Photoreceptor loss causes irreversible blindness in many retinal diseases. The identification of suitable donor cell populations is of considerable interest because of their potential use to replace the photoreceptors lost in disease. Stem or progenitor cells that give rise to neurons and glia have been identified in several regions of the brain, including the embryonic retina and the ciliary epithelium of the adult eye, raising the possibility of autologous transplantation. However, there has been little systematic investigation into precisely which regions of the large mammalian adult eye give rise to such cells. Here, we show for the first time using the porcine eye the presence of progenitor cells in additional regions of the adult eye, including the pars plana and iris, regions that, in the human, are readily accessible during routine eye surgery. When cultured in the presence of growth factors, these cells proliferate to form neurospheres comprised of cells expressing retinal progenitor markers. Using an adherent monolayer culture system, these cells could be readily expanded to increase their number more than 1 million-fold and maintain a progenitor phenotype. When grown on the substrate laminin in the presence of serum, cells derived from both spheres and monolayer cultures differentiated into neurons and glia. These results suggest that a population of cells derived from the adult iris, pars plana, and ciliary body of a large mammalian species, the pig, has progenitor properties and neurogenic potential, thereby providing novel sources of donor cells for transplantation studies.
Disclosure of potential conflicts of interest is found at the end of this article.
Retinal degenerative diseases are the leading causes of untreatable blindness. The most common causes, age-related macular degeneration and retinitis pigmentosa, result in the loss of photoreceptor cells. One attractive therapeutic strategy is the transplantation of stem or progenitor cells into the diseased eye in order to replace the photoreceptor cells lost over the course of the disease. We have recently shown that cell transplantation can restore visual function in the diseased mouse retina, provided that the correct ontogenetic stage of donor cell is used . A subsequent and significant challenge is to find suitable sources of donor stem or progenitor cells and to find ways to manipulate them in vitro to the appropriate developmental stage prior to transplantation.
Transplantation studies conducted in small mammals have used either fetal or early postnatal tissue or embryonic stem cells as sources of donor cells [1, , –4]. Ethical considerations are likely to make it difficult to use fetal tissue as a source of donor cells in humans, and thus an adult source would be preferable. Recent research has suggested that progenitor cells with stem-like properties of self-renewal and multipotentiality might be isolated from the adult human eye , raising the theoretical possibility of autotransplantation if these cells could be expanded and induced to differentiate into photoreceptors. Autotransplantation could circumvent the need for long-term immunosuppression, which is required to prevent rejection of transplants originating from a nonautologous source.
Two populations of proliferative cells have been described in the adult mammalian eye: retinal stem cells and retinal progenitor cells [6, , –9]. During retinal histogenesis, progenitor cells are capable of a finite number of divisions to generate the full complement of retinal neurons and glia. By contrast, stem cells theoretically persist during adult life and retain properties of multipotentiality and self-renewal. However, there is limited evidence of either specific cell markers or assays that permit the distinction between these two cell types [10, 11].
In contrast to mammals, lower vertebrates continue to generate new retinal neurons throughout life from the ciliary marginal zone (CMZ), which is the circumferential region anterior to the neural retina [12, , –15]. The progeny derived from these CMZ cells comprises various cell morphologies and laminar positions, including retinal pigment epithelium (RPE), photoreceptors, and inner retinal neurons, indicating that CMZ progenitors are multipotent . Until recently, it was thought that the adult mammalian retina lacked proliferative or regenerative capacity. However, in vitro experiments suggest that the ciliary epithelium, part of the ciliary body, a structure analogous to the lower vertebrate CMZ, contains a population of progenitor cells [6, 7], but their function in vivo is still unclear . Although mitotically quiescent during adult life, when cultured in vitro these cells demonstrate characteristics typical of stem cells, including multipotentiality and self-renewal [6, 7]. Cells from the ciliary body, but not the sensory neural retina, have been shown to generate neurospheres that express nestin, a marker for somatic neural progenitor cells . Following incubation under differentiation conditions, these cells may differentiate into both neuronal and glial phenotypes, indicating that a proportion of cells in the ciliary body are multipotent. Although there is some evidence to suggest that these cells differentiate into cells that are immunopositive for specific retinal markers , it is unlikely that these represent normal, fully differentiated retinal neurons .
