SEARCH

SEARCH BY CITATION

Keywords:

  • regeneration;
  • eye;
  • lens;
  • retina;
  • transdifferentiation;
  • stem cells

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. LENS REGENERATION
  5. RETINA REGENERATION
  6. EYE REGENERATION IN INVERTEBRATES
  7. PERSPECTIVE
  8. Acknowledgements
  9. REFERENCES

Eye tissues such as the lens and the retina possess remarkable regenerative abilities. In amphibians, a complete lens can be regenerated after lentectomy. The process is a classic example of transdifferentiation of one cell type to another. Likewise, retina can be regenerated, but the strategy used to replace the damaged retina differs, depending on the animal system and the age of the animal. Retina can be regenerated by transdifferentiation or by the use of stem cells. In this review, we present a synthesis on the regenerative capacity of eye tissues in different animals with emphasis on the strategy and the molecules involved. In addition, we stress the place of this field at the molecular age and the importance of the recent technologic advances. Developmental Dynamics 226:211–224, 2003. © 2003 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. LENS REGENERATION
  5. RETINA REGENERATION
  6. EYE REGENERATION IN INVERTEBRATES
  7. PERSPECTIVE
  8. Acknowledgements
  9. REFERENCES

More than a 100 years ago, Philipeaux (1880), Griffini and Marchio (1889), Colucci (1891), and Wolff (1895) provided the first solid documentation for a remarkable type of regeneration. These scientists showed that the retina and the lens of adult newts could be completely regenerated after removal or damage. Over a century since the initial observations, the field of eye regeneration has evolved and expanded in many different animal models. Although lens regeneration seems to be the property of amphibians (Stone, 1967) and occurs exclusively by transdifferentiation, regeneration of the retina is possible in many other species as well, such as fish, avian, and possibly mammals. In addition, regeneration of the retina seems to be achieved by more complex mechanisms, involving both transdifferentiation and stem cell differentiation.

Despite these spectacular properties of regeneration of eye tissues, the field of eye regeneration, as represented by lower vertebrate animals such as amphibia, reacted slowly to the developments in molecular biology techniques. Part of the reason was that amphibia were not (and some species still are not) the best models for transgenesis and molecular manipulation. Because of that, the molecular mechanisms for lens and retina regeneration were virtually unknown. Similar scenarios were true for birds and fish until recently, when these animal models were made more technologically friendly. During the past few years, several laboratories, including ours, have initiated systematic molecular studies in the field of lens and retina regeneration. The purpose of this review is to update the reader of the latest molecular advances in the field of eye regeneration. As it will become apparent, important progress has been recorded in all fronts. In this review, we will examine the different animal models used in the study of lens and retina regeneration. For each model, the known cellular and molecular mechanisms will be mentioned and compared. This synthesis will close with future perspectives in the research of eye regeneration and the impact on basic and medical sciences.

LENS REGENERATION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. LENS REGENERATION
  5. RETINA REGENERATION
  6. EYE REGENERATION IN INVERTEBRATES
  7. PERSPECTIVE
  8. Acknowledgements
  9. REFERENCES

As mentioned above, lens regeneration was first observed in adult newts. These animals have been the major experimental material for the study of lens regeneration. Lens regeneration is also possible in certain frogs, but the process differs considerably from the one in newts. In these frogs, lens regeneration is possible only during the premetamorphic stages of development (see below). Adult frogs are not capable of regeneration. The only adult animals with that capability are some urodeles. In the following section, we will examine the two animal models and compare the mechanisms in both.

Lens Regeneration in Adult Newts

Lens regeneration in the adult newt begins with proliferation and dedifferentiation of dorsal iris pigment epithelial cells (PECs). By dedifferentiation, we mean the loss of characteristics that define the pigment epithelial cells, such as pigmentation. Dedifferentiation initiates molecular events, such as cell cycle re-entry, which are necessary for cell proliferation and the subsequent regeneration of the lens. At approximately 10 days postlentectomy, a lens vesicle is formed from the depigmented dorsal PECs (Fig. 1A). Around 12 to 16 days postlentectomy, the internal layer of the lens vesicle thickens and synthesis of crystallins begins (Fig. 1B). This finding marks the beginning of primary lens fiber differentiation. During days 15 to 19, proliferation and depigmentation of PECs slows down. In the internal layer, the lens fiber complex is formed and in the margin of the external layers nondividing secondary lens fibers appear. By 18 to 20 days, the PECs have stopped proliferating and the lens fibers continue to accumulate crystallins (Fig. 1C). Lens regeneration is considered complete by day 25–30 (Fig. 1D; Eguchi, 1963, 1964; Reyer, 1977; Yamada, 1977; Tsonis, 1999, 2000). Lens regeneration, therefore, is possible by transdifferentiation, which is the transformation of one cell type to another (in this case, pigment epithelial cells to lens cells). A very interesting restriction is that the ventral iris, which is seemingly composed of the same PECs, is not capable of regenerating a lens. Lens regeneration is only a property of PECs from the dorsal iris. The ventral iris has no potential to regenerate a lens. The process of transdifferentiation has been proven beyond any doubt in this system. These processes can also be observed when single PECs are placed in culture (Eguchi et al., 1974; Kodama and Eguchi, 1995; Tsonis et al., 2001). The restrictions that we see in the in vivo newt model do not apply for the in vitro models. PECs from the whole eye (including from the ventral iris) of any species, including aged humans, are capable of transdifferentiating to lens cells under certain conditions. As the PECs proliferate, they become depigmented and then transdifferentiate to lens cells. The transdifferentiation process takes several weeks to occur, and during that time, the cells have established monolayers. Therefore, although in many other regenerative tissues stem cells may play a role, such a possibility is very unlikely for lens regeneration.

thumbnail image

Figure 1. Series of histologic sections depicting the process of lens regeneration in the adult newt. A: Dorsal iris 10 days after lentectomy. The tip of the dorsal iris (arrowhead) has depigmented and dedifferentiated. This finding is the beginning of the formation of the lens vesicle. B: Dorsal iris 15 days after lentectomy. The internal layer of the lens vesicle thickens, and differentiation of the lens fibers starts. C: Regenerating lens 20 days after lentectomy. Lens differentiation is almost complete. The lens epithelium is covering the lens in the anterior, whereas lens fiber differentiation is prominent in the posterior part of the lens. D: Regenerated lens 25 days after lentectomy showing a definite lens. The purple color of the regenerating lens vesicle/lens in A–C is due to detection of FGF-1 expression and, in D, is due to detection of gamma crystallin. Note in D that the lens epithelium is negative for the lens fiber-specific gamma crystallin. From Tsonis, PA (1998; Regeneration of the Vertebrate Lens and Other Eye Structures. In: Encyclopedia of Life Sciences, pp. 164–169. London: Nature Publishing Group. [doi: 10.1038/npg.els.0001102] www.els.net).

