Amphibians have a remarkable capacity to regenerate various organs and tissues including brain, limbs, and liver. Newt retina has been a great model to study neural retinal regeneration. Following removal of the entire neural retina, the eye can regenerate an entire complete and laminated new retina using the surrounding RPE cells (Stone 1950a, b; Stone and Steinitz 1957). The process involves proliferation and dedifferentiation of the RPE cells into an optic placode-like stem cell. These cells then regenerate a new bilayered structure, which has both a new neural retina and a new RPE cell layer. The newly formed retina has all the different cells including photoreceptors, bipolar cells, and ganglion cells as well as the inhibitory interneurons and glial cells. The only difference lies from the fact that the newly developed retina is inverted in orientation (Okada 1980). Recent studies have shown that the RPE transdifferentiation into neural cells occurs only in the presence of underlying connective tissue, i.e., choroid (Mitsuda et al. 2005). The results suggest the possibility of choroid playing an essential role in newt retinal regeneration potentially by means of diffusible substances. Two secretory factors that are upregulated in the choroid following newt retina removal include bFGF and IGF1. There is downregulation of Otx2, a key RPE gene, and upregulation of Notch pathway activity (Nakamura and Chiba 2007; Sakami et al. 2005). Some key regulators of this regeneration process include extracellular matrix components, such as laminin and heparin sulfate proteoglycans (Reh et al. 1987), as well as growth factors FGF-2 and IGF1 (Mitsuda et al. 2005; Sakaguchi et al. 1997), among others.

This phenomenon of RPE transdifferentiation also occurs in the embryonic chicken eyes (Coulombre and Coulombre 1965; Park and Hollenberg 1989; Pittack et al. 1991). When the neural retina is dissected out from the chicken embryos within the first 3–4 days of incubation (E3–E4), the RPE dedifferentiates and proliferates to form retinal progenitors, which then differentiate to form a new retina as in the newt. In the chick embryo, this process is stimulated by bFGFs with contribution from BMPs (Haynes et al. 2007) while activin signaling inhibits the regeneration (Sakami et al. 2008). Both FGFs and activin signaling are known to drive Pax6, a key transcription factor in retinal development. Recent experiments have shown that overexpression of Pax6 is sufficient to stimulate the production of neurons from the embryonic chicken RPE (Azuma et al. 2005). As opposed to the newt, in chickens, this RPE regenerative capacity is restricted to very early embryonic development. The mature chicken RPE does not contribute to retinal repair. So far, the data from mammalian RPE suggests very little contribution, if any, to retinal repair.

The peripheral margin of the retina where it meets the ciliary body has been thought of as another source of retinal stem cells and is called the ciliary marginal zone (CMZ). Retina of fish grows throughout life roughly matching the overall growth of the animal by the addition of new neurons from stem cells in the CMZ (Hollyfield 1968; Straznicky and Gaze 1971). In fact, most of the fish retina is generated from this CMZ as they grow to adult size. The CMZ cells resemble the embryonic retinal progenitor cells. The cells can generate clones of over 100 cells containing all the different retinal cell types (Wetts and Fraser 1988). Teleost fish also retain the ability to regenerate new retinal neurons following neurotoxin damage or surgical lesions throughout life (Hitchcock et al. 2004; Raymond et al. 2006). The chicken retina was largely believed to be completely developed at hatching. However, a study by Fischer and Reh showed that the CMZ is active in posthatch chickens (Fischer and Reh 2000, 2002). These cells at the retinal margin express genes common to embryonic retinal progenitors including Pax6. Injection of a combination of insulin IGF and EGF induces the cells in the CMZ to express progenitor markers and undergo proliferation. Similarly, sonic hedgehog induces proliferation in the adult chicken CMZ (Moshiri et al. 2005). A similar active CMZ has also been demonstrated in adult quails (Kubota et al. 2002).

