Molecular Vision 2006; 12:1606-1614 <http://www.molvis.org/molvis/v12/a184/> Received 10 July 2006 | Accepted 7 December 2006 | Published 20 December 2006

Download Reprint

Sorbitol causes preferential selection of Muller glial precursors from late retinal progenitor cells in vitro

Tasneem Zahir,¹ Henry Klassen,² Minoru Tomita,¹ Michael J. Young¹
 
 

¹Schepens Eye Research Institute, Harvard Medical School, Boston, MA, ²Singapore Eye Research Institute, 168751, Singapore

Correspondence to: Michael J. Young, Schepens Eye Research Institute, Department of Ophthalmology, Harvard Medical School, 20 Staniford Street Boston, MA 02114; Phone: (617) 912 7419; FAX: (617) 912 0101; email: mikey@vision.eri.harvard.edu

Abstract

Purpose: The replacement of glucose by sorbitol in growth medium causes selection of astroglial cells from heterogeneous primary cultures derived from the brains of newborn mice. The present study was undertaken to investigate the effects of sorbitol on in vitro selection of Müller glial precursors from expanded late retinal progenitor cells (RPCs).

Methods: RPCs used in these studies were isolated from the neural retina of postnatal day one green fluorescent protein (GFP) transgenic mice. The resulting GFP positive neurospheres were dissociated into a single cell suspension and grown on poly-D-lysine/laminin coated tissue culture flasks or slides to generate adherent RPCs. These adherent cells were treated with glucose free medium containing 25 mM sorbitol for 7 days and the expression of retinal-specific cell markers was determined by immunocytochemistry, reverse transcriptase polymerase chain reaction (RT-PCR) and immunoblot analysis

Results: In vitro studies showed that sorbitol treatment of late RPCs altered cellular morphology. Immunocytochemical studies showed an increase in the proportion of cells expressing glial cell markers, most of which co-expressed CRALBP, GFAP, and vimentin. An increase in the proportion of cells expressing PKCα, a bipolar cell marker, was also observed. RT-PCR analysis showed down-regulation of nestin transcripts with a concomitant increase in CRALBP, GFAP, vimentin and PKCα. These findings were confirmed by immunoblot analysis, where down-regulation of nestin expression with simultaneous up-regulation of CRALBP, GFAP and PKCα was observed.

Conclusions: Sorbitol treatment of multipotent late RPCs, in the absence of glucose, results in the preferential selection of Müller glial precursors and their subsequent differentiation into cells that morphologically resemble Müller cells and co-express multiple glial markers.

Introduction

The vertebrate retina is comprised of seven distinct cell types that are derived from a common progenitor cell. These include six classes of neurons and one glial cell type, the Müller glial cell [1]. Müller cells are astrocyte-like glial cells that span the width of the retina from the inner to outer limiting membranes [2]. They are considered the principal glial cells of the retina because they perform many of the functions carried out by astrocytes, oligodendrocytes and ependymal cells in other regions of the central nervous system [3]. Functions provided by Müller cells include maintaining the complex retinal architecture, giving structural and metabolic support to retinal neurons. Müller cells are also involved in the visual cycle through the synthesis and renewal of cone visual pigments. As part of this process they bind all-trans retinol, convert it into 11-cis-retinol and then release it into the extracellular space for uptake by cone photoreceptors [4,5].

More recently, another quite different role has been attributed to Müller glia, namely that of generating new neuronal cells during retinal regeneration. Because they also generate glial cells, Müller cells may constitute a previously unidentified stem cell in the vertebrate retina that becomes activated upon injury. Evidence of this comes from studies in the post-hatch chick retina where the endogenous Müller cells dedifferentiate, proliferate, express markers common to retinal progenitors, and produce new neurons in response to neurotoxic damage [6,7]. Furthermore, it has been demonstrated that neurotoxic damage is not essential for activation of retinal regenerative mechanisms since the endogenous glial cells become neurogenic following intravitreal injections of exogenous growth factors [8-10]. In a recent study, researchers showed that the capacity for retinal regeneration is not restricted to lower vertebrates, in that Müller cells generated new retinal neurons in the adult mammalian retina after toxic damage [11]. Furthermore, the fate of these regenerated neurons could be partially controlled via extrinsic factors or intrinsic genes.