Download figure to PowerPoint

The spatial regulation in newt lens regeneration from the dorsal iris has been the focus of molecular work in our laboratory. We reasoned that, if we could identify factors restricted to dorsal (or ventral) iris, we might delineate the molecular mechanisms of lens regeneration and apply those to other animal systems. Our strategy was to examine key regulatory genes for specific expression in the dorsal iris. When these genes are identified, we could use them to induce regeneration from the ventral iris. Successful outcome of this strategy would then open new avenues in experimenting with other animals. During the past few years, we have examined expression and roles of several Hox genes, pax-6, FGFs, FGFRs, RARs, and prox-1. All these factors are expressed in the differentiating lens vesicle and the regenerating lens. However, the strongest evidence for spatial regulation was received from the studies involving pax-6, prox-1, and FGFR-1 (Del Rio-Tsonis et al., 1995, 1997, 1998, 1999; McDevitt et al., 1997; Mizuno et al., 1999b). All of these factors were found to be specific in the dorsal iris. The products of prox-1 and FGFR-1 were found to be regulated in the dorsal iris as well (Del Rio-Tsonis et al., 1998, 1999, Mizuno et al., 1999b). In addition, an inhibitor to FGFR-1 was able to abolish regeneration from the dorsal iris (Del Rio-Tsonis et al., 1998). Another interesting result was obtained by inhibiting the activity of RARs with a specific inhibitor. In some cases, ectopic lens regeneration resulted from sites other than the dorsal iris (Tsonis et al., 2000, 2002). Studies under way in our laboratory now aim at examining the possible role of these genes in the induction of lens regeneration from the ventral iris by means of transgenesis.

To create transgenic ventral irises, however, is not easy. Injections of retroviral vectors to target directly the ventral iris are almost impossible. One possibility is to target vectors by means of biolistics. This strategy is something that we are pursuing; however, the efficiency is very low. Recently, Hayashi et al. have devised an in vitro/in vivo system that recapitulates faithfully the in vivo conditions and is suitable for efficient genetic manipulations (Ito et al., 1999; Hayashi et al., 2001, 2002). Isolated PECs from dorsal or ventral iris are cultured for approximately 2 weeks. Then the cells are aggregated and implanted in a lentectomized eye. As it happens during in vivo lens regeneration, only the dorsal aggregates can transdifferentiate to form lenses (Fig. 2A). The ventral aggregates fail to form lenses (note here that ventral iris PECs have to be cultured for longer periods of time and form a monolayer to transdifferentiate to lens cells; Fig. 2B). By using this system, during the culturing period, transfection of genes with efficiency as high as 80% is possible. Ventral transgenic aggregates, thus, can be created with genes that seem to control transdifferentiation of the dorsal iris. This can bypass the serious problem of in vivo transgenesis that has not been established in the newt. These ventral iris PECs can then be transplanted in host eyes and examined for their ability to transdifferentiate to lens! This method is an excellent system that is likely to yield valuable results in the study of induction of lens regeneration. Likewise, transcription of a gene can be inhibited in vitro by morpholinos, thus creating a knockdown condition in the system. Once a candidate gene has been identified as an inducer of lens regeneration, more rigorous studies can be pursued using other animal systems.

thumbnail image

Figure 2. Implantation of dorsal and ventral aggregates into a lentectomized newt eye collected 20 days later. A: Dorsal aggregate was implanted. B: Ventral aggregate was implanted. Note that the dorsal aggregate differentiated into a lens vesicle (arrow in A), whereas the ventral stayed as a mass of pigmented cells (arrow in B). The host lens (arrowheads in A,B) regenerated normally from the dorsal iris.

Download figure to PowerPoint

Lens Regeneration in Frogs

After lentectomy in Xenopus laevis larvae, the lens can be regenerated from the inner layer of the outer cornea (Freeman, 1963; Filoni et al., 1997). This capacity gradually decreases between stages 50 and 58 and disappears after metamorphosis. This seems to be another case of transdifferentiation from cornea to lens cells. The factor(s) responsible for this type of transdifferentiation seem to reside in the neural retina and accumulate in the vitreous chamber (Filoni et al., 1982). In the normal eye, these factors cannot reach the outer cornea, because the lens and the inner cornea act as barriers. However, upon lentectomy, the factors become accessible. This finding has been shown in experiments where a piece of outer cornea can transdifferentiate to lens when placed in the vitreous chamber even in the presence of the host lens (Reeve and Wild, 1978). The decrease in regenerative capacity after the animals metamorphose could be related to the inhibition of the process by the inner cornea. It is possible that the rapid closure of the inner cornea is responsible for the decrease in the regenerative capacity, because cornea implants lacking the inner region from metamorphosed frogs placed into host larvae eyes were able to form lens fibers (Filoni et al., 1997). Also, in a related species, Xenopus tropicalis, cornea-to-lens transdifferentiation is possible but it occurs at a very low rate because of the rapid recovery of the inner cornea endothelium (Henry and Elkins, 2001). When outer cornea was cultured in vitro, transdifferentiation was induced by the presence of FGF-1. This effect of FGF-1 was not mediated by means of its mitogenic activity, because it was observed even when inhibitors of DNA synthesis were added in the culture medium (Bosco et al., 1997). This finding is interesting because FGF has been linked to the induction of transdifferentiation of the PECs to lens in vitro (Hyuga et al., 1993) and to the morphogenesis of lens during development (De Iong and McAvoy, 1993). Other genes that are expressed during the cornea to lens transdifferentiation are pax-6, prox-1, otx-2, and sox-3 (Schaefer et al., 1999; Henry et al., 2002). All these genes are important and are expressed during embryogenesis as well.

