DEVELOPMENTAL
BIOLOGY
I&3,322-333
Induction
(1991)
of Retinal Regeneration CAROL M. PARK~ANDMARTIN Department
of Anatomy,
University Accepted
of Toronto, August
in Vivo by Growth
Factors
J. HOLLENBERG' Tcwonto,
Ontam’o,
Canada
M5.S lA8
IS, 1991
We have previously reported that basic fibroblast growth factor (bFGF) can induce retinal regeneration in the stage 22-24 chicken embryo. The present study was undertaken to identify the cellular source of the regenerate and to determine whether other growth factors also elicit regeneration in this animal model. Polymer implants containing bFGF were inserted into eyes of chicken embryos immediately after extirpation of the neural retina. The retinal pigment epithelium (RPE) was left intact. Evaluation by light microscopy revealed that in bFGF-treated eyes the new neural retina arose by transdifferentiation of the entire RPE layer. Differentiation of the new neural retina occurred in a sequence similar to that of normal development but proceeded in a reverse (vitread) direction. All retinal laminae had differentiated by Day 15. However, the regenerate displayed reversed polarity, with photoreceptors closest to the lens. The RPE, pecten, and optic nerve were absent. Focal areas of degeneration in the retinal regenerate became evident for the first time on Day 10. Retinal regeneration was also observed after treatment with higher doses of acidic fibroblast growth factor, but not with nerve growth factor-& transforming growth factor-&, insulin, or insulin-like growth factors I or II. These results raise the possibility that FGFs may play a role in retinal differentiation during development. 0 1991 Academic
Press. Inc.
INTRODUCTION
It is apparent from the work of a number of investigators that the neural retina has a remarkable capacity for regeneration. Many vertebrate species, including amphibians (Reyer, 197’7), fish (Dabaghian, 1959), birds (Coulombre and Coulombre, 1965,1970), and mammals (Stroeva, 1960), are capable of regenerating neural retina from the retinal pigment epithelium (RPE) during the early embryonic period. In the 4-day-old chicken embryo (stages 22-24 according to Hamburger and Hamilton, 1951), it has previously been shown that retinal regeneration following surgical removal of the neural retina (retinectomy) occurs if a piece of embryonic chicken neural retina or otocyst (Coulombre and Coulombre, 1965), or embryonic mouse neural retina (Coulombre and Coulombre, 1970), is inserted into the eye. However, the mechanism underlying this process is unknown. There is increasing evidence that polypeptide growth factors may play an important role in development by regulating cell proliferation, cell differentiation, and morphogenesis (Mercola and Stiles, 1988). The presence of certain growth factors and their receptors in the eye has led investigators to speculate that these molecules may be involved in the development, maintenance, and repair of ocular tissues (Barritault et ab, 1981). Both basic and acidic fibroblast growth factors (bFGF and r Present Columbia, 0012.1606/91 Copyright All rights
address: Vancouver,
Department of Anatomy, BC, Canada V6T lW5.
$3.00
0 1991 by Academic Press, Inc. of reproduction in any form reserved.
University
of British
322
aFGF, respectively) have been isolated from neural retina. Studies have shown that bFGF constitutes the major percentage of the mitogenic activity in extracts of adult bovine neural retina (Baird et ah, 1985) and embryonic chicken neural retina (Mascarelli et ab, 1987). More recently, aFGF mRNA has been detected in mouse and bovine neural retina by in situ hybridization (Jacquemin et ah, 1990) and both aFGF and bFGF mRNAs have been found in adult rat retina (Noji et ab, 1990). Other investigations have demonstrated the presence of binding sites for bFGF and aFGF in embryonic mouse retina (Fayein et al, 1990; Jeanny et aL, 1987), and receptors for insulin (Rodrigues et aZ., 1988), insulin-like growth factor I (IGF-I), (Ocrant et al, 1989; Waldbillig et al., 1988; Zick et ah, 1987), and insulin-like growth factor II (IGFII) (Ocrant et al., 1989) have been found in neural retina of rodents and other mammals. Also of relevance to the present study are reports of receptors for insulin (Bassas et ah, 1989) and IGF-I (Bassnett and Beebe, 1990; Bassas et al, 1989) in embryonic chicken neural retina. In a previous investigation, we used ethylene/vinyl acetate copolymer (EVAc) implants to administer bFGF intraocularly following retinectomy in the stage 22-24 chicken embryo (Park and Hollenberg, 1989). Our results showed that bFGF induced retinal regeneration in this animal model in a dose-dependent manner within 7 days. The neural retina that regenerated in bFGFtreated eyes displayed reversed polarity and in all cases the RPE was no longer present. These two observations led us to hypothesize that the retinal regeneration process was initiated by transdifferentiation of the RPE.
PARKANDHOLLENBERG
Retinal
Accordingly, the present investigation was undertaken to determine the cellular source and fate of the regenerate during bFGF-induced retinal regeneration and to determine if aFGF and other growth factors reported to be present in neural and ocular tissue can induce retinal regeneration as well. MATERIALS
AND METHODS
Embryos
Fertilized White Leghorn chicken eggs (Glen Fenelon Farms, Toronto, ON) were incubated in a humidified atmosphere (56%) at 38°C. Embryos were staged according to the criteria of Hamburger and Hamilton (1951).