The efforts of several groups have been focused on demonstrating that mature Müller glia have the potential to acquire the characteristics of retinal progenitors and can serve as a source of regeneration in the vertebrate retina. In contrast, our goal is to develop strategies for generating Müller glia from cultured late retinal progenitor cells. (RPCs) This work has implications for both retinal tissue engineering and the control of Müller cell activation and subsequent gliosis in a range of retinal diseases.

In this study, we determine the effect of sorbitol on the in vitro selection of Müller glial precursors from expanded late RPCs and their subsequent differentiation. The enzymes of the sorbitol pathway, aldose reductase and sorbitol dehydrogenase, along with the sorbitol uptake system, are present in rat astroglia-rich primary cultures but have not been detected in rat neuron-rich cultures [12,13]. Because of the absence of the appropriate enzymes for the uptake and metabolism of sorbitol, mouse brain-derived oligodendroglia, microglia and ependymal cells cannot survive in media containing sorbitol instead of glucose. Hence, the replacement of glucose by sorbitol has been used for the selection of astroglia from heterogeneous primary cultures of newborn mouse brains [14].

The present study was undertaken to investigate the effects of sorbitol on in vitro differentiation of late RPCs. As the progeny of late retinogenesis include rod photoreceptors, bipolar neurons and Müller glia, we sought to determine if sorbitol treatment of late RPCs would preferentially promote Müller cell differentiation, either directly through the induction of late RPCs down the glial pathway or indirectly by selection of Müller glial precursors.

Methods

Isolation and culture of late retinal progenitor cells

RPCs were isolated from the neural retina of postnatal day one GFP transgenic mice [15]. Briefly, retinas from newborn GFP transgenic mice (gift from Dr. Masaru Okabe, University of Osaka, Japan) [16] were surgically removed and the ciliary marginal zone and optic nerve head dissected. The tissue was finely chopped with sterile forceps and digested with 0.1% type I collagenase (Sigma-Aldrich, St. Louis, MO) for 20 min. The supernatant containing dissociated cells was then passed through a 100 mm mesh strainer, centrifuged and seeded in complete medium (CM) containing Neurobasal media (with 25 mM D-glucose; Invitrogen, Carlsbad, CA), B27 neural supplement (Invitrogen, Carlsbad, CA), 2 mM L-glutamine (Sigma-Aldrich), 100 mg/ml penicillin-streptomycin (Sigma-Aldrich) and 20 ng/ml epidermal growth factor (recombinant human EGF; Invitrogen). This cycle was repeated until all retinal tissue was digested. Cells were refed every 2 days. GFP+ neurospheres appeared within 2 to 3 weeks and were passaged at regular intervals. To assess the capacity for self-renewal, spheres were broken up and plated as single cells. Individual cells formed new spheres over a period of 5-7 days. These spheres labeled positively for nestin (a marker for neural progenitor cells; Developmental Studies Hybridoma Bank (DSHB), Iowa City, IA) and Ki67 (a marker for cell proliferation; Novocastra Laboratories, Newcastle, UK) [15].