Xenopus is a suitable animal model for transgenesis studies and superior (at the moment) to newt. Several strategies have been developed to create transgenic frogs, and recently, the use of morpholinos to specifically inhibit gene action has gained popularity (Heasman, 2002). Also, frog embryos are easier to rear and to manipulate than are newt embryos. The disadvantage of the frog animal system is that lens regeneration does not occur in the adult. However, the difference between Xenopus pre- and post-metamorphic regenerative capacities could be beneficial in elucidating regeneration specific molecules and aiding in the determination of the mechanism behind the ability to regenerate. In addition, the two amphibian model systems could be used in conjunction to complement each other. Comparative studies are extremely important and largely lacking in the field.

Lens Regenerative Abilities in Other Vertebrates

Some species of adult fish have been determined to regenerate their lens by means of transdifferentiation of the dorsal iris, similarly to salamanders (Sato, 1961; Mitashov, 1966). Within other classes of vertebrates, chicks have been reported to be able to replace their lens upon removal, but only during a small window of their development. It is not clear if these embryos actually regenerate their lenses by means of transdifferentiation of the iris PECs or if it is a result of an inductive event from the still competent ectoderm by the optic vesicle (Van Deth, 1940; McKeehan, 1961; Génis-Gálvez, 1962; Niazi, 1967; Wedlock and McCallion, 1968). Also there have been several reports claiming that rabbits and cats are able to regenerate their lenses upon removal, but only if the anterior and posterior capsular bags are left intact. The regenerates seem to be generated from growth of lens epithelial cells left in the capsular bags (Stewart and Espinase, 1959; Gwon et al., 1989, 1990, 1992, 1993a, b). It seems that lens regeneration in rabbits is rather the property of lens epithelial cells that differentiate to a viable lens than a real regeneration by transdifferentiation. However interesting these cases of regeneration might be, molecular biology advances in fish (except zebrafish) and rabbit are virtually nonexistent. From that point of view, these animal models are not favorable for molecular studies.

Is Lens Regeneration a Repeat of Lens Development?

This question surfaces every time that a gene known to control lens development is also expressed during regeneration. Obviously, the end result is the same in both instances. Conceivably, genes that control the formation of the lens will be expressed in both processes. But we have to be careful and exercise caution when equating development and regeneration of the lens. This is especially true in the case of the newt, where the lens is regenerated by an already terminally differentiated tissue. During development, the lens is induced by interactions between the ectoderm (not terminally differentiated) and the optic vesicle. It is very possible that what induces the PECs of the dorsal iris to dedifferentiate is quite unique and not related to lens development. Dedifferentiation is the key event for regeneration to occur. After a vesicle has been formed, lens morphogenesis might take its course as in lens development. That is why there seems to be conservation in the expression of many genes that are important for lens morphogenesis, such as, pax-6, FGFs, and FGFRs, prox-1, sox-3, otx-2 (Del Rio-Tsonis et al., 1997, 1999; Schaefer et al., 1999; Henry et al., 2002). Even when it comes to the expression of crystallins, which are markers of lens differentiation, the expression patterns are not exactly identical. In the newt, expression of alpha, beta, and gamma crystallins is very similar in both events. In contrast, during Xenopus cornea-to-lens transdifferentiation, gamma crystallin expression is detected later than in the normal developing lens and is only in morphologically discernible lens fibers (Mizuno et al., 1999a, 2002).

RETINA REGENERATION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. LENS REGENERATION
  5. RETINA REGENERATION
  6. EYE REGENERATION IN INVERTEBRATES
  7. PERSPECTIVE
  8. Acknowledgements
  9. REFERENCES

Retina regeneration takes place in a variety of organisms during a restricted period in early development. These organisms include fish, birds, mammals, and amphibians. Regeneration of the retina in these embryos occurs mainly by transdifferentiation of the pigment epithelial cells of the retina (Fig. 3, #1). Only certain urodeles retain the ability to regenerate their retinas by means of transdifferentiation as adults (Lopashov and Stroeva, 1964; Mitashov, 1996, 1997; Raymond and Hitchcock, 2000; Fischer and Reh, 2001a). Other adult amphibians, fish, and birds have been shown to partially regenerate their retinas by means of the use of precursors present in the ciliary marginal zone (CMZ, Fig. 3, #2), in other areas of the retina (Fig. 3, #3 and #4) or by Müller glia cells (Fig. 3, #5; Lopashov and Stroeva, 1964; Johns and Fernald, 1981; Johns, 1982; Raymond et al., 1988; Braisted et al., 1994; Mitashov, 1996, 1997; Raymond and Hitchock, 1997, 2000; Julian et al., 1998; Reh and Levine, 1998; Fischer and Reh, 2000, 2001b; Vihtelic and Hyde, 2000; Reh and Fischer, 2001; Otteson et al., 2001; Wu et al. 2001). Recently, the possibility that mammals could regenerate their retina has been opened with the discovery of adult pigmented cells from the ciliary margin (PCM) of rats that have the capability to transdiffrentiate in vitro into retinal-specific cells (Fig. 3, #6) (Ahmad et al., 2000; Tropepe et al., 2000). The objective of this section is to present the different regenerative mechanisms that have evolved in the different vertebrate species with an emphasis on the cellular and molecular events that take place and to discuss a perspective on the possible clinical applications of this research.

thumbnail image

Figure 3. Diagram of a vertebrate eye pinpointing the regions of the retina that contribute to the process of retina regeneration in the different model organisms. #1: Retinal pigment epithelium (RPE, red) can transdifferentiate into neural retina in a many embryonic species, but as adult only in the newt. #2: Ciliary marginal zone (CMZ, blue) is a source of stem cells and precursors that contribute to the ongoing growth of the retina as well as during regeneration in many organisms. #3: Rod precursors (red/spherical) contribute to rod photoreceptor regeneration in teleost fish. #4: Intrinsic stem cells (red/fusiform) present in the inner nuclear layer (INL) that proliferate and give rise to neural progenitors that migrate to the outer nuclear layer (ONL) and replace damaged retinal cells. #5: Müller glia cells (red/cell body and axons) that proliferate and migrate to the ONL to replace damaged retina. #6: pigmented cells of the ciliary margin (PCM, pink) cells can transdifferentiate in vitro in mice. GCL, ganglion cell layer; CGZ, circumferential germinal zone.