Previous studies in lower vertebrates and small mammals have indicated that cells other than those derived from the ciliary epithelium may be potential sources of adult progenitor cells, capable of differentiating into neural phenotypes [18, –20]. Furthermore, during development, some cell types are capable of dedifferentiation; following surgical removal of the retina from chick embryos and implantation of slow release fibroblast growth factor (FGF) beads, the remaining RPE dedifferentiates to form an inverted retina in vivo . It therefore seems likely that a number of cell types within the adult mammalian eye are potential sources of cells with progenitor cell properties. However, there has so far been little systematic investigation into exactly which of the many regions of the adult eye give rise to these cells.
We sought to determine the potential for harvesting adult-derived progenitor cells from different regions of the large mammalian eye, particularly those most readily accessible during routine eye surgery, such as the iris and pars plana. The porcine eye provides an excellent model, since its morphology is similar to the human eye and the anatomical regions can be clearly defined and accurately dissected. Previous studies have demonstrated that progenitor cells can be derived from the porcine brain and propagated in culture [22, –24] and demonstrate many similarities with their human counterparts . The genetic similarities between pig and human have already been utilized in the generation of porcine heart valves for transplantation into human subjects .
We demonstrate that neurospheres containing progenitor cells could be derived from not only the ciliary body but also the iris and pars plana, whereas the anterior neural retina failed to give rise to cells with such characteristics. Cells derived from each of the adult porcine ciliary body, iris, and pars plana could be similarly expanded in culture, maintaining a proliferative phenotype, and we show that these cells can be differentiated into cells that express either neuronal or glial markers.
Materials and Methods
Male Landrace pigs were reared according to British farming regulations and were approximately 2 years of age at the time of slaughter. Eyes were obtained from the slaughter house and placed immediately on ice prior to dissection. This occurred within 2 hours of slaughter. Eyes used for control reverse transcription-polymerase chain reaction (RT-PCR) analysis were obtained at the time of slaughter, the lens removed, and the eyes placed in RNA-later (Ambion; Applied Biosystems, Warrington, U.K., http://www.appliedbiosystems.com) for transportation.
Eyes were dissected in Earl's balanced salt solution (Sigma-Aldrich, Poole, U.K., http://www.sigmaaldrich.com). Eyes were hemisected and the posterior half of the eye removed. Taking the anterior half, samples from each of the anterior neural retina, pars plana, ciliary body, and the iris were carefully dissected free from all surrounding tissue.
In Vitro Cultures
Primary Neurosphere Cultures.
Tissue samples were dissociated according to the manufacturer's instructions using a papain-based dissociation system (Worthington Biochemical, Lakewood, NJ, http://www.worthington-biochem.com) (supplemental online Methods). Dissociated cells were resuspended in Dulbecco's modified Eagle's medium-F12/Glutamax (Invitrogen, Paisley, U.K., http://www.invitrogen.com) containing N2 supplement (1:100; Invitrogen), penicillin-streptomycin solution (1:100; Invitrogen), epidermal growth factor (EGF) (20 ng/ml; Peprotech, London, http://www.peprotech.com), FGF-2 (20 ng/ml; Peprotech EC), and heparin (2 μg/ml). Cells were plated at a density of 10–20 cells per microliter in untreated 24-well tissue culture plates. Plating at this density is reported to lead to the formation of clonally derived neurospheres [5, 7]. Fresh growth factors were added every other day, and medium was exchanged every 5–7 days.