Regeneration
by Growth
Factors
323
Histological Procedures
Enucleation was carried out in the fixative solution (2.5% glutaraldehyde in 0.1 M phosphate buffer, pH 7.4). Eyes were bisected along the vertical meridian, immersed in fresh fixative for 2-4 hr, dehydrated in ethanol, and embedded in glycol methacrylate (JB-4 Embedding Kit, Polysciences, Inc., Warrington, PA). Serial sections (2 pm) were cut parallel to the vertical meridian and stained with alkaline toluidine blue for light microscopy. RESULTS
Embryos Treated with bFGF
Examination by light microscopy of eyes fixed immediately after insertion of EVAc implants containing bFGF revealed that removal of the neural retina was Preparation of EVAc implants containing bovine complete in all cases. The RPE remaining within the eye serum albumin (BSA, 60 pug; Sigma Chemical Co., St. Louis, MO) and individual growth factors [bFGF (100 appeared undamaged and intact (Fig. la). Within 24 hr implants, the ng, R & D Systems, Inc., Minneapolis, MN), aFGF (1 after insertion of the bFGF-containing entire RPE had given rise to a germinative neuroepitheng-2 pg, R & D Systems, Inc., Minneapolis, MN), insulin (1 ng-1 yg, Sigma Chemical Co.), IGF-I or -11 (1 ng-1 pug, lium several cells thick (Fig. lb). The following series of changes accompanied this process. IniCollaborative Research, Inc., Bedford, MA), nerve morphological tially (0 hr), the RPE in the central region of bFGFgrowth factor-0 (NGF-P, 1 ng-1 pg, Boehringer-Mannheim Canada, Dorval, PQ), or transforming growth fac- treated eyes appeared as a simple, low columnar epithelium (Fig. lc) and was morphologically indistinguishtor-& (TGF-P,, l-500 ng, R & D Systems, Inc., Bedford, able from that of the corresponding controls. Mitotic MA)] or BSA alone (60 pg) was carried out as described previously (Park and Hollenberg, 1989). Implants were figures were rarely observed in the RPE layer at this stage and rod-shaped pigment granules were confined hydrated in Hanks’ balanced salt solution for 10 min to the basal cytoplasm of the cells. Two layers of pigprior to use. mented columnar cells were present at the periphery The effect of neurotrophic molecules, such as brainand mitotic figures were occasionally seen in this rederived neurotrophic factor (BDNF), neurotrophin-3 (NT-3), and ciliary neurotrophic factor (CNTF) (for re- gion. By 8 hr, RPE cells throughout the eye had acquired an elongated shape and there was a marked vitread view, see Thoenen, 1991), was not examined in this study shift of pigment granules to the apical cytoplasm of the as a commercial source was not readily available. cells (Fig. Id). The entire RPE layer subsequently underwent proliferation, forming two to three layers of Surgical Techniques neuroblasts after 16 hr (Fig. le) and three to four layers Complete surgical removal of the neural retina (retinby 24 hr after implant insertion (Fig. If). Throughout ectomy) was performed in ova on the right eyes of stage the proliferative phase, pigment granules were observed 22-24 chicken embryos according to the procedure of almost exclusively in the innermost layer of cells borCoulombre and Coulombre (1965). The RPE was left in- dering the vitreous cavity; mitotic figures were also contact. EVAc implants were inserted into the eyes of em- fined to this region. bryos immediately after retinectomy. Embryos which Subsequent differentiation of the germinative neuroreceived implants containing bFGF plus BSA, or BSA epithelium into neural retina followed the pattern seen alone (controls), were decapitated immediately after (0 during retinal histogenesis. The order of appearance of hr) and at various intervals (2,4,8, and 16 hr; 1,2,3,4,5, the various retinal cell types was identical to that of 6, ‘7, 10, 13, and 15 days) after surgery and eyes were normal development but the retina differentiated with removed for histological processing as described below. reversed polarity. The first layers to form, lying closest Embryos treated with other growth factors and controls to the choroid, were the ganglion and nerve fiber layers. were decapitated 7 days postoperatively and eyes were These were apparent on the third postoperative day processed for histological evaluation. (Fig. 2a). The inner plexiform layer differentiated next, Preparation
of Implants
324
DEVELOPMENTAL
BIOLOGY
VOLUME
148,1991
a
b FIG. 1. Early stages of bFGF-induced retinal regeneration in the chicken embryo. Eyes are shown at different times following insertion of EVAc implants containing 100 ng bFGF. Asterisks indicate EVAc implants. (a) Whole eye fixed immediately (0 hr) after implant insertion. The neural retina has been completely removed, leaving the RPE intact. (b) Whole eye 24 hr after implant insertion. The entire eye is lined with a germinative neuroepithelium several cells thick. (c) 0 hr. The RPE appears as a simple, low columnar epithelium with pigment granules evident in the bases of the cells. (d) 8 hr. Note that the RPE cells have become elongated and pigment granules (arrows) have shifted to the apices of the cells. (e) 16 hr. Proliferation of the RPE has occurred, forming a layer two or three cells in thickness. (f) 24 hr. The regenerate is three or four cells thick. Mitotic figures (arrows) are evident in its innermost cells. L, lens. Scale bar = 250 pm in (a, b); 20 pm in (c-f).