Download figure to PowerPoint

The Retina

To understand how the retina regenerates, it is essential to understand how this complex tissue is organized. The central retina is divided into multiple layers (Fig. 4; Litzinger and Del Rio-Tsonis, 2001). The first layer, the retinal pigment epithelium (RPE), consists of pigment epithelial cells that are highly specialized to act as supportive cells for the other retinal cells, especially the photoreceptors that lie just above this layer. The photoreceptor layer contains the outer and inner segments of the rods and cones. The nuclei of these photoreceptor cells reside in the outer nuclear layer (ONL) and their axons in the outer plexiform layer (OPL). The inner nuclear layer (INL) that follows contains the nuclei of bipolar cells, horizontal cells, Müller glia cells, and the majority of the amacrine cells. Vertical communication between the bipolar cells and the ganglion cells takes place in the next layer, the inner plexiform layer (IPL). The ganglion cell layer that lies above the IPL contains the cell bodies of the ganglion cells as well as displaced amacrine cells. The dendrites of the ganglion cells extend into the IPL, whereas their axons extend in the opposite direction to make the nerve fiber layer that eventually makes up the optic nerve. The peripheral region of most vertebrate retinas consists of a CMZ that functions in many species to accommodate for the expansion of the growing postnatal eye. This region contains an RPE layer under a simpler epithelium that contains retinal progenitor cells followed by a region of putative stem cells mostly located in the region of the ciliary body, which eventually connect to the iris of the eye (Perron et al., 1998; Fischer and Reh, 2000; Reh and Fischer, 2001).

thumbnail image

Figure 4. Representation of the structure of the vertebrate retina. All retinal layers are represented in the illustration and correspond to the layers of the adjacent section from a newt eye. From Litzinger, TC, and Del Rio-Tsonis, K (2001; Eye Anatomy. In: Encyclopedia of Life Sciences. London: Nature Publishing Group. [doi:10.1038/npg.els.0000108] http://www.els.net/).

Download figure to PowerPoint

Retina Regeneration in Amphibians

Embryonic and larval stages.

During embryogenesis, any damaged retina will most likely be replaced by transdifferentiation of the RPE, as it occurs in several larval amphibians and some adult urodeles (reviewed by Reyer, 1977; Mitashov, 1996, 1997). For example, Rana catesbiana larvae have been shown to undergo transdifferentiation of RPE after retinal damage (Reh and Nagy, 1987). In other species such as Xenopus and some members of the Rana family, the RPE has been induced to transdifferentiate into retina under special conditions, such as transplantation of RPE into eye cavities of hosts of the same species, as well as enhanced in vitro culture conditions (Loposhov and Sologub, 1972; Sologub, 1975; Sakaguchi et al., 1997). The regenerative ability of these amphibians decreases as the animal grows, except in some urodele families, including the Salamandriade family (reviewed by Reyer, 1977; Mitashov, 1996, 1997).

Neurogenesis in the retina of amphibians occurs partly during embryogenesis but mostly takes place postembryonically by means of the addition of new cells from the CMZ (reviewed by Reh and Levine, 1998; Reh and Fischer, 2001). Because most of the retinal growth in amphibians occurs during the larval stages (Hollyfield, 1971), it seems logical that the CMZ could also provide cells to replace damaged retina (Fig. 3, #2). This is the case for some tadpoles where their retinas have been damaged either mechanically or chemically (Dabagyan and Sheresheva, 1966; Dabagyan and Tret'yakova, 1968; Reh and Tully, 1986; Reh, 1987; Reh and Nagy, 1987). There are also reports of a group of proliferating cells that are present in the INL (Fig. 3, # 4) of larval Xenopus retina that have been implicated in retina regeneration (Levine, 1981).

Adult stages.

As mentioned, some urodeles retain the capability to regenerate their retinas throughout their lifetime. These salamanders regenerate their retinas by means of the transdifferentiation of retinal pigmented epithelial cells (rPEC) cells (Fig. 3, #1; Fig. 5), even though some regeneration by means of the CMZ (Fig. 3, #2) does occur at the periphery of the eye (Mitashov, 1968; Keefe, 1971, 1973; Reyer, 1971). Also, proliferating cells in the INL of newts (Fig. 3, #4 or #5) have been reported to replace damaged retina (Grigorian et al., 1996). On the other hand, in salamanders such as the axolotl and in frogs such as Xenopus, the retina will only partially regenerate as a small patch at the periphery of the retina by means of the use of precursors in the CMZ (Mitashov, 1968; Levine, 1981; Mitashov and Maliovanova, 1982; Svistunov and Mitashov, 1983, 1985). Because the major source of regeneration in adult urodeles is the RPE, we will concentrate on the process of transdifferentiation.

thumbnail image

Figure 5. Retina regeneration in adult newts. A: Intact retina section showing all the layers of a mature retina. B: Retinectomized newt eye (5 days postretinectomy). Here, dedifferentiation and proliferation of the retinal pigmented epithelial cells begin. Note that the cells shown by arrowheads have dedifferentiated or are in the process of dedifferentiation. C: After approximately 2 weeks postretinectomy a neuroepithelial (ne) cell layer forms that will eventually give rise to all the cells of the retina. D: A month postretinectomy, the regenerated differentiated retina stratifies into the different retinal layers: the outer nuclear layer (o), the inner nuclear layer (i), and the ganglion cell layer (g). The retinal pigment epithelium (RPE) has also been renewed, and the orientation of the newly formed retina is the same as the intact. Sections were stained with hematoxylin and eosin.