In vitro differentiation of late retinal progenitor cells

GFP+ neurospheres were dissociated into a single cell suspension and grown on poly-D-lysine and laminin coated tissue culture plates to generate adherent RPCs that were subsequently exposed to sorbitol treatment. Briefly, passage 20 cells from confluent T75 flasks were trypsinized and dissociated into a single cell suspension (200,000 cells/ml). These cells were then seeded onto Biocoat poly-D-lysine/mouse laminin coated eight well culture slides or flasks (BD Biosciences, Bedford, MA) and allowed to grow for 12-16 h at 37 °C in CM. The CM containing EGF was then removed and the cells washed with Ca²⁺ and Mg²⁺ free HBSS (Invitrogen, Carlsbad, CA). Selection/Differentiation medium containing DMEM F12 modified (with 2 mM L-glutamine, with 25 mM D-sorbitol, without glucose, custom ordered from SAFC BioSciences, Lenexa, KS), B27 neural supplement (Invitrogen, Carlsbad, CA), and 100 mg/ml penicillin-streptomycin (Sigma-Aldrich), but no EGF, was then added to the experimental wells or flasks. Controls were treated with CM containing glucose but no sorbitol. The medium was changed every 3 days for up to 7 days. The cells grown in flasks were used for RT-PCR and immunoblot analysis whereas the cells grown on culture slides were fixed at 7 days post-sorbitol treatment for immunocytochemical studies.

Immunocytochemistry

RPCs cultured on Biocoat poly-D-lysine/mouse laminin culture slides were processed according to standard protocols. Briefly, cells were fixed in 4% paraformaldehyde (pH 7.2) in 0.1 M phosphate buffered saline (PBS) and blocked in PBS containing 2% BSA and 0.5% Triton X-100. They were then incubated with appropriate primary antibodies (Table 1) for 2 h at room temperature. Antibody incubations were conducted in PBS containing 2% BSA. Cells were washed three times for 10 min each with PBS and incubated with species specific secondary antibodies conjugated with Cy2 or Cy3 (Jackson ImmunoResearch Laboratory, Bar Harbor, ME) for 1 h at room temperature. Subsequently, cells were washed three times with PBS, coverslipped in 2.5% PVA (polyvinyl alcohol)/DABCO (1,4 diazabicyclo/2,2,2 octane) and examined under an epifluorescence microscope (Nikon Eclipse, E800). For double-labeling studies, cells were treated with 100% methanol for 30 min at -20 °C, prior to fixing with 4% paraformaldehyde. This was followed by several washes of 20 min each with PBS to eliminate GFP fluorescence. Note that for double labeling studies the cells were not incubated with a mixture of primary and secondary antibodies. Instead they were incubated with one set of primary and secondary antibodies followed by the second set of primary (rabbit polyclonal GFAP and goat polyclonal vimentin, Table 1) and secondary antibodies. Semi-quantitative immunocytochemistry was performed by counting a total cell number of at least 200 cells per well in randomly selected fields using 4',6-Diamidino-2-phenylindole (DAPI) labeling. Each determination was performed in quadruplicate (n=4), and each experiment was repeated at least three times.

Reverse transcriptase polymerase chain reaction analysis

Total RNA was isolated from RPCs (10⁶ cells per sample), grown in the absence or presence of sorbitol, using an RNAqueous-4PCR kit (Ambion, Austin, TX). Reverse transcription reactions were set up using the RETROscript kit (Ambion). A 2-step RT-PCR approach was used, involving an initial step of heat denaturation of RNA to obtain cDNA followed by a step of PCR amplification using gene specific primers (Table 2). The thermocycler program used for PCR amplification included a hot start (95 °C for 1 min) followed by denaturation at 94 °C for 30 s, annealing at specific temperatures (Table 2) for 30 s for 30 cycles, extension at 72 °C for 30 s for 30 cycles and final extension at 72 °C for 5 min. RT-PCR reactions (cDNA) were run on 2% agarose gels (containing ethidium bromide at a final concentration of 0.5 mg/ml) against a 100 bp ladder and products visualized under UV light. The amount of cDNA was normalized based on the signal from constitutively expressed S15 mRNA, which encodes a small ribosomal subunit protein.