PARK AND HOLLENBERG
Retinal
Regeneration
by Growth
FIG. 2. Differentiation of the regenerating neural retina. Sections of chicken embryo eyes are implants containing 100 ng bFGF. (a) 3 days. At this stage, the ganglion cell (GCL) and nerve (arrows) can be seen in the innermost cells of the regenerate. (b) 5 days. Formation of the inner inner nuclear (INL), outer nuclear (ONL), and outer plexiform (OPL) layers have formed. (d) distinguishable, with the exception of the nerve fiber layer which is absent. The photoreceptor droplets at their distal ends. Scale bar = 25 Frn in (a-c); 27 pm in (d).
making its appearance on the fifth day (Fig. 2b). Formation of the outer plexiform layer and the photoreceptors began on the sixth day. On the seventh postoperative day, the outer plexiform layer could be readily discerned and the developing inner segments of the photoreceptor cells appeared as small spherical buds protruding through the external limiting membrane (Fig. 2~). By this time, the neural retina had achieved a stage of differentiation appropriate to the stage of the embryo in which it was growing. Thus, differentiation of the neural retina during bFGF-induced retinal regener-
Factors
325
shown at various times after insertion of EVAc fiber (NFL) layers have formed. Mitotic figures plexiform layer (IPL) is evident. (c) 7 days. The 15 days. The various retinal laminae are readily cells appear well-differentiated and display oil
ation occurred more rapidly than during normal development. Small areas of degeneration were evident for the first time in the central area of the neural retina on the 10th postoperative day. These foci of degeneration were more numerous and extensive in eyes examined 13 and 15 days postoperatively (Figs. 3a and 3b). In these patches, pyknotic nuclei were abundant in the inner and outer nuclear layers and, with the exception of the ganglion cell and inner plexiform layers, the retinal laminae could no longer be distinguished. However, neural retina flank-
326
DEVELOPMENTAL BIOLOGY
VOLUME 148,199l
FIG. 3. Degeneration of the regenerated neural retina. Section of a chicken embryo eye at 13 days after treatment with 100 ng bFGF. (a) Whole eye. Patches of degenerating neural retina (arrows) can be seen. Note that the RPE, pecten, and optic nerve are absent. (EVAc implant not present in section shown.) (b) Patch of degenerating neural retina shown in (a), at higher magnification. Ganglion cell and inner plexiform layers remain intact. Other retinal layers are disorganized and some cells display pyknotic nuclei. L, Lens; SC, scleral cartilage. Scale bar = 800 pm in (a); 50 pm in (b).
ing these regions was intact and had achieved an advanced stage of differentiation by Day 15 (Fig. 2d). Well-differentiated photoreceptor cells, displaying oil droplets and both inner and outer segments, were evident. The inner and outer nuclear layers and inner and outer plexiform layers were also readily distinguishable but the nerve fiber and ganglion cell layers were attenuated (Fig. 2d). The subsequent fate of the neural retina could not be determined because all surgically manipulated embryos died prior to hatching. There was no evidence of formation of the optic nerve or pecten in any of the bFGF-treated eyes examined. Also, neural retina was not observed regenerating from the extreme periphery of the eye cup. However, at the margins, cells began to be added at 8 hr postoperatively, forming a sparsely pigmented, or often unpigmented, simple, columnar epithelium, which by the seventh day after surgery completely covered the vitreal surface of the lens.
Embryos Treated with aFGF A significant increase in the incidence of retinal regeneration compared to that of controls was found in embryos treated with 100 ng of aFGF (Table 1). Administration of aFGF at a dose of 1 or 2 pg per implant resulted in retinal regeneration in all embryos examined. In these cases, the regenerated neural retina lined almost the entire eye, displayed reversed polarity, and had achieved a stage of differentiation appropriate to the stage of development of the embryo (Figs. 4a and 4b). An additional loop of neural retina, which was continuous with the RPE at the margins of the eye cup and extended into the vitreous cavity, was observed in 3 of 19 embryos treated with 1 pg aFGF and in 2 of 11 embryos treated with 2 pg aFGF (Fig. 4~). The neural retina comprising these loops displayed normal polarity and was apposed to the RPE over part of its length (Figs. 4d and 4e). Its stage of differentiation was not as advanced as that of the neural retina lining the eye cup.
PARKANDHOLLENBERG TABLE
Retinal
1
INCIDENCEOFRETINALREGENERATIONINGROWTH FACTOR-TREATEDANDCONTROLCHICKENEMBRYOS Treatment aFGF Control 1w 10 ng 100 ng 1 P8 2 Pg TGF-/3, Control 1 wz 10 ng 100 ng 500 ng Insulin Control 1 ng 10 ng 100 ng 1 llg IGF-I Control 1w 10 ng 100 ng 1 Pg IGF-II Control 1w 10 ng 100 ng 1 P!z NGF-P Control 1 ng 10 ng 100 ng 1 fig
N
Retinal
regeneration
7 5 8 13 19 11
o/7 o/5 O/8 7/13 19/19 ll/ll
(0) (0) (0) (54)* (loo)** (loo)***
6 3 7 16 13
O/6 o/3 o/7 O/16 o/13
(0) (0) (0) (0) (0)
6 9 10 12 3
O/6 o/9 o/10 o/12 o/3
(0) (0) (0) (0) (0)
11 14 11 16 2
o/11 o/14 o/11 O/16 o/2
(0) (0) (0) (0) (0)
6 10 11 16 2
O/6 o/10 o/11 O/16 o/2
(0) (0) (0) (0) (0)
7 6 14 12 8
o/7 O/6 o/14 o/12 O/8
(0) (0) (0) (0) (0)
Note. Proportions indicate the number of embryos in which retinal regeneration was observed in the total number examined (N); the percentage of eyes showing regeneration is given in parentheses. The significance of differences in the incidence of retinal regeneration in growth factor-treated versus control embryos was tested using a onetailed Fisher’s exact test. * P = 0.02. ** P = 0.00002. *** P = 0.00003.