Download figure to PowerPoint

Transdifferentiation of the RPE

Retina regeneration was first described in adult newts more than a hundred years ago (Philipeau, 1880; Griffini and Marcchio, 1889). Since then, a large number of studies have been done to try to understand this amazing process. Upon retinectomy (removal of the retina) or ischemia-induced retinal degeneration, performed by enucleating the eye and resetting it back unto the eye socket, the pigment epithelial cells of the retina undergo a process of transdifferentiation. This process entails the retinal pigment epithelial cell (Fig. 3, #1; Fig. 5B), a “terminally differentiated cell type,” loosing its characteristics of origin (loss of pigment and detachment from the basement membrane, for example) and entering the cell cycle to form a neuroepithelial layer of cells (Fig. 5C) that eventually will differentiate into all the different cell types of the retina (Fig. 5D; Stone, 1950a; Lopashov and Sologub, 1972; Reyer, 1977; Stroeva and Mitashov, 1983; Reh and Nagy, 1987). During transdifferentiation, rPEC pass through a transitional state at which time they express both RPE and neuronal markers (Reh and Nagy, 1987; Klein et al. 1990; Grigorian and Anton, 1995; Grigorian, 2001). The rPEC not only give rise to the neuroepithelial layer of cells, but they also renew themselves by forming a monolayer of cells that lie under the new neuroepithelium. This neuroepithelium then differentiates, recapitulating the appearance of retinal cells during development (Fig. 5A,D; reviewed in Reyer, 1977; Hitchcock and Raymond, 1992; Mitashov, 1996, 1997; Raymond and Hitchcock, 2000; also see Cheon et al., 2001; Sakakibara et al., 2002). This sequential retinal differentiation has been followed by using retinal cell-specific markers, including opsin and visinin (photoreceptors), sodium channel alpha subunit (newt ganglion cells), and glial fibrillar acidic protein for Müller glia cells (Bugra et al., 1992; Negishi et al., 1992; Mitashov et al., 1995; Cheon et al., 1998; Sakakibara et al., 2002). If the eye of the newt underwent ischemia-induced retinal degeneration, the existing retina must first degenerate and recruit additional cells to clear up the retinal debris. Migrating macrophages are joined by some of the existing pigment cells to phagocytose cell debris (Keefe, 1973b; and reviewed by Mitashov, 1996, 1997).

A glance into the molecular age of transdifferentiation

The process of retina regeneration has been extensively studied at the histologic and electrophysiological/biochemical level in amphibians (Wachs, 1920; Stone 1950a, b; Hasegawa, 1958; Dabagian and Oganessian, 1970; Keefe, 1973a, c; Lam, 1977; Sarthy and Lam, 1983; Kaneko and Saito, 1992; Kaneko et al., 1993; Saito et al., 1994; Chiba and Saito, 1995; Chiba et al., 1997; Sakai and Saito, 1997). In addition, in vitro and in vivo studies have pinpointed key regulators of this regeneration process, including extracellular matrix components such as laminin (Reh et al., 1987) and heparin sulfate proteoglycans (Nagy and Reh, 1994), as well as growth factors such as FGF-2 (Sakaguchi et al., 1997). However, the detailed cellular and molecular events that drive this process still need to be clearly dissected. The field has taken some forward steps with the increase availability of tools, such as antibodies and cDNAs (Klein et al., 1990; Bugra et al., 1992; Negishi et al., 1992; Mitashov et al., 1995; Saito et al., 1994; Chiba et al., 1997; Cheon et al., 1998; Kaneko et al., 1999, 2001; Sakakibara et al., 2002). Further steps need to be taken to bring this field into the molecular age, including the use of new technology, such as morpholinos and RNAi to create knockdown animals (Heasman, 2002; Zhou et al., 2002), the use of viral vectors (Burns et al., 1994; Roy et al., 2000; Kumar et al., 2000) to bypass the production of transgenic animals (very difficult to produce in urodeles) by creating “on site transgenics,” or by the use of tissue/organ cultures that could be manipulated with the addition of key factors or genes (Ikegami et al., 2002). Also, expressed sequence tag studies, which can characterize gene expression patterns during the process of regeneration, are important and should reveal new genes responsible for the process (now under way in several labs, especially with Xenopus; Henry et al., 2002).

Retina Regeneration in Fish

Embryonic stages and larval stages.

Neurogenesis in the retina of fish is similar to that of amphibians in that most of the development occurs postembryonically (reviewed by Reh and Levine, 1998; Raymond and Hitchcock, 2000; Reh and Fischer, 2001). Fishes do have a CMZ (Fig. 3, #2), also referred to as the circumferential germinal zone (CGZ), that allows the retina to expand as development of the retina proceeds and can produce all cell types of the retina except rod photoreceptors, which are produced by rod precursors present in the ONL (Fig. 3, #3; Müller, 1952; Lyall, 1957; Johns, 1977, 1982; Meyer, 1978; Johns and Fernald, 1981). In larval goldfish, these rod precursors travel from the INL to the ONL, where they continue to proliferate and give rise to rod photoreceptors (Johns, 1982; Raymond and Rivlin, 1987). Like in other species with active CMZ, damaged peripheral retina in larvae can be replaced by the CGZ precursor cells (Fig. 3, #2; reviewed by Reh and Levine, 1998; Raymond and Hitchcock, 2000). Regeneration of the retina in the embryonic fish, on the other hand, has been reported to occur by means of transdifferentiation of RPE (Fig. 3, #1; Dabagyan, 1959, 1960; Sologub, 1975a). This ability is not observed in adult fish, where neural stems cells seem to play a special role (see below).

Adult stages.

Adult fish have been described to regenerate their retinas by means of different cell sources, including neural precursors from the CGZ, rod precursors present in the ONL, quiescent stem cells distributed throughout the retina, and possibly Müller glia cells (reviewed in Hitchcock and Raymond, 1992; Reh and Levine, 1998; Raymond and Hitchcock, 1997, 2000; Reh and Fischer, 2001; also see Wu et al., 2001). On the other hand, transdifferentiation of RPE into neural retina has not been observed (Knight and Raymond, 1995).