Immunoblot analysis

RPCs, grown in the absence or presence of sorbitol, and B6 mouse retinas (6 weeks), were homogenized separately in lysis buffer (1% Triton-X 100, 10 mM Tris-HCl (pH 7.4), 5 mM EDTA, 50 mM NaCl, 50 mM NaF) containing protease inhibitor cocktail (1:100 dilution; Sigma-Aldrich) and phosphotase inhibitor (1:100 dilution; Sigma-Aldrich). Protein levels in total cell lysates were quantified using a protein assay kit (Bio-Rad, Hercules, CA). The protein samples (20 mg) were separated on poly-acrylamide gels (NuPage; Invitrogen) for 40 min at 160 V and transferred to polyvinylidene difluoride (PVDF) membranes (Invitrolon; Invitrogen) for 60 min at 30 V. After transfer, the membranes were blocked in 5% nonfat dry milk in TBS-T (10 mM Tris HCl, pH 7.4, 150 mM NaCl, 0.1% Tween-20) for 30 min. The blots were incubated with the following primary antibodies: nestin (1:20 dilution; DSHB), CRALBP (1:10,000 dilution; ABR), GFAP (1:1000 dilution; Chemicon), PKCα (1:200 dilution; Santa Cruz Biotechnology, Santa Cruz, CA) and recoverin (1:1000 dilution; Chemicon, Temecula, CA). Subsequently, blots were incubated with horseradish peroxidase (HRP) conjugated species-specific secondary antibodies and the signals visualized with the ECL Western blotting detection system (GE Healthcare Bio-Sciences, Piscataway, NJ).

Results

Sorbitol-induced changes in morphology of late retinal progenitor cells

Adherent RPCs generated from dissociated RPC neurospheres continued to proliferate and formed a monolayer. Most of these cells remained morphologically undifferentiated, although a small number extended short processes (Figure 1A). After sorbitol treatment the proliferation of adherent RPCs decreased dramatically. Many of these cells displayed a bipolar, spindle shaped morphology, often giving rise at one pole to relatively wide cytoplasmic projections resembling endfeet and at the opposite pole to a long smooth process terminating in a growth cone-like structure (Figure 1B).

Semi-quantitative evaluation of effect of sorbitol on protein expression using immunocytochemistry

The effect of sorbitol on the expression of proliferative, progenitor, glial and bipolar cell markers was determined by immunocytochemistry (Table 3). In the adherent RPC control population, 30.2% of cells expressed Ki67 (Table 3; Figure 2A) compared to 4.5% of adherent RPCs after sorbitol treatment (Table 3; Figure 2B). After 7 days of treatment, the percentage of RPCs expressing CRALBP, GFAP and vimentin had increased from 0 to 32.2%, 7.2% to 52.2%, and 0 to 49.3%, respectively (Table 3; Figure 2E-J). Sorbitol also had an inductive effect on PKCα expression. The percentage of cells expressing PKCα increased from 9% to 22.3% (Table 3; Figure 2K-L). In contrast to the increase observed in the percentage of cells expressing bipolar and glial markers, sorbitol treatment decreased the percentage of cells that stained positive for nestin from 92.5% to 46.7% (Table 3; Figure 2C-D).

Double-labeling with glial markers for cell identification

The expression of different glial markers (CRALBP, GFAP, vimentin) by RPCs grown in the presence of sorbitol led us to test whether these RPCs co-labeled with different combinations of glial markers (CRALBP and GFAP, CRALBP, and vimentin, vimentin, and GFAP). After 7 days of treatment 100% of the cells that expressed CRALBP (32.2% of total cell population) also expressed GFAP and vimentin (Figure 3G-I, Figure 3J-L). However, not all the GFAP-positive cells were CRALBP-positive (32.2%%±1.9 versus 52.2%±2.3 of total cell population, Figure 3G-I). In contrast, all the GFAP-positive cells were vimentin-positive (52.2%±2.3 versus 49.4%±3.4 of total cell population, Figure 3M-O). In addition, the sorbitol treated cells were also examined for co-expression of Ki67 and GFAP, nestin and GFAP, as well as PKCα and GFAP. These cells did not co-express Ki67 and GFAP or PKCα and GFAP (Figure 3A-C, Figure 3P-R). In contrast, co-localization of nestin and GFAP immunoreactivity was observed although the overlap was not 100% (Figure 3D-F).