Embryos Treated with Other Growth Factors
No evidence of retinal regeneration was observed in eyes of embryos treated with TGF-/3,, insulin, IGF-I, IGF-II, or NGF-P at any of the doses tested (Table 1). Further, no morphological differences were noted between these growth factor-treated eyes and those of controls, with the exception of eyes treated with the
Regeneration
by Growth
Factors
327
highest dose of TGF-&. Administration of high doses of TGF-0, had an inhibitory effect on RPE cell proliferation. An intact, single layer of heavily pigmented RPE cells was not observed lining the vitreous cavity of eyes treated with 500 ng TGF-& for 7 days, as was the case in control eyes or those treated with other growth factors or lower doses of TGF-0,. Instead, the tissue surrounding the EVAc implant consisted of closely packed mesenchymal cells; interspersed among these were clusters of pigment-containing cells (Figs. 5a and 5b). Control Embryos
No evidence of neural retina was found in eyes of control embryos up to 15 days after retinectomy and insertion of EVAc implants containing BSA alone. In eyes examined immediately after insertion of the implants (0 hr) (Figs. 6a and 6b), The RPE consisted of a single, intact layer of low columnar epithelial cells, except at the margins of the eye cup, where two layers of these cells were evident. A few rod-shaped pigment granules could be distinguished in the basal cytoplasm of the RPE cells at this stage. By the second postoperative day, RPE cells in the central region of the eye were heavily pigmented and had assumed a cuboidal shape. At the periphery, cells added from the margins of the eye cup as a result of mitosis in this region formed a pigmented, simple, cuboidal epithelium which eventually covered the vitreal surface of the lens. By 15 days postoperatively, a single layer of heavily pigmented, cuboidal cells completely lined the vitreous cavity (Figs. 6c and 6d). No evidence of transdifferentiation of RPE cells was seen in any of the control eyes examined. DISCUSSION
The present study has demonstrated that the new neural retina which regenerates following retinectomy and intraocular administration of bFGF in the stage 22-24 chicken embryo arises by transdifferentiation of the RPE. In addition, our examination of the effects of several other growth factors in this animal model has shown that aFGF also induces retinal regeneration, while TGF-& appears to have an inhibitory effect on RPE cell proliferation. In our investigation, bFGF-induced retinal regeneration occurred exclusively by transdifferentiation of the entire RPE layer. The morphological changes observed in RPE cells following treatment with bFGF were identical to those described previously by Coulombre and Coulombre (1965) in patches of RPE that transdifferentiated into neural retina following retinectomy and insertion of fragments of the excised neural retina. The latter investigators also reported addition of normally oriented neural retina from the margins of the eye cup.
328
DEVELOPMENTALBIOLOGY
V0~~~~148,1991
PARKANDHOLLENBERG
Retinal
Re.qeneratir*n
by Growth Factors
329
FIG. 5. Effect of TGF-P, on chicken embryo eye development after ret .inectomy. Section of an eye from an embryo 7 days after treatment with 500 ng TGF-0,. Asterisk indicates EVAc implant. (a) Whole eye. The EVAc implant is surrounded by mesenchymal tissue. An intact, single layer of RPE cells is absent. (b) Area indicated by arrow in (a) shown :it higher magnification. Densely packed mesenchymal cells and clusters of pigment-containing cells are evident. L, lens. Scale bar = 300 pm in 1 (a); 30 pm in (b).
However, in our study with bFGF, we did not observe this second mode of forming new neural retina. Tissue arising from the cup margins consisted of a single layer of cells which often, but not always, contained pigment granules. One explanation for this discrepancy is that there may have been other growth factors or molecules released from the fragments of neural retina used to induce retinal regeneration in the Coulombres’ experiment. Perhaps one or a combination of these molecules stimulated proliferation and differentiation of those cells at the margin of the eye cup that are responsible for the growth of the neural retina during normal eye development. In bFGF-treated eyes, retinal regeneration was not accompanied by formation of the optic nerve or pecten, and the ganglion and nerve fiber layers of the regenerated neural retina appeared attenuated at later stages. Failure of the optic nerve to develop in bFGF-treated eyes may reflect the fact that the appropriate molecular signals required for axonal guidance are no longer present in sufficient quantity at the time at which the reti-
nal ganglion cells (RGCs) of the regenerate begin to differentiate. Possibly, since the RGCs failed to innervate their target, the optic tectum, they were deprived of target-derived factor(s) required for their survival and this resulted in axon loss and RGC death. During normal development of the chicken embryo, large numbers of RGCs and their axons are produced and subsequently eliminated (Provis and Penfold, 1988). Experiments performed on chicken embryos have shown that early ablation of the tectum greatly enhances this normally occurring cell death and it has been postulated that trophic substances supplied by the tectum are necessary for RGC survival (Hughes and LaVelle, 1975; Hughes and McLoon, 1979). In our study, degeneration of the other laminae of the regenerated neural retina was also evident at later stages of development. This may have been due to the fact that the RPE was no longer present and the pecten failed to form in eyes in which retinal regeneration had occurred. Since the chicken neural retina is avascular, it is dependent on the RPE for the delivery of metabolites from the choriocapillaris. The pecten
FIG. 4. Effect of aFGF on chicken embryo eye development after retinectomy. Sections of eyes from embryos 7 days after treatment with 2 fig aFGF. (EVAc implants not present in sections shown.) (a) Whole eye. The regenerated neural retina lines the eye cup. (b) High power view of the regenerated neural retina shown in (a). Note that the polarity of the regenerate is reversed: developing photoreceptors (arrows) protrude into the vitreous cavity. (c) Whole eye. In this eye, a loop of neural retina which extends into the vitreous cavity is present in addition to the neural retina lining the eye cup. (d) High power view of normally oriented neural retina from loop shown in (c). (e) High power view of anterior region of eye shown in (c). The loop of neural retina, which appears to have arisen at the margins of the eye cup, displays normal polarity and is apposed to the RPE over part of its length. L, lens; SC, scleral cartilage; NR, neural retina; RPE, retinal pigment epithelium; ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GCL, ganglion cell layer; NFL, nerve fiber layer. Scale bar = 500 wrn in (a) and (c); 20 pm in (b); 25 pm in (d); 120 pm in (e).