Dissecting the Source: Precursors and Stem cells

In teleost fish, regeneration of retinal cell types depends on the damage elicited. If rod photoreceptors are selectively destroyed in the goldfish retina, rod precursors intrinsic to the ONL will replace them by means of enhanced mitotic activity (Braisted et al., 1994). If the retinal damage includes cells from both the INL and the ONL or just the ONL (including rods and cones), then clusters of proliferating neuroepithelial cells will form around the wound and will eventually replace all lost retinal cell types. These proliferating cells are thought to originate from several sources, including rod precursors (Fig. 3, #3; Raymond et al., 1988; Braisted and Raymond, 1992; Hichtcock et al., 1992), intrinsic stem cells in the INL (Fig. 3, #4; Raymond and Hitchcock, 1997; Julian et al., 1998; Otteson et al., 2001; Reh and Fischer, 2001; Wu et al., 2001), and Müller glia cells (Fig. 3, #5) that proliferate and migrate to the ONL (Braisted et al., 1994; Raymond and Hitchcock, 1997, 2001; Reh and Fischer, 2001; Wu et al., 2001). These proliferating neuroepithelial cells express genes known to be involved in early neurogenesis such as pax-6, Notch, and N-Cadherin (Hitchcock et al., 1996; Sullivan et al., 1997; Wu et al., 2001). Eventually these cells give rise to the regenerated missing neurons.

On the other hand, if only cells of the INL or ganglion cells are damaged, no regeneration is observed (Negishi et al., 1988; Hitchcock, 1989). Therefore, it seems that damage to the ONL is necessary to elicit a regenerative response (Hitchcock and Raymond, 1992; Braisted and Raymond, 1996). In addition, it is important to note that, for regeneration to take place, the RPE must be present (Schmidt et al., 1978).

Recently, a model has been suggested that delineates the retinal precursors' lineage (Otteson et al., 2001). Long-term BrdU-systemic labeling of intact goldfish retina supports the presence of slow dividing neural stem cells in the INL, which are spherical in shape and express pax-6 (Fig. 3, #4). These cells can give rise to neural progenitors (fusiform shaped cells and pax-6 negative; Fig. 3, #4) that eventually travel to the ONL and become rod precursors (Fig. 3, #3). A similar population of dividing cells has been detected in zebrafish that followed light-induced photoreceptor degeneration (Vihtelic and Hyde, 2000) and in rapidly growing juvenile and adult rainbow trout retina (Julian et al., 1998). In a more recent study, this population of stem cells was challenged to replace damaged cones and rods. By using a variety of cell markers and BrdU labeling, it was concluded that cones are regenerated from multipotent progenitors originally present in the INL (Fig. 3, #4). These progenitors divide and eventually migrate to the ONL along with proliferating Müller glia nuclei, where they replace the damaged cones. Damaged rods are later replaced by rod precursors present in the ONL (Fig. 3, #3; Wu et al., 2001).

Where to go from here?

Knowledge obtained in the field of retina regeneration in the teleost fish is limited to studies at the cellular and electrophysiological level (reviewed by Hitchcock and Raymond, 1992; Mitashov, 1996; Raymond and Hitchcock, 2000; also see Hitchcock and Cirenza, 1994; Hitchcock, 1997, Cameron et al., 1999). It seems that we are coming closer to understanding the process of retina regeneration in teleost fish as more cDNAs and antibodies are made available, but what really lacks in this field, like in the amphibian retina regeneration field, is the availability of technology to manipulate the system to dissect the molecular mechanism behind this amazing process. With the use of zebrafish as a retina regeneration model (Vihtelic and Hyde, 2000), more avenues will open because transgenic technology, retroviral-mediated insertional mutagenesis, as well as knockout technology has been established (Higashijima et al., 1997; Xie et al., 1997; Amsterdam and Hopkins, 1999; Linney et al., 1999; Halloran et al., 2000; Nasevicius and Ekker 2000; Heasman, 2002).

Retina Regeneration in Chicks

Embryonic stages.

In the chick, during stages 22–24 of development (according to Hamburger and Hamilton, 1951), retinal cells regenerate after retinectomy by means of transdifferentiation upon fibroblast growth factor 2 (FGF-2) treatment (Park and Hollenberg, 1989, 1991, 1993) (Fig. 6). During this process, there is first a vitread shift of pigment granules to the apex of the rPEC, followed by proliferation and dedifferentiation of the rPEC into several layers of neuroblast cells in the first 24 to 48 hr (Fig. 6D). Differentiation of the regenerating retina is observed by the appearance of nerve fibers and ganglion cells by the third day of regeneration. The retina regenerates in a similar sequence to that of normal development, but with reverse polarity (Fig. 6 E,F). As a result, the rods and cones of the photoreceptor layer are located in the innermost layer of the retina, which is closest to the lens. A disappearance of the rPEC is also observed during this regenerative process (Fig. 6D,F; Coulombre and Coulombre, 1965; Park and Hollenberg, 1989, 1991, 1993). Retina regeneration from the anterior region of the retina, most likely from the ciliary marginal zone (Fig. 3, #2), has also been observed in the embryonic chick eye, but only under certain conditions (Fig. 6F; Coulombre and Coulombre, 1965; Park and Hollenberg, 1991). Upon retinectomy, neural retina begins to appear from the anterior margin of the RPE as early as 1 day postretinectomy, as long as a source of FGF-1 or a piece of neural retina is included in the eye cup. This retina differentiates very rapidly with the photoreceptor outer segments facing away from the lens in the normal orientation (Fig. 6F; Coulombre and Coulombre, 1965; Park and Hollenberg, 1991, 1993). This process appears to occur by means of the use of neural precursors (Willbold and Layer, 1992).

thumbnail image

Figure 6. Retina regeneration in embryonic chick eyes. A: Intact chick eye at day 4 of development (stages 22–24 of development according to Hamburger and Hamilton, 1951). Note the retina (r) is not fully formed yet and the retinal pigment epithelium (RPE) does not appear pigmented. B: Retinectomized chick eye at day 4 of development. The entire neural retina has been removed, and the RPE layer has thickened. C: Intact eye at day 7 of development. The neuroepithelial layer has thickened (represented here by r), but no layers have been formed yet. D: At 3 days postretinectomy (day 7 of development), the RPE at the posterior region of the eye has transdifferentiated (t) to form a neuroepithelium similar to the one present in the intact eye at the equivalent stage (C). E: Intact eye at 11 days of development. Note all the retinal layers are nicely formed by now: the outer nuclear layer (o), the inner nuclear layer (i), and the ganglion cell layer (g). F: Regenerating retina 7 days postretinectomy (11 days of development), at which time, new retina has been formed by means of the transdifferentiation of the RPE (t) or by means of the use of neural precursors from the ciliary marginal zone (m). Note that the inner loop of regenerated retina contains all the retinal cell layers in the original orientation, whereas the retina regenerated by means of transdifferentiation of the RPE has a reversed or mirrored orientation. l, lens. Sections were stained with hematoxylin and eosin.