Effect of sorbitol on transcription of glial, bipolar, and progenitor cell markers

Semi-quantitative RT-PCR analysis was carried out to determine the effects of sorbitol on the preferential selection of Müller glial precursors from late RPC and their subsequent differentiation (Figure 4A). RPCs grown in the absence of sorbitol (referred to as RPCs in Figure 4A) showed high levels of nestin transcripts. After 7 days of sorbitol treatment there was down-regulation of nestin transcripts and a simultaneous up-regulation of CRALBP, GFAP and vimentin transcripts. The down-regulation of nestin transcripts, along with the concomitant increase in CRALBP, GFAP, vimentin, and PKCα transcripts, suggested that sorbitol increased the number of RPCs expressing markers of glial and bipolar neurons. In addition, the complete absence of recoverin confirmed the lack of induction of photoreceptor or cone bipolar differentiation by sorbitol. The RT-PCR results were corroborated by immunoblot analysis of late RPCs grown in the absence or presence of sorbitol (Figure 4B). Late RPCs grown in the absence of sorbitol (referred to as RPCs in Figure 4B) expressed high levels of nestin protein. In addition, low levels of GFAP and PKCα protein expression were also detected in these cells at baseline levels. Treatment with sorbitol for 7 days resulted in an induction of CRALBP and GFAP expression along with a dramatic decrease in nestin expression. Furthermore, RPCs treated with sorbitol showed an increase in PKCα expression (Figure 4B). In this study retinas from B6 mice (4 weeks) were used as negative control for nestin expression and positive control for CRALBP, GFAP, PKCα and recoverin expression. The data from immunoblot analysis confirmed that sorbitol treatment induced the late RPCs to up-regulate glial and bipolar markers. As with the RT-PCR analysis, the absence of recoverin expression excluded the possibility of differentiation into photoreceptors or cone bipolar cells.

Discussion

Primary cultures derived from the brains of newborn rodents contain neurons, oligodendrocytes, ependymal and other cell types, but are dominated by astroglial cells. Sorbitol, a long-chained polyol, is taken up in glial cells by diffusion through a proteinaceous channel-like structure that does not involve a carrier or direct penetration of the lipid bilayer [12]. Since the enzymes of the sorbitol pathway and sorbitol uptake system,are present in astroglia but not neurons [12,13], the replacement of glucose by sorbitol has been used for astroglial selection from cultured mouse brain [12,14]. The effect of sorbitol on in vitro selection and differentiation of Müller glial precursors from expanded late RPCs has not been studied previously.

The RPCs studied here are multipotent and have the capacity to differentiate into the cell types generated during late retinogenesis [1,17,18]. Results from our immuncytochemical, RT-PCR and immunoblot analyses indicated that sorbitol induced a significant number of late RPCs to differentiate, either directly through the induction of late RPCs down the glial pathway or indirectly by selection of Müller glial precursors, into cells that morphologically resembled Müller cells and also expressed the glial markers CRALBP, GFAP and vimentin. In addition to up-regulation of glial markers, an increase in cells expressing the bipolar cell marker PKCα was also observed. To our knowledge, this is the first study demonstrating that sorbitol has an inductive effect on bipolar cell differentiation. In contrast to its effects on glial and bipolar markers, sorbitol did not induce RPCs to express rod photoreceptor markers.

In the case of nestin and Ki67, sorbitol treatment decreased the percentage of RPCs that stained positive for both markers. Nestin [19], an established marker for CNS progenitor cells, is expressed by these cells in their proliferating, undifferentiated state [15,20,21]. When these cells undergo their final mitotic division and begin to differentiate into mature neurons, nestin expression is down-regulated. Similarly, Ki67 is a marker for cell proliferation (like BrdU pulse labeling) and its expression is used to determine the proportion of cells that are mitotically active. It is a prototypic cell cycle related nuclear protein, expressed by proliferating cells in all phases of the active cell cycle (G1, S, G2, and M phase). The expression of Ki67 is down-regulated in RPCs upon differentiation.