Spheres isolated from each region at 7 days in vitro were plated in plastic 24-well tissue culture plates. Spheres were cultured in the presence of 5% fetal calf serum (FCS) and mitogens for an additional 7 days. During this time, spheres became adherent to the culture dish, and cells spread out and migrated away from the sphere to form monolayers. Monolayer cultures were split every 7 days using trypsin-EDTA solution (Sigma-Aldrich) and replated into fresh plates. The number of cells was determined using a hemocytometer to count the number of viable cells in a 20 μl sample of cell suspension immediately following passaging and prior to replating.
To assess the differentiation potential of cells within neurospheres, individual spheres were plated in 24-well plates on a substrate of poly-l-ornithine (10 μg/ml)/laminin (100 ng/ml) (both Sigma-Aldrich) in medium supplemented with 10% FCS. Medium was changed every 3–4 days, and the cells were allowed to migrate and differentiate over the course of 21 days. During the first 3 days of differentiation, FGF-2 (20 ng/ml) was added to each well.
Characterization of Neurospheres
At 7 days in vitro, cultures were assessed for neurosphere number. Neurospheres were defined as free-floating, with a diameter of >40 μm and a clearly defined outer boundary. This feature readily distinguishes neurospheres from any smaller aggregations of cells, which have uneven boundaries. All spheres in a given well were counted. Results are expressed as number of neurospheres per 50,000 cells. To assess size and pigmentation, neurospheres were visualized using an inverted Leica DMIL microscope (Leica, Heerbrugg, Switzerland, http://www.leica.com) fitted with a camera and image capture system and analyzed off-line. Neurospheres were selected at random from the image set, which had previously been recoded to ensure blind assessment. Neurosphere diameter was measured twice, the second measurement perpendicular to the first, and an average of the two was calculated.
Pigmentation was assessed using the images captured for size analysis and Image-Pro software. Black was given a pixel value of 0, whereas white had a value of 255. Image-Pro software calculates the average pixel intensity of a given region. The regions were selected manually to encompass the whole neurosphere, as determined by drawing around the perimeter of the sphere. Neurospheres were selected at random from the masked image set.
Spheres cultured for 7 days in vitro were placed on poly-l-lysine coated glass slides (BDH; VWR, Leicestershire, U.K., http://uk.vwr.com) and allowed to settle prior to fixation with 4% paraformaldehyde (PFA) for 20 minutes at room temperature (RT). Spheres were preblocked in Tris-buffered saline (TBS) containing normal goat serum (1%), bovine serum albumin (1%), and 0.5% Triton X-100 for 2 hours at RT before being incubated with primary antibody overnight at 4°C. After rinsing three times for 30 minutes each time (3 × 30 m) with TBS, spheres were incubated with secondary antibody for 4 hours at RT, rinsed (3 × 30 m), and counterstained with Hoechst 33342. Negative controls omitted the primary antibody.
Monolayer cultures were fixed at 7 days in vitro with 4% PFA for 10 minutes at RT. Cells were preblocked in TBS containing normal goat serum (1%), bovine serum albumin (1%), and 0.1% Triton X-100 for 1 hour at RT before being incubated with primary antibody overnight at 4°C. After rinsing 3 × 30 m with TBS, cells were incubated with secondary antibody for 2 hours at RT, rinsed (3 × 30 m), and counterstained with Hoechst 33342. Negative controls omitted the primary antibody.
Plates were fixed in 4% PFA. Cells were preblocked, as above, before being incubated with primary antibody overnight at 4°C. After rinsing 3 × 10 m with TBS, cells were incubated with secondary antibody for 2 hours at RT, rinsed (3 × 10 m), and counterstained with Hoechst 33342. Negative controls omitted the primary antibody. Positive controls comprised staining of frozen porcine retinal sections. The percentage of cells positive for a given marker was determined by counting cells with a signal greater than background from more than three regions per well and more than three independent experiments.
