Characteristics of Progenitor Cells Derived from Adult Ciliary Body in Mouse, Rat, and Human Eyes

Abstract

purpose. To isolate and characterize progenitor cells derived from adult mammalian ciliary body.

methods. The authors isolated progenitor cells from the ciliary body of adult mice, rats, and human cadaver eyes and determined quantitative growth characteristics of groups of progenitor cells called neurosphere (NS) cells, including individual cell diameter, NS diameter, percentage of NS-forming cells, and cell number per eye in mouse, rat, and human eyes. The immunolabeling and ultrastructure of NS cells were investigated by confocal and transmission electron microscopy.

results. Average diameters of individual cells and neurospheres after 1 week in culture were similar in mice, rats, and humans (cell diameters: 22 ± 1.1, 21 ± 0.3, 25 ± 0.4 μm; NS diameters: 139 ± 22, 137 ± 9, 141 ± 11 μm, respectively). Mean numbers of cells per NS were estimated to be 1183 in mice, 5360 in rats, and 685 in humans. Molecules that were identified by immunolabeling in NS cells included nestin, Chx-10, vimentin, GFAP, and Pax-6. Thy-1 was expressed in some NS cells. Ultrastructurally, NS cells displayed abundant rough endoplasmic reticulum and many cellular processes but no characteristics of mature retinal neurons or glia.

conclusions. Progenitor cells from adult mammalian ciliary body have significant, but limited, proliferation potential and express markers characteristic of other progenitor cells and seen during early retinal development. The ciliary body could be a source of cells for transplantation in experimental rodent eyes and for autotransplantation in human eyes.

Glaucoma and other optic nerve diseases of hereditary and traumatic etiology cause loss of vision resulting from irreversible damage to retinal ganglion cells. In this study, we isolated progenitor cells from the ciliary body of adult mice, rats, and human cadaver eyes by modifications of methods reported by Tropepe et al. ¹ Research to generate new ganglion cells might allow restoration of visual function in patients with glaucoma, the second leading cause of blindness worldwide. ² Although the adult mammalian central nervous system, including the retina, is generally considered to have minimal regenerative capacity, potential sources of replacement neurons for the retina have been studied by several laboratories. Embryonic tissues may proliferate significantly more than adult cells, but they must change from an undifferentiated state to the desired tissue and cell type. Further, embryonic tissues must be acquired from a donor source that differs from the recipient immunologically, leading to the possibility of graft rejection. Pieces of retina and dissociated retinal cells have been transplanted into the retina. ^{3 4} Instead of orderly incorporation into retinal structure, ^{5 6 7 8} undesirable rosette formation occurs, a result that clearly prevents the formation of proper connections with host cells. Correct orientation is better preserved with the transplantation of retinal sheets, but their size is of necessity limited and the acquisition of retinal tissues for clinical use is problematic. ⁹  

It was previously thought that central nervous system neurons were not replaced during adult life, but the brain has resting cells that multiply and produce new neurons. ^{10 11 12 13 14} One review defines stem cells as multipotential (capable of producing differentiated cells of different types) and self-renewing for the lifetime of the organism. ¹⁵ Progenitor cells are defined as lacking in one or more of these characteristics. The removal of stem or progenitor cells from the organism and their multiplication in culture is facilitated by exposure to epithelial and fibroblast growth factors. ¹⁶ Although much of the work on neuronal stem cells was accomplished in rodents, monkey ¹⁷ and human brains ^{18 19} have cells with similar potential. 

The ciliary body, iris, and retina of the eye are derived from the neural ectoderm of the optic vesicle. Some fish produce new retinal tissue throughout life from the peripheral retinal zone. Recently, progenitor cells have been identified in the adult mammalian ciliary body, ¹ opening the possibility that persons could be the source of their own retinal replacement neurons, with reduced probability of immune rejection. It may be easier to reactivate the genetic program typical for retinal cells in progenitor cells from ciliary body than from embryonic stem or germ line cells. Ciliary body progenitor cells may have less proliferation capability than stem cells. If their proliferation capacity is sufficient to produce enough new cells but less than would produce uncontrolled, tumorous growth, this would be an additional advantage. 

In the present study, we examined detailed features of progenitor cells generated from ciliary body of mice, rats, and humans that have not previously been reported. This included the size, ultrastructural morphology, growth characteristics, and expression of a wider range of molecular species. 