In addition to the first set of immunoreactivity studies, where cells were stained with a single marker, double-labeling studies were also carried out. After 7 days of sorbitol treatment, many of the cells that expressed CRALBP also expressed GFAP and vimentin. The co-expression of CRALBP (cellular retinaldehyde-binding protein) and vimentin (intermediate filament protein) strongly supports Müller cell differentiation, since both these proteins have been extensively used to identify Müller cells. The co-localization of GFAP, an intermediate filament protein that is primarily expressed in retinal astrocytes, with CRALBP, a marker for Müller cells in the adult retina, was not altogether unexpected. While resting Müller cells in the mammalian retina express little or no GFAP in vivo, this marker is strongly up-regulated in reactive Müller cells following retinal detachment, retinal injury and photoreceptor degeneration [22-24]. The up-regulation of GFAP expression observed in this study may be a result of the in vitro culture system. This possibility is supported by previous reports of increased GFAP expression by cultured Müller glia [25-27]. Thus our finding that Müller cells in culture express GFAP suggests that they are more similar to reactive than resting Müller cells.

Although all the CRALBP positive cells were GFAP-positive, the converse was not the case. However, all GFAP-positive cells were vimentin-positive. We can only speculate as to the identity of the subset of GFAP and vimentin co-expressing cells that did not express CRALBP (approximately 20% of total number of cells). It is tempting to simply call them retinal astrocytes, since both GFAP and vimentin are well-established astrocyte markers [26-29] However, the lineage of neuroretinal cells does not include retinal astrocytes, which instead are derived from the brain, and migrate into the retina along the optic nerve during development. We instead favor a description of these GFAP/vimentin positive cells as immature Müller cells that do not express all of the markers of mature Müller glia.

In addition, sorbitol treated RPCs did not co-express GFAP and either Ki67 or PKCα. Co-localization of nestin and GFAP was observed, although the overlap was not complete. From these findings we conclude that the cells generated by treatment of RPCs with sorbitol are similar to reactive Müller cells because of up-regulation of GFAP and co-expression of nestin by the same cells [30,31].

Semi-quantitative RT-PCR analysis was carried out to further investigate the effects of sorbitol on late RPC differentiation. The findings of down-regulation of nestin transcripts with concomitant increases in CRALBP, GFAP, vimentin, and PKCα suggested that sorbitol increased the number of RPCs expressing markers of glial and bipolar neurons. Additionally, the complete absence of recoverin indicated the lack of induction of photoreceptor or cone bipolar differentiation by sorbitol.

The RT-PCR results were corroborated by immunoblot analysis. These data confirmed that sorbitol treatment induced the late RPCs to up-regulate glial and bipolar markers. The absence of recoverin expression again argued against the possibility of induction of differentiation of RPCs by sorbitol into photoreceptors or cone bipolar cells.

In the sorbitol treated cultures where CM was replaced with differentiation media after 12-16 h and the cells were grown for 7 days, 30% of the total cells survived compared to the cells in control cultures that were grown in CM media for 12-16 h. Although sorbitol may in fact favor the survival of Müller and bipolar cell precursors over other cell types that exist within the total RPC population (i.e rod photoreceptors), the results of the subsequent differentiation studies offer clues as to the relative importance of this mechanism. As the bipolar cell marker PKCα increased in expression from 9% to 22.3% of the surviving cells upon treatment with sorbitol and subsequent differentiation, this may be wholly attributed to the survival or increased proliferation of bipolar cell progenitors over other cell types. In contrast, Müller cell markers are increased from undetectable levels to 33-50% of the surviving cells after treatment with sorbitol. Even a profound increase in proliferation of glial precursors could not fully explain such an increase in Müller cell phenotypes. Furthermore, sorbitol tended to decrease, rather than increase, markers of cell division. Taken together these data point to an inductive role for sorbitol in Müller cell differentiation from RPCs.