The following antibodies were used: Pax6 (mouse; 1:100; Developmental Studies Hybridoma Bank, Iowa City, IA, http://www.uiowa.edu/∼dshbwww) and Sox2 (rabbit; 1:200; Abcam, Cambridge, U.K., http://www.abcam.com) for stem/progenitor cells, β-III tubulin (mouse; 1:1,000; Promega, Madison, WI, http://www.promega.com) and NeuN (goat; 1:100; Chemicon, Temecula, CA, http://www.chemicon.com) for neurons, Brn-3b (goat; 1:100; Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com) for ganglion cells, Rho4D2 (rabbit; 1:100; kind gift from R. Molday) and recoverin (rabbit; 1:100; Chemicon) for photoreceptors, protein kinase C (PKC) (mouse; 1:100; Sigma-Aldrich) for rod bipolar cells, and RPE65 (mouse; 1:100) for RPE cells. The appropriate Alexa-tagged (Molecular Probes, Eugene, OR, http://www.probes.invitrogen.com) secondary antibodies were used. Staining was visualized using either a Zeiss (LSM 510; Carl Zeiss, Jena, Germany, http://www.zeiss.com) or Leica (SP2) confocal microscope.
Incorporation of 5-Bromo-2′-deoxyuridine In Vitro
At 7 days in vitro, monolayer cells were pulse labeled with 5-bromo-2′-deoxyuridine (BrdU) (0.5 μM) for 4 hours. The plates were fixed in 4% PFA for 20 minutes before processing for BrdU immunohistochemistry. Briefly, cells were exposed to 2 M HCl for 30 minutes at 30°C to denature cellular DNA. HCl was neutralized by application of 0.1 M Na-borate prior to rinsing with phosphate-buffered saline. Rat anti-BrdU (1:500; Abcam) and then Alexa-546 goat anti-rat (1:250; Molecular Probes) were used for BrdU staining. Hoechst 33342 was included in the final wash to label all cell nuclei. The percentage of cells that were BrdU-positive was determined by counting the number of BrdU-labeled and Hoechst-labeled cells in more than five randomly selected fields of view from each well and more than six wells per region. The results presented are from N = 3 (passage 1) and N = 4 (passage 11) independent experiments.
Neurospheres were mounted on glass coverslips and imaged using an inverted confocal microscope (Zeiss LSM510). The fluorescence of Hoechst 33342, Alexa-488, and Alexa-546 were excited with the 350-nm line of the UV laser, the 488-nm line of the argon laser, and the 543-nm line of the HeNe laser, respectively. Images show projections of multiple single confocal sections taken at approximately 5–10 μm steps.
RNA was isolated from fresh tissue or cells using the TRIzol reagent (Invitrogen) in accordance with the manufacturer's instructions. cDNA was generated from 1 μg of RNA using the QuantiTect reverse transcription kit (Qiagen, Hilden, Germany, http://www1.qiagen.com) following the manufacturer's instructions. Two hundred ng of cDNA was used in each polymerase chain reaction (PCR). Porcine retinal cDNA or porcine genomic DNA (200 ng per reaction) were used as positive or negative controls. Previously published primers that amplify specific target porcine sequences were used according the conditions used in the original publication . In each case, PCR primers were designed to flank at least one intron.
Results are presented as the mean ± SEM. Where appropriate, N indicates the number of eyes and n the number of neurospheres investigated. The different regions were compared using a one-way analysis of variance (ANOVA) test with Dunnett's correction for multiple comparisons, where appropriate.
Proliferative Potential of Different Regions of the Porcine Eye
The ability to form clonal spheres, whereby a stem cell undergoes proliferation to form a free-floating neurosphere containing stem/progenitor cells, is well documented and provides an indication of the number of endogenous stem-like cells in a given tissue [5, 7, 28, 29]. To examine the different regions of the porcine eye, primary cells from the iris, ciliary body, pars plana, and anterior neural retina (Fig. 1A) were isolated and dissociated into single cells. Free-floating neurospheres could be easily derived from single dissociated cells of the adult porcine iris, ciliary body, and pars plana but not the anterior neural retina (Fig. 1B). Neurospheres were visible at 3 days in vitro, and the number and size of neurospheres was assessed at 7 days in vitro. Significantly, the number of neurospheres isolated from both the iris and the pars plana was approximately two- and threefold greater, respectively, than the number isolated from the ciliary body (N = 9; Fig. 1C). Neurosphere size provides an indirect measure of proliferation; neurospheres from all three regions were of a similar size, with no significant differences among groups (Fig. 1D). Only rarely were sphere-like structures derived from the anterior retina. These were extremely small and did not grow with additional culturing (data not shown).