Methods

Animals

FVB.Cg-Tg(GFPU) 5Nagy mice and C57BL/6 mice (Jackson Laboratory, Bar Harbor, ME) 6 to 8 weeks of age and male Wistar rats weighing 250 g were treated in accordance with the ARVO Statement for Use of Animals in Ophthalmic and Vision Research, in protocols approved and monitored by the Animal Care Committee of the Johns Hopkins University School of Medicine. Animals were housed with a 14-hour light/10-hour dark cycle, with standard chow and water provided ad libitum. 

Human Eyes

Eyes were obtained from eye banks after informed consent was obtained and in accordance with the tenets of the Declaration of Helsinki and with the approval of the Joint Committee on Clinical Investigation, The Johns Hopkins University School of Medicine. In all, six eyes from five persons were used; each person had no known history of eye disease. Donor age ranged from 78 to 85 years (mean, 81.4 years), and time from death to enucleation varied from 3 to 8 hours, with each eye divided for ciliary body culture within 24 hours of death. 

Isolation of Progenitor Cells

Ciliary body cells were isolated by a published method ¹ from adult mice expressing green fluorescent protein (GFP) with chicken β-actin promoter, from adult male albino Wistar rats, and from human cadaver eyes. Briefly, eyes were removed after sacrifice in mice by cervical dislocation and in rats by sodium pentobarbital overdose. Tissues were dissected in artificial cerebral spinal fluid (ASCF; 124 mM NaCl, 5 mM KCl, 1.3 mM MgCl₂, 2 mM CaCl₂, 26 mM NaHco ₃, and 10 mM d-glucose). Ciliary bodies from human eyes received within 24 hours of death from eye banks were dissected in the same medium. The ciliary body was separated, treated with dispase for 10 minutes at room temperature, and treated with a mixture of 1.33 mg/mL trypsin, 0.67 mg/mL hyaluronidase, and 0.2 mg/mL kynurenic acid in modified ACSF with 3.2 mM MgCl₂ and 0.1 mM CaCl₂ for another 10 minutes at 37°C. Cells were collected by centrifugation at 1500 rpm for 5 minutes. The enzyme mixture was replaced with serum-free medium (SFM; modified DMEM/F12 media) containing 1 mg/mL trypsin inhibitor. Cells were triturated and centrifuged again. The SFM–trypsin inhibitor solution was replaced with growth factor medium (GFM; SFM containing 10 ng/mL basic fibroblast growth factor [FGF2], 20 ng/mL epithelial growth factor [EGF], and 2 μg/mL heparin [Sigma-Aldrich, St Louis, MO]). 

Cells were cultured at a concentration of less than 50 cells/μL in GFM at 37°C and 5% CO₂ for 7 to 10 days, by which time they had formed round, adherent collections of cells known as neurosphere (NS) cells. 

Quantitative Analysis of Neurospheres

The number of ciliary cells isolated per eye and the mean number of cells in a dissociated neurospheres after 1 week in culture were estimated using a hemacytometer. Cell viability was determined with the trypan blue exclusion assay. Mean cell counts and diameter measurements made with an eyepiece micrometer were averaged from multiple preparations. The proportion of ciliary cells capable of forming an NS was determined by counting the number of neurospheres that grew in GFM after 1 week in wells containing 100, 200, 500, 1000, 2000, and 4000 cells (four replicates per preparation; multiple preparations per species). 

The mean number of cells within a typical NS was obtained from cell counts of 10 dissociated neurospheres per preparation. Mean diameters were obtained from 25 cells or 25 neurospheres per preparation at 1-day (mice and human) or 2-day (rat) intervals to generate the growth curves. Growth curves were obtained from cells grown through three cycles of growth and dissociation. Photographs of cells and neurospheres were taken (TE2000-U [Nikon, Tokyo, Japan] and Spot RT [Image Systems, Columbia, MD] cameras) at 200× magnification. 