Several reports have shown that cells isolated from the ciliary epithelium of several mammals, including mice, pig, and humans, can give rise to proliferative clones (neurospheres) when cultured in vitro (Ahmad et al. 2000; Asami et al. 2007; Coles et al. 2004; Kohno et al. 2006; MacNeil et al. 2007; Sun et al. 2006; Tropepe et al. 2000). The cells express a variety of neural progenitor markers and have been shown to undergo extensive proliferation. Pax6, neural progenitor gene, is required for the generation of these neurospheres (Xu et al. 2007). Upon exposure to differentiation conditions, these spheres were shown to express rod photoreceptors, bipolar and Müller glial genes. The sphere proliferation can be enhanced by exposure to Wnts, SCF, and PEDF (Inoue et al. 2006; Das et al. 2004, 2006a; De Marzo et al. 2010). Despite this early hope, some recent reports have suggested that these spheres may not completely resemble retinal stem cells. A detailed analysis by Mike Dyer’s group showed that all cells in the CMZ-derived spheres, including those expressing neural progenitor markers, were still pigmented and exhibited other characteristics of normal pigmented cells (Cicero et al. 2009). Thus, the jury is still out on the existence of a CMZ in humans, which can be exploited for native retinal repair.

As discussed above, the fish retina contains a CMZ at the retinal margin, and it contributes to retinal growth as the fish increase in size. However, this zone is not thought to contribute to regeneration in central retina. Instead, fish contains two other quiescent progenitor populations, the rod progenitor cells and the Müller glia (Johns 1982). Upon injury, fish retinal Müller glia slowly start dividing to generate the rod progenitor cells (Bernardos et al. 2007). Using transgenic fish with GFP expressed in Müller glia, several labs have demonstrated that these cells are the primary source of regenerated neurons in fish central retina. Following damage to retinal neurons, the retina Müller glia reenter the cell cycle and dedifferentiate toward a progenitor-like phenotype (Raymond et al. 2006; Bernardos et al. 2007; Faillace et al. 2002; Fausett and Goldman 2006; Fimbel et al. 2007; Kassen et al. 2007; Thummel et al. 2008; Vihtelic et al. 2006; Wu et al. 2001; Yurco and Cameron 2005). A number of groups are now attempting to better understand the processes involved in reentry of the glial cells into cell cycle with the hopes of applying it to mammalian regeneration (Raymond et al. 2006; Kassen et al. 2007; Cameron et al. 2005; Yurco and Cameron 2007). It has been demonstrated that not all glial cells have the capacity to regenerate the retina but that these cells are asymmetrically distributed (Julian et al. 1998). Dan Goldman’s group has identified Ascl1, a basic helix-loop-helix gene expressed highly in retinal progenitors, as one that gets upregulated in Müller glia upon retinal damage (Fausett et al. 2008). They showed using morpholinos that inhibition of Ascl1 blocks Müller glial dedifferentiation and their subsequent reentry into the cell cycle. This is also interesting as a recent study in mammalian retina demonstrated the Ascl1 expression defines a subpopulation of lineage-restricted progenitors in the mammalian retina (Brzezinski et al. 2011). The process is mediated by EGF/MAPK/Notch cascade (Wan et al. 2012) and Wnt/β-catenin signaling (Ramachandran et al. 2011). A recent study looked for the inducer of Müller glial proliferation following damage. They found that TNF-α was produced by dying retinal neurons and was necessary for expression of ASCL1 in Müller glia (Nelson et al. 2013). Another secreted factor, which has been shown to induce glial proliferation in the absence of injury, is CNTF. It does this through a STAT3-dependent mechanism (Kassen et al. 2009). Injury-induced Ascl1 directly regulates lin-28, a microRNA-binding protein. Lin-28 then represses let-7 miRNA, which is normally involved in repression of regeneration-associated genes including ASCL1, PAX6, OCT4, KLF4, and c-MYC (Ramachandran et al. 2010). A detailed characterization study by Pam Raymond’s group into the genes involved in the zebrafish regenerative process showed that cells upregulated the Notch–Delta pathway as well as key retinal progenitor genes, Pax6a, Rx, and Chx10, upon reentry into the cell cycle (Raymond et al. 2006). These data suggest that regeneration and development use the similar molecular mechanisms and revoking a similar pathway in mammalian glial cells may restore regenerative capacity.