330
DEVELOPMENTALBIOLOGY V0~~~~~148,1991
FIG. 6. Sections of eyes from control embryos treated with BSA (60 rg) alone. Asterisks indicate EVAc implants. (a) Whole eye immediately after (0 hr) implant insertion. No neural retina is present within the eye. (b) Area indicated by arrow in (a) shown at higher magnification. The RPE remaining within the eye has the appearance of a sparsely pigmented, simple, low columnar epithelium. (c) Whole eye 15 days after implant insertion. No evidence of retinal regeneration can be seen. (d) Area indicated by arrow in (c) shown at higher magnification. The RPE lining the vitreous cavity appears as a heavily pigmented, cuboidal epithelium. L, lens; SC, scleral cartilage. Scale bar = 250 pm in (a); 20 pm in (b) and (d); 600 pm in (c).
is also thought to have a nutritive function (Romanoff, 1960). In the absence of these structures, the regenerated neural retina may have been deprived of a normal supply of nutrients and thus underwent degeneration. Of the other growth factors we examined, aFGF also induced retinal regeneration in a close-dependent manner. However, a lo-fold greater quantity of aFGF was required to elicit the same response observed with bFGF. This observation is in keeping with reports that aFGF is lo- to lOO-fold less potent than bFGF (Walicke and Baird, 1989). While we have not yet examined the early events of aFGF-induced retinal regeneration, the fact that the RPE was absent in aFGF-treated eyes suggests that this process also involved transdifferentia-
tion of the RPE. In a few eyes treated with 1 or 2 pg aFGF, regeneration of a normally oriented neural retina occurred from the cup margins, in addition to regeneration of neural retina of reversed polarity. A study of the sequence of morphological changes that occur following retinectomy and administration of higher doses of aFGF in this animal model will be required to determine if this effect of aFGF is dose dependent and to confirm the cellular source of the two types of regenerates. Perhaps in the previous experiments of Coulombre and Coulombre (1965), bFGF and/or aFGF released by the implanted neural retina fragment induced transdifferentiation of patches of the RPE into neural retina with reversed polarity while, in addition, aFGF elicited
PARKANDHOLLENBERG
Retinal Regeneration by Grmth
regeneration of neural retina with normal polarity from the margins. It would be of interest to determine whether the normally oriented neural retina apposed to the RPE which was seen in some aFGF-treated eyes is capable of long-term survival. In contrast to our findings with FGFs, none of the other growth factors examined in the present study elicited retinal regeneration in the stage 22-24 chicken embryo. With the exception of TGF-&, treatment with these growth factors did not result in any observable morphological changes in the RPE or other ocular tissues. High doses of TGF-&, however, appeared to have an inhibitory effect on the growth of RPE cells after retinectomy while stimulating local accumulation of mesenchymal cells. These findings are in accordance with reports that TGF-P is a potent inhibitor of the proliferation of many cell types in vitro, particularly of epithelial cells, and is also known to enhance proliferation of cells of mesenchymal origin @porn et ah, 1987; Lyons and Moses, 1990). Further, this growth factor has been shown to have antimitotic effects on cultured human fetal retinal cells (Kimchi et al, 1988). Another interpretation is that TGF-/3, induced transdifferentiation of RPE cells into mesenchymal cells. Further studies on the early morphological changes that occur following administration of TGF-& are required to ascertain the fate of the RPE and the origin of the mesenchymal cells. Little is known about the regulation of retinal differentiation during the early stages of eye development in vertebrates. However, on the basis of the findings presented here, one may speculate that FGFs may play a role in this process. It is possible that during normal ocular development, those retinal progenitors that are able to respond to FGFs do so by differentiating into neural retina while those that are unable to respond form RPE. Which pathway of differentiation the precursor cells follow may depend on whether sufficient concentrations of these growth factors are present in the local microenvironment and, assuming that the action of FGFs is direct, on whether the precursor cells express FGF receptors. The actual source of FGFs and their concentration levels within the eye during development is unknown. It is possible that these molecules could be produced by the presumptive neural retina and/or RPE as their presence in these tissues has been demonstrated by several investigators (Mascarelli et al., 1987; Noji et al., 1990; Schweigerer et ah, 1987). Studies have also identified FGF binding sites in the embryonic rodent retina (Fayein et al., 1990; Jeanny et al., 1987) but their pattern of distribution during retinal development in the chicken embryo has not been determined. Although the outer layer of the optic cup normally differentiates as RPE, its cells have the capacity to form neural retina under certain conditions. There is evi-
