virus)-infected iris cell of rat (1) and monkey (2). The response waveforms were averaged 39 and 11 times in waveforms 1 and 2, respectively. No obvious responses were detected.

The authors thank Anand Swaroop for Nrl cDNA, Constance L. Cepko and Takahisa Furukawa for Crx cDNA, and Tal Kafri (University of North Carolina) for pcDNA-VSV-G expression vector.