Spheres could be partially dissociated and grew to form secondary neurospheres, but passaging did not lead to expansion of cell number. Thus, neurospheres from each of the three regions had limited self-renewal potency, consistent with previous reports of studies using neurospheres derived from either the chicken iris  or rat ciliary body . Previous studies have suggested that the cells that give rise to neurospheres in the murine ciliary epithelium are pigmented and that pigmentation is lost during the proliferation of the neurospheres . Here, we found that, typically, neurospheres derived from the ciliary body and pars plana contained a core of pigmented cells, which were surrounded by nonpigmented cells. By contrast, neurospheres derived from the iris were highly pigmented and failed to depigment even during prolonged culture periods. To assess pigmentation, neurospheres were quantified using a linear scale from black (0) to white (255) and showed a significant difference in the pigmentation of neurospheres derived from the iris (average 37) compared with those derived from the ciliary body (98) and pars plana (98) (Fig. 1E; N > 9 for each region; one-way ANOVA test p < .0001).
Cellular Composition of Porcine Retinal Neurospheres
Since the above data indicated some phenotypic differences between neurospheres derived from the different regions of the porcine eye, we examined the cellular make-up of the neurospheres using immunocytochemistry and a variety of progenitor cell markers. At 7 days in vitro, neurospheres from all three regions (iris, ciliary body, and pars plana) expressed the stem/progenitor cell marker Sox2 and the neural progenitor marker Pax6 (Fig. 2). Occasionally, small numbers of cells, usually at the outer edges of the spheres, labeled for the neuronal marker β-III-tubulin, suggesting that some differentiation may occur even under proliferative culture conditions (Fig. 2).
RT-PCR analysis was used to confirm the presence of additional markers of retinal and neural progenitor cells (Table 1). Neurosphere cultures were shown to express RNA encoding Hes1, which is expressed in immature central nervous system progenitors in the pig  and the eye field genes Dach1, Lhx2, and Six3. Lhx2 and Dach1 are expressed in the forebrain, optic vesicle, and developing neural retina of the mouse [31, –33]. During eye development, Six3 is initially expressed throughout the optic vesicle before becoming restricted to the neural retina and lens . The glial-associated intermediate filament proteins glial fibrillary acidic protein (GFAP) and vimentin were also detectable in spheres from all regions. RT-PCR failed to detect expression of doublecortin, indicating a lack of immature neuronal cells within spheres in serum-free culture. Together, these data indicate that the majority of cells within spheres from each region were progenitor-like cells expressing markers characteristic of immature retinal cells.
Table Table 1.. Expression of additional retinal progenitor markers in neurosphere cultures
Expansion of Adult Porcine Progenitor Cells in Monolayer Cultures
Given the limited self-renewal potential of neurospheres, we assessed the use of adherent monolayer cultures for expanding cell number. Neurospheres derived from each of the three regions grew readily as monolayers. Whereas monolayer cells could be maintained for 1–2 passages in the absence of serum, sustained expansion only occurred in the presence of low serum (4%). Cells from the ciliary body increased in number from 2.0 × 104 to 5.8 × 109 over nine passages, iris cells expanded from 2.0 × 104 to 3.9 × 108 cells, and cells from the pars plana grew in number from 2.0 × 104 to 4.3 × 107 cells over the same period (Fig. 3A). Cells from each region continued to proliferate throughout the period studied and could be passaged 25 times, although they failed to expand significantly after this time (data not shown). The growth rate of monolayer cultures from all three regions was also investigated by incubation in the presence of BrdU. At passage 1, cells derived from the pars plana and iris showed proliferation rates equivalent to those derived from the ciliary body, as determined by BrdU uptake (N = 3; Fig. 3B). By passage 11, the percentage of BrdU uptake for cells from each region had increased approximately threefold (N = 4; Fig. 3B).