Factors
331
dence that the ability of these cells to respond to differentiation signals, such as growth factors, may be influenced by the cells’ physical configuration. Early studies on the chicken embryo have indicated that during formation of the optic cup, close apposition of the inner layer (presumptive neural retina) to the outer layer (presumptive RPE) represses the tendency of the latter to form neural retina. If contact between these two layers is prevented by insertion of a thread at the optic vesicle stage, the outer layer differentiates as neural retina (Orts-Llorca and Genis-Galvez, 1960). Even after the cells of the outer layer of the optic cup overtly display the morphological characteristics of RPE cells, they are capable of transdifferentiating into neural retina if the developing neural retina is removed and a fragment of the excised tissue is reinserted into the eye (Coulombre and Coulombre, 1965,197O) or if FGFs are administered intraocularly, as was shown in the present investigation. Thus, separation of the two layers appears to be necessary for the initiation of RPE transdifferentiation although this action is in itself an insufficient stimulus for this process. This was shown in the studies of Coulombre and Coulombre (1965) as well as in our previous study (Park and Hollenberg, 1989): retinal regeneration in the stage 22-23 chicken embryo is not observed after complete retinectomy. Other experiments have shown that neural retina develops where the cells of the eye primordium form aggregates and RPE forms where they are spread into a single layer and are surrounded by mesenchyme (for review, see Lopashov and Stroeva, 1964). For example, Lopashov (1945,196O) observed that optic vesicles from amphibian embryos explanted into saline contracted into a dense aggregate of cells which differentiated only as neural retina. However, when the anterior neural plate was wrapped in ectoderm and surrounded by mesenchyme and explanted into saline, it gave rise to vesicles composed of a single layer of RPE. Dorris (1938) individually cultured pieces of the outer and the inner wall of the optic cup from 70- to 80-hr-old chicken embryos on the surface of plasma clots and observed that both types of cells spread as a thin layer and difierentiated into RPE. However, when explants of the outer wall of the optic cup were cultured in saline, the cells formed an aggregate which sometimes gave rise to neural retina. More recent studies have shown that RPE explants, but not dissociated RPE cells, from early chicken embryos undergo transdifferentiation into retinal progenitors when cultured in the presence of bFGF (T. Reh, personal communication). This would indicate that the ability of RPE cells to transdifferentiate in response to bFGF is governed by the physical configuration of the cells. An interesting observation in this regard is that the binding of FGF to several different non-
332
DEVELOPMENTAL BIOLOGY
transformed cell lines in culture decreases as the cell density increases and that this reduction in FGF binding is due to a reduction in FGF receptor number (Veomett et aZ., 1989). Perhaps contact between the neural retina and the RPE following invagination of the optic vesicle in some way alters the physical configuration of the RPE cells such that FGF receptor expression is inhibited. It is possible that if this contact between the RPE and the neural retina is prevented, for example, by introduction of a thread at the optic vesicle stage, as in the experiment of Orts-Llorca and Genis-Galvez (1960), or is disrupted by retinectomy, as in the present study, RPE cells express receptors for FGFs and respond to these growth factors by differentiating as neural retina. In summary, the present investigation has demonstrated that the state of differentiation of RPE cells can be altered in vivo by certain growth factors. Specifically, this work shows that at stages 22-24 of development embryonic chicken RPE cells will transdifferentiate into neural retina in the presence of certain concentrations of bFGF. It is likely that aFGF-induced retinal regeneration occurs from the RPE by a similar process and that, in addition, aFGF may stimulate cells at the margins of the optic cup to produce neural retina. As with bFGF, these responses to aFGF appear to be highly concentration dependent. The results presented here raise the distinct possibility that variations in the local production of FGFs and their receptors in the eye during development may, in part, regulate the pathway of differentiation of RPE and neural retina precursors. The expert technical assistance of Aina Tilups is gratefully acknowledged. This research was supported by a grant to M.J.H. from the RP Eye Research Foundation of Canada.
VOLUME 1481991
DABAGHIAN, N. V. (1959). Regulative properties of the eye in the embryos of Ascipenseridae. Dokl. Akad Nauk SSSR 125,938-940. DORRIS, F. (1938). Differentiation of the chick eye in vitro. J. Exp. Zool. 78,385-415. FAYEIN, N. A., COURTOIS, Y., and JEANNY, J. C. (1990). Ontogeny of basic fibroblast growth factor binding sites in mouse ocular tissues. Exp.
Cell Res. 188, 75-88.