Thus it remains unclear as to whether the pro-differentiation effects of sorbitol on PKCα expressing cells arise primarily from selectively enhancing the survival of bipolar precursor cells or by more directly influencing terminal differentiation, our results suggest that the induction of Müller cell differentiation of RPCs in part underlies the effect of sorbitol on vimentin and CRALBP expression. Single cell or clonal analyses of lineage experiments are needed to make this determination, and unfortunately, such studies have proven difficult with cultured RPCs.

In summary, our results show that sorbitol can promote expression of glial and bipolar cell markers along with simultaneous down-regulation of progenitor and proliferative markers. Up-regulation of cell-specific markers is accompanied by morphological differentiation consistent with glial phenotypes. Although much is known about sorbitol metabolism in the retina as it relates to diabetic retinopathy, this is the first study to investigate its relationship to Müller cells. While the ability to generate Müller glia in vitro from passaged late RPCs may have potential clinical implications, a better understanding of the changes underlying Müller cell activation and gliosis is required. We also suggest that the accumulating evidence for an important role of Müller cells in retinal regeneration demand that we further probe the developmental potential of this cell type.

Acknowledgements

This study was supported by a kind gift from Richard and Gail Siegal, and grants from US Dept. of Defense, National Institute of Health NEI 09595 (MJY), and Minda de Gunzburg Research Center. HK was supported by a grant from the Hoag Foundation. We thank the DSHB, University of Iowa for the nestin antibody; and Arshak Alexanian, Kameran Lashkari, Chiara Gerhardinger, Dong Feng Chen and Masumi Takeda for advice and assistance.

References

1. Cepko CL, Austin CP, Yang X, Alexiades M, Ezzeddine D. Cell fate determination in the vertebrate retina. Proc Natl Acad Sci U S A 1996; 93:589-95.

2. Hollander H, Makarov F, Dreher Z, van Driel D, Chan-Ling TL, Stone J. Structure of the macroglia of the retina: sharing and division of labour between astrocytes and Muller cells. J Comp Neurol 1991; 313:587-603.

3. Newman E, Reichenbach A. The Muller cell: a functional element of the retina. Trends Neurosci 1996; 19:307-12.

4. Bridges CD, Alvarez RA, Fong SL, Gonzalez-Fernandez F, Lam DM, Liou GI. Visual cycle in the mammalian eye. Retinoid-binding proteins and the distribution of 11-cis retinoids. Vision Res 1984; 24:1581-94.

5. Das SR, Bhardwaj N, Kjeldbye H, Gouras P. Muller cells of chicken retina synthesize 11-cis-retinol. Biochem J 1992; 285:907-13.

6. Fischer AJ, Reh TA. Exogenous growth factors stimulate the regeneration of ganglion cells in the chicken retina. Dev Biol 2002; 251:367-79.

7. Fischer AJ, Reh TA. Muller glia are a potential source of neural regeneration in the postnatal chicken retina. Nat Neurosci 2001; 4:247-52.

8. Fischer AJ, McGuire CR, Dierks BD, Reh TA. Insulin and fibroblast growth factor 2 activate a neurogenic program in Muller glia of the chicken retina. J Neurosci 2002; 22:9387-98.

9. Fischer AJ, Schmidt M, Omar G, Reh TA. BMP4 and CNTF are neuroprotective and suppress damage-induced proliferation of Muller glia in the retina. Mol Cell Neurosci 2004; 27:531-42.

10. Fischer AJ, Omar G, Eubanks J, McGuire CR, Dierks BD, Reh TA. Different aspects of gliosis in retinal Muller glia can be induced by CNTF, insulin, and FGF2 in the absence of damage. Mol Vis 2004; 10:973-86 <http://www.molvis.org/molvis/v10/a116/>.

11. Ooto S, Akagi T, Kageyama R, Akita J, Mandai M, Honda Y, Takahashi M. Potential for neural regeneration after neurotoxic injury in the adult mammalian retina. Proc Natl Acad Sci U S A 2004; 101:13654-9.

12. Stahl B, Wiesinger H, Hamprecht B. Characteristics of sorbitol uptake in rat glial primary cultures. J Neurochem 1989; 53:665-71.