A common concern regarding the use of stem cells cultured for long periods in cellular therapies is the potential for immortalization and tumorogenesis. Typically, immortalized cells overcome contact inhibition and continue to divide, having reached confluency. By contrast, the cells described here ceased to divide upon achieving confluency and required regular passaging to maintain expansion. Furthermore, our finding of novel progenitor populations within the porcine eye is based on the analysis of more than ten independently derived cell cultures and so is highly unlikely to be due to immortalization events. Expansion of cell number was not seen beyond passage 25, also indicating that these cells are not immortal. Finally, proliferating cells cease to divide upon removal of mitogens and the addition of serum (10%). Such patterns of behavior are typical for primary progenitor cell populations grown in monolayer cultures [35, 36] and suggest that these cells senesce after moderate expansion.
Cellular Composition of Porcine Retinal Monolayer Cultures
To determine whether monolayer cultures derived from the different regions of the eye retain the ability to express the progenitor markers found in primary neurospheres, we used immunohistochemistry to look for expression of Sox2 and Pax6 in passage 4 monolayer cultures (Fig. 4). Staining confirmed the presence of Sox2 and Pax6 expression in cultures from all three regions. Many of the cells staining with Sox2 and Pax6 were located within densely packed regions within each well and were surrounded by cells with lower levels of Sox2 and Pax6 expression (Fig. 4). Analysis of cells from later passages (passage 11) revealed that Sox2 expression was detectable in 70% of cells of pars plana origin, 55% of ciliary body cells, and 20% of iris derived cells. However, we were unable to demonstrate robust Pax6 immunostaining in these cells (data not shown). To determine the prevalence of more mature progenitor cells within the passage 11 monolayer cultures, immunohistochemistry for β-III-tubulin was performed (Figs. 4, 6A). Positive staining was present in 9%, 15%, and 47% of ciliary body, pars plana, and iris cells, respectively.
Together, these findings indicate that the cells within monolayer cultures derived from the porcine ciliary body, pars plana, and iris have some self-renewal potency, similar to that reported for cells from the rat ciliary body [30, 37] and chicken iris . Both monolayer culture cells and sphere cultures from all three regions exhibited morphological characteristics consistent with other mammalian neural stem/progenitor cultures, including analogous cells from humans . However, the decreased expression of Pax6 staining and prevalence of the neuronal marker β-III-tubulin at later passages suggest that significant differentiation occurs in tissue culture.
Differentiation Potential of Adult Porcine Progenitor Cells
To assess the neurogenic potential of progenitor cells derived from the adult ciliary body, pars plana, and iris, neurospheres from each region were plated on laminin-coated tissue culture plates in the presence of 10% serum and FGF-2 (20 ng/ml). During this time, cells migrated out of the neurospheres and proliferated. After 3 days, the medium was replaced with serum-containing mitogen-free medium. By 4 weeks, >90% of the cells were immunopositive for the neuronal marker β-III-tubulin and had morphologies consistent with those of immature neurons (Fig. 5). Similarly, cells grown and passaged as adherent monolayers could be induced to differentiate and express neuronal markers. Dissociated monolayer cells (passage 11) taken from each region were plated on laminin-coated tissue culture plates under differentiation conditions as described above. At 4 weeks, Sox2 expression was no longer detectable (Fig. 6B). We observed robust expression of β-III-tubulin in 28%–40% of cells (Figs. 5, 6A). To further confirm the neuronal phenotype of these differentiating cells, immunostaining for NeuN was also performed. NeuN staining was prevalent in samples derived from all three eye regions and appeared very specific to individual cells within each culture (Fig. 5). These findings indicate that progenitor cells from the porcine iris, pars plana, and ciliary body may be expanded and are each capable of yielding neuronal cells. However, under the conditions used here, differentiated cells failed to show convincing immunoreactivity for rhodopsin, phosducin, recoverin, PKC, or RPE65 (data not shown), suggesting, as has been described in other species , that fully differentiated retinal cell types were not produced.