HAMBURGER, V., and HAMILTON, H. L. (1951). A series of normal stages in the development of the chick embryo. J. Morphol. S&49-92. HUGHES, W. F., and LAVELLE, A. (1975). The effects of early tectal lesions on development in the retinal ganglion cell layer of chick embryos. J. Comp. Neural. 163,265-284. HUGHES, W. F., and MCLOON, S. C. (1979). Ganglion cell death during normal retinal development in the chick: Comparisons with cell death induced by early target field destruction. Exp. Neural. 66, 587-601. JACQUEMIN, E., HALLEY, C. ALTEIRO, J., LAURENT, M., COURTOIS,Y., and JEANNY, J. C. (1990). Localization of acidic fibroblast growth factor (aFGF) mRNA in mouse and bovine retina by in situ hybridization. Neurosci. Lett. 116, 23-28. JEANNY, J.-C., FAYEIN, N., MOENNER, M., CHEVALLIER, B., BARRITAULT, D., and COURTOIS,Y. (1987). Specific fixation of bovine brain and retinal acidic and basic fibroblast growth factors to mouse embryonic eye basement membranes. Exp. Eye Res. 171,63-75. KIMCHI, A., WANG, X.-F., WEIBERG, R. A., CHEIFETZ, S., and MASSAGUI?, J. (1988). Absence of TGF-/3 receptor and growth inhibitory responses in retinoblastoma cells. Science 240,196-199. LOPASHOV, G. V. (1945). 0 nekotorykh prosteishikh organizatsii zaehatkov glaz i mozga u zarodyshei khvostatykh amfibii (Some simple processes of organization of eye and rudiments in urodele embryos). Dokl. Akad. Nauk SSSR 48,634-637. LOPASHOV, G. V. (1949). 0 roli razlichnykh protessov v vosstanovlenii glaz u amfibii (The role of various processes in the restoration of the eyes in amphibians). Dokl. Akad. Nuuk SSSR 69E, 865-868. LOPASHOV, G. V. (1960). Mekhaizmy razvitiya zachatkov glaz v embriogeneze pozvonchnykh (The mechanisms of development of eye rudiments in vertebrate embryogenesis). USSR Academy of Sciences Press, Moscow. LOPASHOV, G. V., and SOLOGUB, A. A. (1972). Artificial metaplasia of pigmented epithelium into retina in tadpoles and adult frogs. J. Embryo1
REFERENCES BAIRD, A., ESCH, F., GOSPODAROWICZ,D., and GUILLEMIN, R. (1985). Retina- and eye-derived endothelial cell growth factors: Partial molecular characterization and identity with acidic and basic fibroblast growth factors. Biochemistry 24,7855-7860. BARRITAULT, D., ARRUTI, C., and COURTOIS,Y. (1981). Is there a ubiquitous growth factor in the eye? Proliferation induced in different cell types by eye-derived growth factors. Diflerentiation 18.29-42. BASSAS, L., GIRBAU, M., LESNIAK, M. A., ROTH, J., and DE PABLO, F. (1989). Development of receptors for insulin and insulin-like growth factor-I in head and brain of chick embryos: Autoradiographic localization. Endocrinology 125,2320-2327. BASSNETT, S., and BEEBE, D. C. (1990). Localization of insulin-like growth factor-I binding sites in the embryonic chick eye. Invest. OphthalmoL
Visual
Sci 31,1637-1643.
COULOMBRE, J. L., and COULOMBRE, A. J. (1965). Regeneration of neural retina from the pigmented epithelium in the chick embryo. Dev. Biol.
12, 79-92.
COULOMBRE, J. L., and COULOMBRE, A. J. (1970). Influence of mouse neural retina on regeneration of chick neural retina from chick embryonic pigmented epithelium. Nature 228,559-560.
Exp.
Morphol.
28,521-546.
LOPASHOV, G. V., and STROEVA, 0. G. (1964). “Development of the Eye: Experimental Studies.” Israel Program for Scientific Translations, Jerusalem. LYONS, R. M., and MOSES, H. L. (1990). Transforming growth factors and the regulation of cell proliferation. Eur. J. Biochem. 18’7, 467473. MASCARELLI, F., RAULAIS, D., COUNIS, M. F., and COURTOIS,Y. (1987). Characterization of acidic and basic fibroblast growth factors in brain, retina and vitreous of the chick embryo. B&hem. Biophys. Res. Commun.
146,478-486.
MERCOLA, M., and STILES, C. D. (1988). Growth factor superfamilies and mammalian embryogenesis. Develoment 102,451-460. NOJI, S., MATSUO, T., KOYAMA, E., YAMAAI, T., NOHNO, T., MATSUO, N., and TANIGUCHI, S. (1990). Expression pattern of acidic and basic fibroblast growth factor genes in adult rat eyes. B&hem. Biophys. Res. Commun.
168,343-349.
OCRANT, I., VALENTINO, K. L., KING, M. G., WIMPY, T. H., ROSENFELD, R. G., and BASKIN, D. G. (1989). Localization and structural characterization of insulin-like growth factor receptors in mammalian retina. Endocrinology 125,2407-2413. ORTS-LLORCA, F., and GJ?.NIS-GALVEZ,J. M. (1960). Experimental production of retinal septa in the chick embryo: Differentiation of pigment epithelium into neural retina. Acta. Anat. 42,31-70.
PARK AND HOLLENBERG
Retinal
PARK, C. M. and HOLLENBERG, M. J. (1989). Basic fibroblast growth factor induces retinal regeneration in viva. Dev. Biol. 134,201-205. PROVIS, J. M., and PENFOLD, P. L. (1988). Cell death and the elimination of retinal axons during development. Prog. Neurobiol. 31,331347. REYER, R. W. (1977). The amphibian eye: Development and regeneration. In “Handbook of Sensory Physiology” (Crescitelli, F., Ed.), Vol. VII/5, pp. 309-390. Springer-Verlag, Berlin. RODRIGUES, M., WALDBILLIG, R. J., RAJAGOPALAN, S., HACKETT, J., LEROITH, D., and CHADER, G. J. (1988). Retinal insulin receptors: Localization using a polyclonal anti-insulin receptor antibody. Brain
Res. 443,389-394.