Preparation of Neurospheres and Eye Tissues for Immunohistochemical Study

Neurospheres collected after 7 days of growth were fixed with 2% paraformaldehyde for 15 minutes. Eye tissues were prepared by a modified method, as previously described. ²⁰ Briefly, mice were anesthetized with a mixture of ketamine, xylazine, and acepromazine at 75, 13 and 2 mg/kg, respectively, and underwent intracardiac perfusion with 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.2) at a rate of 10 mL/min for 12 minutes. Then eyes were processed by serial exposure to 2% paraformaldehyde in 5% sucrose in 0.1 M phosphate buffer (pH 7.2) for 60 minutes, graded sucrose in phosphate buffer at 5%, 10%, 12.5%, and 15% for 30 minutes each, and 20% sucrose in phosphate buffer overnight. Neurospheres and eye cups were embedded in a 1:2 mixture of OCT compound (Sakura Finetek USA, Torrance, CA) and 20% sucrose in phosphate buffer. Cryosections 8 μm thick were collected onto slides (SuperFrost Plus; Fisher Scientific, Pittsburgh, PA) and stored at –80°C before immunolabeling. 

Immunohistochemistry

For immunohistochemical analysis, frozen sections of neurospheres and ocular cross-sections were blocked with 10% normal serum (same species as secondary antibody) for 1 hour before incubation with primary antibodies overnight at 4°C with 0.3% Triton X-100. Primary antibodies used were rabbit anti-nestin (1:100; Abcam, Cambridge, UK), sheep anti-Chx10 (1:150; Exalpha Biologicals, Watertown, MA), goat anti-Brn3b (1:100; Santa Cruz Biotechnology, Santa Cruz, CA), rabbit anti-Thy1 (1:100; Santa Cruz Biotechnology), rabbit anti-Pax6 (1:200; Chemicon, Temecula, CA), rabbit anti-vimentin (1:500; Abcam), rabbit anti-GFAP, rabbit anti-calbindin (1:500; Chemicon), rat anti-opsin (1:10,000; Sigma-Aldrich), and rabbit anti–Ki-67 ²¹ (1:750; Vector Laboratories, Burlingame, CA). Sections were washed and incubated with 1:500 Alexa Fluor secondary antibodies (Molecular Probes, Eugene, OR) at room temperature for 1 hour and were mounted with Vectashield with DAPI (Vector Laboratories). Negative control experiments included nonimmune serum of the same species as the primary antibody, at the same protein concentration and incubation in buffer alone. Tissue sections were imaged (LSM 510 Meta confocal camera; Carl Zeiss Inc., Thornwood, NY) at 63× using the 405 nm and 543 nm lasers to excite DAPI and Alexa Fluor 546. Standard filter sets for DAPI and rhodamine were used. 

Transmission Electron Microscopy

To minimize tissue loss during processing, we centrifuged groups of neurospheres after 3 and 7 days in culture, before and after each of the after processing steps: fixation in 4% paraformaldehyde with 1% glutaraldehyde in 0.1 M phosphate buffer, pH 7.2, postfixation in 1% buffered osmium tetroxide, dehydration in graded ethanols, and embedding in LX112 epoxy resin (Ladd Research Industries, Burlington, VT). Blocks were thin-sectioned and stained by 5% uranyl acetate and Reynolds lead citrate before examination in the transmission electron microscope (Hitachi H7600; Hitachi High-Technologies Corp., Tokyo, Japan). 

Statistical Analysis

One-way analysis of variance (ANOVA) and t tests were performed for evaluation of study results. 

Results

Progenitor Cell Growth into Neurospheres

Average diameters of ciliary cells immediately after dissociation and of NS cells after 1 week in culture were similar in mice, rats, and humans (Table 1) . Experiments in which cells were cultured after successive dilution showed that 1 cell in 101 ± 15 (SD) among those isolated from the ciliary body developed into an NS cell (Table 1) . This varied from 1 in 40 ± 11 cells in human eyes to 1 in 117 ± 42 for rat eyes. 

For growth curves, cells were cultured for 7 to 9 days before dissociation and reculture, with diameter measurements every 2 days in mice and humans and every third day in rats. Diameters of 25 NS cells were measured at each time point. Growth of NS cell diameter was logarithmic in each passage, with correlations to a logarithmic fit varying from R ² = 0.84 to 0.98 (Fig. 1) . Diameter of NS cells in the third passage were smaller in all three species than in the first passage, and in mice there was a monotonic fall in diameter from the first growth period to the third. At the end of 1 week, the estimated number of cells within a NS was 1183 cells in mouse NS, 5360 cells in rat NS, and 685 cells in human NS. The average diameter of a NS was similar across species at 1 week, indicating a different packing density of cells within neurospheres. In a few NS cells, histologic evaluation showed areas without viable cells within the interior of the NS. 