Differentiated primary neurospheres and differentiated cells from late passage (P12) monolayer cultures were also examined by RT-PCR. Expression of the eye field markers Dach1, Lhx2, and Six3 was found in all cultures (Table 2). However, in contrast to undifferentiated neurosphere cultures grown without serum but in the presence of mitogens, differentiating cells from both late passage monolayer cultures and neurosphere cultures were positive for doublecortin, a marker of immature neuronal cells. The glial-associated markers GFAP and vimentin were also detectable. Together, these results suggest that progenitor cells derived from the iris, pars plana, and ciliary body of the adult porcine eye are able to generate new neurons and glial cells in culture.
Table Table 2.. Expression of additional markers in differentiated neurosphere and monolayer cultures
Here we have shown that the iris and pars plana regions of the adult porcine eye yield cells that are capable of significant proliferative expansion in EGF/FGF-2 supplemented media in vitro in a manner virtually indistinguishable from those derived from the ciliary body. Robust passaging and expansion could be achieved with monolayer cultures but could not be readily established with cultures of free-floating neurospheres, possibly because cells in the latter configuration are not easily dissociated. Proliferating cells from both sphere and monolayer cultures expressed the developmentally regulated transcription factors Sox2 and Pax 6 together with eye field and retinal markers including Lhx2, Dach1, and Six3. Finally, with serum-induced differentiation, cells expressed the glial markers GFAP and vimentin or the neuronal markers β-III-tubulin, doublecortin, and NeuN, although no photoreceptor specific markers could be elicited using these protocols. Single cells isolated from all three regions of the porcine anterior uvea (but not retina) can therefore undergo clonal expansion for up to 25 passages in vitro and subsequently express neuronal features after induced differentiation with serum based medium. Such cells are readily accessible during ophthalmic surgery and may provide alternative sources of donor cells for retinal or neuronal autotransplantation.
A population of stem-like cells has been described for the CMZ of lower vertebrates and the mammalian equivalent, the ciliary epithelium. However, there has been little systematic study of the different regions of a larger mammalian eye. The populations of cells that we isolated from the adult porcine iris, pars plana, and ciliary body displayed a number of properties similar to those described previously for stem/progenitor cells in the pigmented ciliary margin in smaller mammals where these regions are less easy to define [6, 7]. Neurospheres derived from all three regions expressed the stem cell marker Sox2 and the progenitor marker Pax6. The number of neurospheres generated from both the iris and the pars plana was 2–3 times greater than that isolated from the ciliary body, suggesting that these regions may be particularly rich in stem/progenitor cells. Individual neurospheres from each of the three regions were similar in size, indicating that they have comparable expansion potential when grown as neurospheres. In contrast to other regions of the porcine eye, the anterior retina of the adult pig lacks progenitor cells capable of expansion under the conditions used in this study.
To utilize cells for autologous transfer, large numbers of cells may be needed. Neurospheres generated from the ciliary body and iris of other species have previously been shown to have limited potential for secondary neurosphere formation and passaging . Similarly, we also found limited secondary sphere formation from these regions, together with the pars plana, in the porcine eye. However, it was possible to achieve rapid and significant expansion of neurosphere-derived progenitor cells using adherent monolayer cultures in the presence of FGF-2 and EGF, conditions that led to an approximately 1 million-fold increase in cell numbers. Neurospheres from the ciliary body, pars plana, and iris could be efficiently expanded, forming depigmented monolayers.