Like the fish retina, chicken retina displays some regenerative capacity (Fischer and Reh 2001). The earliest demonstration of this was using neurotoxic damage of inner retinal neurons with N-methyl-d-aspartate (NMDA) in posthatch chicks (Fischer and Reh 2001). Müller glia reenter the cell cycle and dedifferentiate into retinal progenitors (Fischer and Reh 2001, 2003; Fischer et al. 2002; Hayes et al. 2007). These Müller glial-derived retinal progenitors differentiate into amacrine cells, bipolar cells, and ganglion cells. There was no evidence of photoreceptor cell differentiation. A close examination of the Notch–Delta pathway revealed a dual role in the regenerative response (Hayes et al. 2007). Notch upregulation was indeed necessary for the glia to dedifferentiate and reenter cell cycle. However, after this initial need, Notch inhibition is necessary for differentiation of the Müller glial-derived retinal progenitors to form new neurons.

As opposed to fish and avian retinas, the mammalian Müller glia do not spontaneously reenter the cell cycle following retina damage (Chang et al. 2007; Close et al. 2006; Dayer et al. 2005; Sahel et al. 1990; Zhao et al. 2005). In order to get them to proliferate following damage, they require additional stimulation including growth factors. Ooto et al. used NMDA to induce damage in the rat inner retina (Ooto et al. 2004). They found that Müller glia underwent proliferation, though it was limited. They detected some of the Müller glial-derived progeny, which expressed markers of bipolar cells and photoreceptors. This number could be increased by co-injecting retinoic acid. Alternatively, overexpression of certain transcription factors including Math3, NeuroD1, Pax6, and Crx could generate bipolar and amacrine cells and photoreceptors. A number of additional publications have also reported similar BrdU-expressing photoreceptor cell generation following either retinal damage or growth factor treatment in adult rodents (Close et al. 2006; Das et al. 2006b; Osakada et al. 2007; Wan et al. 2007; Wan et al. 2008; Del Debbio et al. 2010). To gain a better handle on understanding the regenerative process in mouse retinas, Karl et al. conducted a detailed examination of the progeny derived from glia following toxic damage and EGF stimulation in various transgenic strains of mice to track various lineages (Karl et al. 2008). Using this protocol, up to 10 % of Müller glia entered the cell cycle within 3 days. They found that almost all of the Müller glial-derived progenitors differentiated to amacrine cells. Consistent with this observation, they found that FoxN4, a transcription factor necessary for the amacrine cell differentiation, was one of the most highly upregulated genes in both the mouse and chick retina after NMDA-induced damage. Other studies using Wnt3a or Shh following damage have reported photoreceptor cell differentiation (Osakada et al. 2007; Wan et al. 2007, 2008). There have been no studies of functional regeneration in either the chicken or the mammalian retina. Thus, it is still not known whether Müller glial-derived cells functionally integrate into the retinal circuitry.

Recently, the field of direct reprogramming has come into a lot of focus. Direct reprogramming or induced transdifferentiation refers to the process that allows cells to directly progress from one cell fate to another without an intermediate stem cell fate. The earliest landmark paper describing this was the direct conversion of fibroblast cells to myoblasts using a single transcription factor MyoD1 (Davis et al. 1987). More recently, a similar transdifferentiation strategy was used to generate neurons from fibroblast using ASCL1, BRN2, and MYT1L (Vierbuchen et al. 2010). A recent publication from Tom Reh’s lab attempted to employ the strategy to restore retinal stem cell characteristics to mammalian Müller glia (Pollak et al. 2013). As described above, Ascl1a is required for retinal regeneration in the fish and is rapidly upregulated in proliferating Müller glia. However, Ascl1 is not upregulated in the mouse retina after N-methyl-d-aspartate (NMDA)-induced damage (Karl et al. 2008). Therefore, the authors induced expression of ASCL1 in Müller glia in either dissociated cultures or explant retinas (Pollak et al. 2013). They found that ASCL1 was sufficient to induce neural differentiation of the glial cells. ASCL1 remodeled the chromatin at retinal progenitor gene sites activating their expression while at the same time downregulating glial genes. The reprogrammed cells expressed early postmitotic retinal markers including Otx2, Islet1, Neurod1, Prox1, and Crx. Thus, direct reprogramming could serve as an alternate strategy to induce retinal regeneration in mammalian retinas.