P1 represents the first week of NS growth, P2 represents growth after dissociation of the 1-week NS into individual cells and a further week in culture, and P3 represents the third week of growth after second dissociation. Numbers of initial dissections (sets of eyes) for these data are as follows: mouse, P1, n = 6; P2, n = 6; P3, n = 2; rats: P1, n = 4; P2, n = 4; P3, n = 3; humans: P1, n = 8; P2, n = 6; P3, n = 6. 

We hypothesize that each NS begins with a single progenitor ciliary cell and grows through binary division. If so, each initial cell and its daughters would undergo 10 divisions to reach the number of cells seen per NS before reaching the limiting size that we observed at 7 to 10 days. An alternative hypothesis is that one progenitor cell gives rise to sequential daughter cells but that the daughter does not divide further. This scenario is not compatible with the observed time periods and numbers of cells, nor is it compatible with the Ki-67 data. 

Neurospheres varied in diameter (Table 1) , and the degree of pigmentation also varied between cells and within neurospheres from human eyes (Fig. 2) . Mouse and rat eyes studied in these initial experiments came from albino animals. If the melanin present in one original progenitor cell was the only melanin available, all human neurospheres would be essentially unpigmented. Hence, continued synthesis of pigment must occur during NS enlargement, along with cell multiplication. 

Figure 2Ashows individual cells from dissociated neurospheres. The cells have various amounts of melanin. Figure 2Bshows a human NS at lower power, with variations in pigmentation among the cells (bar represents 50 μm). 

Immunolabeling of Neurospheres and Ocular Tissues

Immunolabeling of Ki-67 was performed to investigate the proliferative capacity of the NS. After 3 days in culture, 60.1% ± 11.6% of cells were Ki-67 positive, whereas at day 7, 21.3% ± 8.7% were positive for Ki-67 (Fig. 3 ; P < 0.001, one-way ANOVA, Student-Newman-Keuls test). Neurospheres maintained in culture for 10 days showed no detectable cells positive for Ki-67. 

Percentages of cells in mouse neurospheres at 3 and 7 days after initiation of floating culture conditions show that most express the marker Ki-67 at 3 days, indicating active proliferation; by 7 days, only approximately 20% are Ki-67 positive. At 10 days in culture, few cells remained Ki-67 positive (*P < 0.001 vs. day 3). 

We found that 87.1% ± 5.5% of NS cells expressed nestin, a molecule considered characteristic of progenitor and stem cells (Table 2 ; Fig. 4A1 ). Many NS cells also expressed Chx10 and Pax 6 (Figs. 4B1 4C1) . In addition, 83.5% ± 8% of NS cells labeled positively for vimentin, 82.9% ± 7.4% expressed Thy-1, and 97.4% ± 1% were positive for GFAP (Table 2 ; Figs. 4D1 , 5A1 5B1 ). We did not detect labeling for calbindin, opsin, and Brn3b, molecules characteristic of mature horizontal cells, rods, and retinal ganglion cells, respectively (Figs. 5H1 5I1 5J1) . 

Immunolabeling of the ciliary body and the retina was also performed in frozen sections of normal mouse eyes with the same antibodies tested with neurospheres. We found no labeling for nestin in either. Ciliary body epithelial cells labeled positively for Chx10, Pax6, and vimentin (Fig. 4) . Other markers, such as Thy1, GFAP, calbindin, opsin, and Brn3b, were not expressed in the ciliary body (Fig. 5) . Labeling occurred in expected retinal layers for Chx10, Pax6, vimentin, Thy-1, GFAP, calbindin, opsin, and Brn3b (Figs. 4 5) . 

Immunolabeling is shown for neurospheres (left column), retina (middle column), and ciliary body in situ (right column) with each antibody. Neurospheres are strikingly positive for nestin, but ciliary body and retina have no nestin labeling (A1–A3). Labeling for Chx10 is present in neurospheres and in bipolar cells of the mature retina and in ciliary body (B1–B3). Pax-6 is identified in neurospheres, in the inner nuclear layer of adult retina, and in ciliary epithelium (C1–C3). Neurospheres are positive for vimentin, as are Müller cells of mature retina and ciliary epithelium (D1–D3). Neurospheres at 7 days in culture showed a minority of cells positive for Ki-67, but ciliary body and retina showed no staining for Ki-67 (E1–E3; DAPI counterstain; scale bar in A1 represents 40 μm; magnification is identical in all images). 