Download figure to PowerPoint

Is retina regeneration possible in adult birds? Neural progenitors vs. Müller glia cells.

Work by Reh and coworkers has eluded that adult birds have/could have the ability to regenerate their retina upon damage. It has been shown that adult chickens posses a CMZ (Fig. 3, #2) that contains progenitor cells as well as neural stem cells that could potentially replace damaged retina. These cells are mitotically active (Morris et al., 1976; Fischer and Reh, 2000) and respond to growth factors by increasing their proliferation (Fischer and Reh, 2000; Fischer et al., 2002). These same cells coexpress cell markers present in neural progenitors such as pax-6 and chx10 (Belecky-Adams et al., 1997; Fischer and Reh, 2000; Reh and Fischer 2001). These progenitor cells eventually give rise to amacrine and bipolar cells but not photoreceptors or ganglion cells. It appears that the repertoire of cells that can be produced by the CMZ is limited to its microenvironment, because exposing postnatal chick eyes to a combination of growth factors, such as FGF-2 and insulin, generates ganglion cells (Fischer et al., 2002). Even though the cells of the CMZ in adult chickens share some of the properties of cells of the CMZ from fish and amphibians, the chick cells are unable to replace damaged retina (Fischer and Reh, 2000; Reh and Fischer, 2001; Fischer et al., 2002).

The pigmented cells present in the ciliary margin or PCM (Fig. 3, #6) are also mitotically active and express cell markers that are transiently expressed in the developing RPE, including pax-6 and Mitf (Fischer and Reh, 2001a). Exposure to certain growth factors induce their proliferation and depigmentation— the initial signs of transdifferentiation. Unfortunately, the postnatal microenvironment does not provide essential factors that could drive these cells to replace damaged retina. As a matter of fact, the population of depigmented cells eventually undergoes cell death (Fischer and Reh, 2001a).

On the other hand, cells from the INL in the central retina, specifically Müller glia cells respond to acute retinal damage (Fig. 3, #5). These cells re-enter the cell cycle, dedifferentiate producing neural precursors that express pax-6, chx10, and cash1 (cell markers for embryonic neural precursors), and eventually differentiate into neurons and glia cells (Fischer and Reh, 2001b; Reh and Fischer, 2001).

Future directions for studies in embryonic and adult birds.

In vitro models have been set up to try to understand the transdifferentiation potential of the RPE and to try to manipulate it (Pittack et al., 1991, 1997; Eguchi and Kodama, 1993; Kodama and Eguchi 1995; Yan and Wang, 1998). With the use of these and other technologies, such as the use of retroviral vectors electoroporation (Morgan and Fekete, 1996; Oruga, 2002), to manipulate the expression of putative genes during retinal regeneration in embryonic chicks in vivo as well as the use of alternative technologies to knockouts, such as morpholinos (Heasman, 2002), the molecular pathway of retina regeneration by means of transdifferenatiation could be dissected. It is also essential to understand the cues and environment that promote Müller glia cells to transdifferentiate to retinal and glial cells so similar approaches should be used. Culture systems for glia cells have been established that could be exploited to understand the factors and molecular pathways that control this process (Degenstein and Moscona, 1986; Okada, 1981; Kodama and Eguchi, 1995; Fischer and Reh, 2001b).

Retina Regeneration in Mammals

Embryonic stages.

Neural retina progenitors are present in mammals during embryogenesis just as in any other organism. In rats, neural retina progenitors have been isolated from embryonic day (E) 17 rats and cultured in vitro. These cells have the ability to differentiate into neurons and glia in vitro (Ahmad et al., 1999), and when transplanted in the subretinal space of 2-week-old rats (Chacko et al., 2000), these cells express photoreceptor-specific markers, suggesting that these progenitor cells have the potential to become photoreceptors and could eventually be used to replace damaged/diseased photoreceptors. On the other hand, embryonic RPE from rats has been shown to transdifferentiate when whole optic vesicles or cups from E11.5 to E14.5 embryos were transplanted into the anterior chamber of adult rat eyes (Stroeva, 1960). Later studies by Zhao et al. (1995) have shown that embryonic RPE (E13–E14) treated with FGF-2 can transdifferentiate in vitro to form stratified layers of cells that express retinal cell-specific markers, confirming the appearance of ganglion, amacrine, and even photoreceptor cell types. RPE from later stages, when cultured in vitro, were unable to fully transdifferentiate (Neill and Barnstable, 1990; Neill et al., 1993; Zhao et al., 1995). Based on these data, it seems that the transdifferentiation potential of the RPE in rodents is limited to a small window during its development.

Postnatal stages.

Until now, birds are the only warm-blooded vertebrates with the capacity to regenerate their neural retina as adult organisms by means of the use of Müller glia cells (Fischer and Reh, 2001b; Reh and Fischer, 2001). Recent reports have opened the possibility that mammals could, under special circumstances, replace damaged retina. As mentioned earlier, adult PCM of mice have the capability to transdifferentiate in vitro into retinal-specific cells (Ahmad et al., 2000; Tropepe et al., 2000). The cells from the PCM have been found to dedifferentiate and form colonies with progenitor cells that self-renew and express cell markers specific to embryonic progenitor cells. These cultured progenitor cells can give rise to a variety of retinal cells, including rod photoreceptors, bipolar neurons, and even Müller glia cells. Populations of PCM cells have also been isolated from postmortem adult bovine and human ciliary margin tissues but have not been further investigated (Tropepe et al., 2000).