13. Wiesinger H, Thiess U, Hamprecht B. Sorbitol pathway activity and utilization of polyols in astroglia-rich primary cultures. Glia 1990; 3:277-82.

14. Wiesinger H, Schuricht B, Hamprecht B. Replacement of glucose by sorbitol in growth medium causes selection of astroglial cells from heterogeneous primary cultures derived from newborn mouse brain. Brain Res 1991; 550:69-76.

15. Shatos MA, Mizumoto K, Mizumoto H, Kurimoto Y, Klassen H, Young MJ. Multipotent stem cells from the brain and retina of green mice. Journal of Regenerative Medicine 2001; 2:13-5.

16. Okabe M, Ikawa M, Kominami K, Nakanishi T, Nishimune Y. 'Green mice' as a source of ubiquitous green cells. FEBS Lett 1997; 407:313-9.

17. Turner DL, Cepko CL. A common progenitor for neurons and glia persists in rat retina late in development. Nature 1987; 328:131-6.

18. Klassen HJ, Ng TF, Kurimoto Y, Kirov I, Shatos M, Coffey P, Young MJ. Multipotent retinal progenitors express developmental markers, differentiate into retinal neurons, and preserve light-mediated behavior. Invest Ophthalmol Vis Sci 2004; 45:4167-73.

19. Lendahl U, Zimmerman LB, McKay RD. CNS stem cells express a new class of intermediate filament protein. Cell 1990; 60:585-95.

20. Ahmad I, Tang L, Pham H. Identification of neural progenitors in the adult mammalian eye. Biochem Biophys Res Commun 2000; 270:517-21.

21. Tropepe V, Coles BL, Chiasson BJ, Horsford DJ, Elia AJ, McInnes RR, van der Kooy D. Retinal stem cells in the adult mammalian eye. Science 2000; 287:2032-6.

22. Okada M, Matsumura M, Ogino N, Honda Y. Muller cells in detached human retina express glial fibrillary acidic protein and vimentin. Graefes Arch Clin Exp Ophthalmol 1990; 228:467-74.

23. Bignami A, Dahl D. The radial glia of Muller in the rat retina and their response to injury. An immunofluorescence study with antibodies to the glial fibrillary acidic (GFA) protein. Exp Eye Res 1979; 28:63-9.

24. Ekstrom P, Sanyal S, Narfstrom K, Chader GJ, van Veen T. Accumulation of glial fibrillary acidic protein in Muller radial glia during retinal degeneration. Invest Ophthalmol Vis Sci 1988; 29:1363-71.

25. Roberge FG, Caspi RR, Chan CC, Kuwabara T, Nussenblatt RB. Long-term culture of Muller cells from adult rats in the presence of activated lymphocytes/monocytes products. Curr Eye Res 1985; 4:975-82.

26. Wakakura M, Foulds WS. Immunocytochemical characteristics of Muller cells cultured from adult rabbit retina. Invest Ophthalmol Vis Sci 1988; 29:892-900.

27. Lupien C, Brenner M, Guerin SL, Salesse C. Expression of glial fibrillary acidic protein in primary cultures of human Muller cells. Exp Eye Res 2004; 79:423-9.

28. Guidry C. Isolation and characterization of porcine Muller cells. Myofibroblastic dedifferentiation in culture. Invest Ophthalmol Vis Sci 1996; 37:740-52.

29. Seigel GM, Mutchler AL, Imperato EL. Expression of glial markers in a retinal precursor cell line. Mol Vis 1996; 2:2 <http://www.molvis.org/molvis/v2/a2/>.

30. Mizumoto H, Mizumoto K, Shatos MA, Klassen H, Young MJ. Retinal transplantation of neural progenitor cells derived from the brain of GFP transgenic mice. Vision Res 2003; 43:1699-708.

31. Wojciechowski AB, Englund U, Lundberg C, Warfvinge K. Migratory capacity of the cell line RN33B and the host glial cell response after subretinal transplantation to normal adult rats. Glia 2004; 47:58-67.