Immunolabeling was as shown in Figure 4with neurospheres (left column), retina (middle column), and ciliary body (right column). Neurospheres showed extensive Thy-1 labeling in the ganglion cell layer of the retina and in cells of the inner portion of the inner nuclear layer, perhaps representing displaced ganglion cells, but none was seen in ciliary body (A1–A3). Neurospheres were positive for GFAP. Cells labeled in normal retina appeared to be Müller cells, but no label was seen in ciliary body (B1–B3). Calbindin labeling was absent in neurospheres and ciliary body but present in the outer plexiform layer of the retina (C1–C3). Opsin was absent in neurospheres and ciliary body, but outer segments in the retina were labeled (D1–D3). Brn-3b did not label neurospheres or ciliary epithelium but did label ganglion cells in the retina (E1–E3; DAPI counterstain; scale bar in A1 represents 40 μm; magnification is identical in all pictures). 

Transmission Electron Microscopy of Neurospheres

Ultrastructural appearance of day 3 mouse neurospheres showed protrusions from the cell membrane and a prominent nucleolus (Fig. 6A) . Developing groups of cells clumped together in neurospheres were present, but a few individual cells were not in contact with the neurospheres. The latter showed a variety of morphologies, including some that formed closed tubular profiles, simulating capillary vessel structure (not shown). After 7 days in culture, only large groups of connected cells in neurospheres were seen. Cells within these neurospheres were closely apposed to each other (Fig. 6B) . The cells had abundant rough endoplasmic reticulum, indicative of active intracellular synthesis and occasional lysosomes. Junctional complexes developed between cells in the neurospheres (Fig. 6C) . No cell processes resembled dendrites or axons, nor did any specializations occur such as receptor outer segments or synapses. Occasional cells within the interior of neurospheres showed features compatible with apoptotic cell death, including clumped densification of nucleus, blebbing of nuclear membrane, and loss of internal organelles (Fig. 6D) . 

Transmission electron microscopy of NS cells on day 3 (Fig. 6A)showed cells with many processes and a prominent nucleolus. By day 7, cells within neurospheres were closely apposed to each other and had abundant rough endoplasmic reticulum and lysosomes (Fig. 6B) . In some cells, we found junctional complexes between cells (Fig. 6C , arrows). We also found rare apoptotic cells (Fig. 6D)within the groups of cells. Scale bars indicate 2 μm (Figs. 6A 6B 6D)and 0.3 μm (Fig. 6C ; uranyl acetate/lead citrate staining). 

Discussion

A small number of cells in the adult ciliary body are now known to have the capacity to proliferate when subjected to appropriate stimuli. ^{1 22 23} Tropepe et al., ¹ Ahmad et al., ²² and Kohno et al. ²⁴ found that cells isolated from the ciliary body of adult mice and rats proliferate when stimulated with epidermal and fibroblast growth factors, though they estimated that only about 1 in 100 to 1 in 500 ciliary body cells have this proliferative potential. ^{1 25} We confirm that no more than 1% to 2% of ciliary epithelial cells generate neurospheres in mice and rats, and we provide further detailed information on similar behavior in the ciliary body of human postmortem eyes, as initially reported by Coles et al. ²⁶ For all three species, growth of neurospheres follows a logarithmic multiplication rate. The growth rate of neurospheres was faster in the first passage than in the second and third growth phases after dissociation of the primary cultured cells, suggesting that these cells have a relatively limited proliferative capacity. Although investigations have been able to continue reforming neurospheres for several passages in mice ¹ and human ciliary cells, ²⁶ we found, as did Coles et al., ²⁶ that growth was less vigorous with multiple passages. Indeed, we determined that after the second passage, the total number of cells produced at each new passage was no greater than the starting number after the first passage. Although this indicated that a small number of cells retained proliferative potential after multiple divisions, the major increase in cell number occurred in the first burst of activity. 