We have recently demonstrated that, in contrast to previously held opinion, the optimal donor cell for retinal transplantation is not an uncommitted progenitor or stem cell but rather one that has already committed to the photoreceptor fate . For stem and progenitor cells from either embryonic or adult sources to be viable donor cells, it will be necessary to determine ways in which to generate such committed precursor cells in vitro. Wu et al.  reported a priming procedure using FGF-2 to treat fetal human neural stem cells in vitro before further differentiation that allowed them to obtain cholinergic neurons in vitro. Recently, Merhi-Soussi et al.  used a modified protocol to generate significant numbers of murine stem cells committed to the neuronal lineage. Here, we used a similar technique, priming cells with FGF in the presence of serum, followed by the complete withdrawal of growth factors. We found that, upon differentiation, cells from all three regions of the porcine eye, in both neurosphere and monolayer cultures, showed robust expression of the neuronal marker β-III-tubulin and exhibited the morphology of immature neurons. Such morphology was consistent with the relatively lower levels of labeling for NeuN (a more mature neuronal marker) observed in differentiated monolayer culture.
Previous reports have indicated that the addition of exogenous FGF-2 to dissociated P0 rat retinal cells grown as monolayers caused a marked increase in the number of cells that express rhodopsin . However, this effect was apparently absent at more immature developmental stages (embryonic day 16) , a finding replicated in the mouse retina . Lineage-tracing studies of single progenitor cells in vivo have shown that multipotent progenitors give rise to all the cell types of the retina, including photoreceptors . When placed in culture, these cells may also have the capacity to differentiate and express genes specific to mature cell types, such as photoreceptors . After long-term expansion, these cells maintain the capacity to generate cells committed to the photoreceptor pathway. However, the proportion of photoreceptors generated in vitro is very low compared with the expected numbers generated in vivo . Similarly, despite assessing a wide array of retinal-specific markers, we did not observe any convincing immunoreactivity for retinal neurons, including photoreceptors, suggesting that our culture conditions maintain a large number of cells in a primitive neuronal precursor stage. The loss of robust Pax6 staining from monolayer cultures at passage 11 may indicate a loss of retinal identity after prolonged expansion of cells in tissue culture. Such findings are in accordance with previous reports using smaller mammals, which have also found that progenitor cells derived from the iris or ciliary margin fail to differentiate into retinal-specific neurons . Furthermore, expression of retinal-specific markers such as the photoreceptor pigment rhodopsin could only be obtained following viral transduction of the photoreceptor transcription factor Crx . Other reports have suggested the presence of retinal markers in differentiated stem/progenitor cell cultures, as determined by immunohistochemistry [5, 19]. However, there is increasing evidence to suggest that such positive staining is unlikely to represent true differentiation into mature retinal phenotypes . Indeed, undifferentiated cells have been reported to not only express differentiated cell markers, but can also express markers corresponding to more than one lineage [43, 44]. Such findings highlight the difficulties in assessing the potential of stem and progenitor cell populations and the need for cautious and thorough assessments.
A promising strategy for directing the differentiation of a population of progenitor cells is to utilize viral gene transfer to express key transcription factors. Studies using rat and primate iris tissue have been successful in generating cells with a photoreceptor-like phenotype following transduction with Crx and Crx/NeuroD, respectively . Further studies will be needed to establish whether similar techniques may be used to guide the differentiation of adult-derived iris, ciliary body, and pars plana progenitor cells that have undergone multiple passages and expansion in tissue culture. Ideally, by using gene transfer, the appropriate combination of factors could be delivered to generate functional photoreceptor precursor cells with the potential for integration into a recipient retina.
The results presented here suggest that a population of cells derived from each of the adult porcine iris, pars plana, and ciliary body have retinal progenitor cell characteristics and neurogenic potential. Further studies are needed to understand more about the factors that induce cell-specific differentiation. The ability to generate large numbers of neural precursor cells in a controlled manner from an accessible region of a human might provide a valuable source of donor cells for autotransplantation.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.
We would like to thank P. Buch for technical assistance, R. Molday for providing antibodies, and members of R.R.A.'s lab for helpful discussions. This work was supported by Grants from the Medical Research Council U.K. (G03000341), the Royal Blind Asylum & School, the Scottish National Institute for the War Blinded, and the Special Trustees of Moorfields Eye Hospital. R.E.M. is a clinician scientist supported by the Health Foundation.