In search for the mammalian paradigm.

Because some cells of the adult mammalian retina have the capability to transdifferentiate in vitro, then we could expect that, under the appropriate conditions, those cells could also transdifferentiate in vivo. One way to approach this could be to mimic the in vitro conditions. This could be done with approaches as simple as manipulating the animal's diet to as extreme as overexpressing or knocking out key genes involved in switching on the process of retina regeneration. We look forward to more extensive studies on PCMs.

EYE REGENERATION IN INVERTEBRATES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. LENS REGENERATION
  5. RETINA REGENERATION
  6. EYE REGENERATION IN INVERTEBRATES
  7. PERSPECTIVE
  8. Acknowledgements
  9. REFERENCES

Some invertebrates, such as snails, have eyes remarkably similar to those of vertebrates, i.e., being composed of pigment epithelium, retina, and lens. The eyes are found at the tip of cephalic eyestalks. These animals posses very good eye regenerative abilities if the whole eye is removed by amputation through the mid-eyestalk. The integumentary epithelium of the stalk invaginates at the apex of the eyestalk and produces a shallow cleft or eyecup. The eye is regenerated by transdifferentiation of these epithelial cells. First retina is differentiated followed by the lens. The whole process takes approximately 2 weeks (Miller Bever and Borgens, 1988).

In other invertebrates such as planarians, the eye (eye spot) consists of two cell types: a bipolar nerve cell with photoreceptive properties, and a cup-shaped structure of pigmented cells. Upon head amputation, these cells differentiate from neoblasts, totipotent cells found in the adult eye. Pax-6 is expressed in the cells of the eye spots and, after amputation, appears in a group of pigmented cells, which are the earlier signs of the eye spot regeneration (Callaerts et al., 1999). Obviously, much can be learned from studies with invertebrates, because they use different strategies for regeneration of the eye structures.

PERSPECTIVE

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. LENS REGENERATION
  5. RETINA REGENERATION
  6. EYE REGENERATION IN INVERTEBRATES
  7. PERSPECTIVE
  8. Acknowledgements
  9. REFERENCES

It is apparent from the synthesis that was presented in this review that eye tissues retain a very good potential for regeneration if damaged or removed. The intriguing fact is that regeneration of the lens and the retina takes place by means of different mechanisms. Although transdifferentiation is the only way for the lens to regenerate, retina can be replaced by stem cells as well. Different animal systems use different strategies, depending on the age of the animal. It is important to discuss here the possible connection of the dedifferentiation process seen during transdifferentiation and the generation of stem cells. In other words, does dedifferentiation create stem cells? The answer to this question depends on how we define “stem cells.” If we define stem cells as a population of cells derived from undifferentiated progenitors that have the ability to differentiate into different cell types, then the population of cells generated by means of dedifferentiation from terminally differentiated cells would not fit the definition. On the other hand, if we ignore the origin of the stem cell population and we concentrate on the sole property of stem cells to differentiate into different cell types, then cells generated by means of dedifferentiation could be a suitable source of stem cells. We have suggested before (Tsonis, 2000, 2002) that mesenchymal stem cells derived from the bone marrow destined to become chondrocytes, and blastema cells created by dedifferentiation of the bone or muscle during limb regeneration destined to become chondrocytes as well, should be considered identical at some point. According to this statement, all cells that are involved in repair are stem cells, but their creation could be triggered by different mechanisms. In addition, in many cases, this “pool of stem cells” generated by means of dedifferentiation as well as its original cell source have the ability to replace themselves as they proliferate (just as classic stem cells behave) and eventually differentiate into the different cells that make up the damaged tissue. For example, PECs of the adult newt can give rise, by means of transdifferentiation, to lens cells and to retina cells while the population of PECs is also replaced.

It is also interesting to compare the eye regeneration models with others outside the eye field. A good candidate is the limb regeneration model in the newt. Upon limb amputation, cells left in the stump (such as muscle and bone) dedifferentiate, proliferate, and eventually differentiate into the different cell types necessary to replace the lost part (Tsonis, 1996). Just as in the case of lens and retina regeneration in the newt, transdifferentiation is the mode of limb regeneration. Similar molecules to the ones involved in lens and retina regeneration have also been reported to play a role during limb regeneration such as FGFs and their receptors (D'Jamoos et al., 1998; Yokoyama et al., 2001). Notably, the regenerating limb blastema has the necessary factor(s) to induce lens regeneration when dorsal irises are transplanted away from the eye in a regenerating limb (Ito et al., 1999). When it comes to the eye, studying different systems will complement work that will lead to the ultimate goal, which is to apply the knowledge to treat certain diseases. Lens regeneration could be the ultimate therapy for cataracts. Recent methods that involve the replacement of the cataractous lens with a synthetic lens, often suffer with the development of secondary cataracts due to the proliferation of lens epithelial cells left in the lens capsule (Apple et al., 1992; Ibaraki, 1997). The application of this knowledge in retinal diseases could be even more compelling. Several degenerative diseases such as cone-rod dystrophy (CRD), Stargardt disease, retinopathy of prematurity (ROP), Leber aongenital amaurosis (LCA), juvenile retinoschisis, and Best disease destroy the retina and can lead to permanent blindness. Success in inducing retina regeneration from RPE or by means of neuronal precursors could be of enormous applications in these degenerative diseases. Undoubtedly, delineation of the mechanisms that control induction of lens and retina regeneration will provide solutions to these diseases as we have now entered a new century where regeneration research has taken an important place in the medical world.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. LENS REGENERATION
  5. RETINA REGENERATION
  6. EYE REGENERATION IN INVERTEBRATES
  7. PERSPECTIVE
  8. Acknowledgements
  9. REFERENCES

Special thanks to Elizabeth Hague, Mayur Madhavan, Natalia Vergara, and Mindy Call for their contribution to the artwork. We also thank Mindy Call for critical reading of the manuscript. P.A.T. and K.D.R.T. were funded by the NIH.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. LENS REGENERATION
  5. RETINA REGENERATION
  6. EYE REGENERATION IN INVERTEBRATES
  7. PERSPECTIVE
  8. Acknowledgements
  9. REFERENCES