Despite the limited proliferation capacity of ciliary-derived progenitor cells, these cells seem to have sufficient capacity to generate material for reimplantation in future experiments. Our proliferation data are compatible with the hypothesis that soon after dissociation from the ciliary body, most cells in neurospheres divide to produce daughter cells, whereas after 7 days in culture, only one fifth have proliferative capacity, as judged by Ki-67 labeling. This reduced proliferative capacity may decrease the chance of tumorigenesis after transplantation and, hence, is a potential advantage. The presence of apparently apoptotic cells in neurospheres indicates a balance between the production of new cells and the death of existing cells. 

After 7 to 10 days in culture, each ciliary-derived progenitor had produced from 1000 to 5000 daughter cells, with the estimated number of cells from rat and mouse eyes larger than that from human eyes. This may be an inherent species difference, or it may derive from the fact that rodent eyes were dissected from freshly killed animals; human tissues were from eye bank material obtained several hours after death. Furthermore, relative age may play a role in this difference because the rats and mice were young but human eye bank eyes were from elderly adults. It is interesting to calculate how many potential new cells could be obtained and produced for transplantation from human eyes. Cyclectomy of a living eye could yield a specimen as large as 5 mm (approximately 10% of the ciliary body circumference) without the loss of structural integrity of the eye for later reimplantation of new cells. Based on data from our eye bank eyes, we estimate that 360,000 progenitor cells are present in a 5-mm cyclectomy specimen because 1 in 40, or a total of 9000, progenitors would be obtained, each producing a NS with approximately 685 cells. This would give a total of more than 6 million injectable cells for autotransplantation in a human patient. Given that these cells would be reimplanted into the same person, it is likely that the immune reaction would be negligible unless the conditions of culture altered the histocompatibility status of the cells during preparation. This approach could serve partially to repopulate the neurons in a human eye or to provide continuous delivery devices for molecules produced spontaneously by the transplanted cells. In addition, it is conceivable that the cells could be genetically manipulated before transplantation so that they would produce therapeutic molecules for long-term delivery. ²⁷  

Without artificial stimulation by growth factors, it has been thought, the adult ciliary body and retina do not have cells that show features of stem or progenitor cells from other organs. Specifically, no cells had been identified as expressing stem or progenitor markers, such as nestin, ²² a marker that is characteristic of stem and progenitor cells. Nestin is expressed in fetal eyes ²⁸ and can be reactivated by injury or after injection of growth factors, but it is uncommonly expressed in adult eyes. ²⁹ Recent experiments with transgenic nestin–GFP mice suggest that a small proportion (0.7%) of ciliary epithelial cells may express nestin even without growth factor stimulation (judged by GFP measurement by fluorescent cell sorting). When dissociated ciliary epithelium or ciliary body in situ ³⁰ is exposed to growth factors (EGF, FGF2, or insulin), however, the susceptible progenitor cells alter their phenotype and begin to express nestin and glial fibrillary acidic protein, as previously reported. ³¹ We report for the first time that ciliary progenitors express vimentin. Previous studies had demonstrated evidence of active cell division in progenitors by incorporation of BrdU. ²² Our labeling experiments with Ki-67 provide more specific evidence of the course of cell division, which parallels that of the diameter of the neurospheres themselves. Many cells in neurospheres undergo a burst of rapid division for 3 days; a decreasing proportion continues to divide at 1 week, and active division ceases by 10 days under the conditions used in these floating cultures. 

Recent experiments with mouse ciliary and central nervous system progenitor cells in floating culture suggest that such spheres can expand with the addition of individual cells present in the same culture or by fusion of two spheres when the density of initial cells is 100 to 200 cells/mL, densities much higher than those used in our investigation. ^{24 32} However, each of these investigations confirmed that single cells produced many daughter cells by clonal expansion and expressed nestin. In addition, we found empty areas within some spheres when they were studied histologically and apoptotic features in cells by electron microscopy. This may explain differences in diameter and total cell numbers between rodent and human spheres; older human cells may be more likely to undergo cell death than those of young animals. 

In addition to expression of general progenitor components, ciliary progenitor cells have been reported to express homeobox genes that are prominent in normal retinal development, ³³ including Chx10 and Pax6. ^{1 22 25 30 31} We provide new quantitative, histochemical data showing that a high proportion of NS progenitor cells express these genes. Chx10 plays an important role in normal eye development, and mutations in this gene cause an ocular retardation phenotype in mice and a dramatic decrease in retinal progenitor cell proliferation. ³⁴ The transcription factor Pax6 is expressed in brain and retina during development and plays an important role in the regulation of cell proliferation and the determination of neuronal fate. ^{35 36 37 38} Pax6 mutation is associated with anophthalmos or small eye phenotype in mammals. ³⁹ The expression of Chx10 and Pax6 can be taken as evidence that ciliary-derived progenitors are committed to development into retinal tissues. 

It is logical that ciliary progenitors in their proliferating stage in neurospheres would be relatively devoid of specific features of mature retinal neurons. Our electron microscopic photomicrographs indicate their high intracellular activity and their nonspecific features. Ciliary progenitors did show some junctional complexes, as also reported by Kohno et al., ²⁴ that are shared by ciliary epithelium, though these were not observed in all cells. Immunohistochemical data on NS cells revealed that almost none of the markers typical for mature retinal cells were demonstrated, including markers of photoreceptors, horizontal cells, amacrine cells, or ganglion cells (Brn3b). We report here, for the first time to our knowledge, the expression by ciliary progenitors of Thy-1, a molecule present on mature retinal ganglion cells that is also expressed in the eye during development. ^{40 41} Other investigators have reported the expression of mature retinal cell markers in ciliary progenitors; in every case, however, this was true only when the NS was placed in contact with culture plates, was cocultured with embryonic retinal tissue, or underwent withdrawal of growth factor stimulation. These conditions allow the transition toward cells expressing molecules seen in adult retina, such as neurofilament protein, β-tubulin, Brn3b, opsin, and protein kinase C. ^{1 25} It will be important as a next step in this research to develop protocols that shift phenotypic expression toward the production of specific retinal neurons. We plan microarray studies of gene expression by progenitor cells to guide a strategy that may allow the successful generation of new retinal neurons. 

 Supported by National Eye Institute Grant P30EY001765 and by the Leonard Wagner Charitable Trust.

 Submitted for publication August 31, 2006; revised October 17, 2006; accepted January 19, 2007.

 Disclosure: H. Xu, None; D.D. Sta. Iglesia, None; J.L. Kielczewski, None; D.F. Valenta, None; M.E. Pease, None; D.J. Zack, None; H.A. Quigley, 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.

 

Quantitative Characteristics of Progenitor Cells

	Mouse	n	Rat	n	Human	n
Cell diameter (μm)	22 ± 1.1	7	21 ± 0.3	4	25 ± 0.4	8
NS diameter (1 week, μm)	139 ± 22	7	137 ± 9	4	141 ± 11	8
Cells isolated per eye	3.3 × 104	6	2.0 × 105	4	1.8 × 106 , †	6
NS-forming cells (%)*	1.0	4	0.9	4	2.5	7
Cell number per NS	1183	3	5360	1	685	2

 Diameters are expressed as mean ± SD.

*  From serial dilution of ciliary cells in individual wells.

†  In rat and mouse, all cells obtainable from each eye were used. In humans, pieces of ciliary body averaging 2.5 mm long were used from each eye. n = number of ciliary body preparations from which multiple counts of cells and neurospheres were performed.

Figure 1.

 

Neurosphere growth curves for mouse, rat, and human ciliary body cells.

Figure 2.

 

Neurospheres in human eyes.

Figure 3.

 

Ki-67 labeling in mouse neurosphere.

Table 2.

 

Positive Immunostaining of Markers in Neurosphere Cells

Immunostaining Markers	Positive NS Cells (%)	SD	No. NS Counted
Nestin	87.1	5.5	8
Chx10	81.7	11.1	9
Pax-6	85.4	10.1	4
Vimentin	83.5	8.0	3
Thy-1	82.9	7.4	10
GFAP	97.4	1.0	3
Ki-67	21.3	8.7	6

Figure 4.

 

Immunochemistry examination of mouse neurosphere, retina, and ciliary body.

Figure 5.

 

Immunochemistry examination of mouse neurosphere, retina, and ciliary body.

Figure 6.

 

Transmission electron microscopy images of mouse neurosphere.

The authors thank Brenda Coles and Derek van der Kooy for their extensive help in developing methods to study ciliary progenitor cells, Rhonda Grebe for TEM assistance, and Ruben Adler for manuscript review.