ROMANOFF, A. L. (1960). The Avian Embryo. Macmillan Co., New York. SCHWEIGERER, L., MALERSTEIN, B., NEUFELD, G., and GOSPODAROWICZ, D. (1987). Basic fibroblast growth factor is synthesized in cultured retinal pigment epithelial cells. Biochem. Biophys. Res. Cmnmun,. 143, 934-940. SPORN, M. B., ROBERTS, A. B., WAKEFIELD, L. M., and DE CROM-
Regeneration
by Growth
Factors
333
BRUGGHE, B. (1987). Some recent advances in the chemistry and biology of transforming growth factor-beta. J. Cell Biol. 105,10391045. STROEVA, 0. G. (1960). Experimental analysis of the eye morphogenesis in mammals. J. Evnbryol. Exp. Morphol. 8,349-368. THOENEN, H. (1991). The changing scene of neurotrophic factors. Trends Neurosci. 14,165-170. VEOME~, G., KUSCZYNSKI, C., KAZAKOFF, P., and RIZZINO, A. (1989). Cell density regulates the number of cell surface receptors for fibroblast growth factor. B&hem. Biophys. Res. Cwmmun. 159,694-700. WALDBILLIG, R. J., FLETCHER, R. T., SOMERS, R. L. and CHADER, G. J. (1988). IGF-I receptors in the bovine neural retina: Structure, kinase activity and comparison with retina insulin receptors. Exp. Eye
Res. 47,587-607.
WALICKE, P. A., and BAIRD, A. (1989). Trophic effects of fibroblast growth factor on neural tissue. Prog. Brain Res. 78,333-338. ZICK, Y., SPIEGEL, A. M., and SAGI-EISENBERG, R. (1987). Insulin-like growth factor I receptors in retinal rod outer segments. J. Biol. Chem.
262,10259-10264.
BIOLOGY
I&3,322-333
Induction
(1991)
of Retinal Regeneration CAROL M. PARK~ANDMARTIN Department
of Anatomy,
University Accepted
of Toronto, August
in Vivo by Growth
Factors
J. HOLLENBERG' Tcwonto,
Ontam’o,
Canada
M5.S lA8
IS, 1991
We have previously reported that basic fibroblast growth factor (bFGF) can induce retinal regeneration in the stage 22-24 chicken embryo. The present study was undertaken to identify the cellular source of the regenerate and to determine whether other growth factors also elicit regeneration in this animal model. Polymer implants containing bFGF were inserted into eyes of chicken embryos immediately after extirpation of the neural retina. The retinal pigment epithelium (RPE) was left intact. Evaluation by light microscopy revealed that in bFGF-treated eyes the new neural retina arose by transdifferentiation of the entire RPE layer. Differentiation of the new neural retina occurred in a sequence similar to that of normal development but proceeded in a reverse (vitread) direction. All retinal laminae had differentiated by Day 15. However, the regenerate displayed reversed polarity, with photoreceptors closest to the lens. The RPE, pecten, and optic nerve were absent. Focal areas of degeneration in the retinal regenerate became evident for the first time on Day 10. Retinal regeneration was also observed after treatment with higher doses of acidic fibroblast growth factor, but not with nerve growth factor-& transforming growth factor-&, insulin, or insulin-like growth factors I or II. These results raise the possibility that FGFs may play a role in retinal differentiation during development. 0 1991 Academic
Press. Inc.
INTRODUCTION
It is apparent from the work of a number of investigators that the neural retina has a remarkable capacity for regeneration. Many vertebrate species, including amphibians (Reyer, 197’7), fish (Dabaghian, 1959), birds (Coulombre and Coulombre, 1965,1970), and mammals (Stroeva, 1960), are capable of regenerating neural retina from the retinal pigment epithelium (RPE) during the early embryonic period. In the 4-day-old chicken embryo (stages 22-24 according to Hamburger and Hamilton, 1951), it has previously been shown that retinal regeneration following surgical removal of the neural retina (retinectomy) occurs if a piece of embryonic chicken neural retina or otocyst (Coulombre and Coulombre, 1965), or embryonic mouse neural retina (Coulombre and Coulombre, 1970), is inserted into the eye. However, the mechanism underlying this process is unknown. There is increasing evidence that polypeptide growth factors may play an important role in development by regulating cell proliferation, cell differentiation, and morphogenesis (Mercola and Stiles, 1988). The presence of certain growth factors and their receptors in the eye has led investigators to speculate that these molecules may be involved in the development, maintenance, and repair of ocular tissues (Barritault et ab, 1981). Both basic and acidic fibroblast growth factors (bFGF and r Present Columbia, 0012.1606/91 Copyright All rights
address: Vancouver,
Department of Anatomy, BC, Canada V6T lW5.
$3.00
0 1991 by Academic Press, Inc. of reproduction in any form reserved.
University
of British
322
aFGF, respectively) have been isolated from neural retina. Studies have shown that bFGF constitutes the major percentage of the mitogenic activity in extracts of adult bovine neural retina (Baird et ah, 1985) and embryonic chicken neural retina (Mascarelli et ab, 1987). More recently, aFGF mRNA has been detected in mouse and bovine neural retina by in situ hybridization (Jacquemin et ah, 1990) and both aFGF and bFGF mRNAs have been found in adult rat retina (Noji et ab, 1990). Other investigations have demonstrated the presence of binding sites for bFGF and aFGF in embryonic mouse retina (Fayein et al, 1990; Jeanny et aL, 1987), and receptors for insulin (Rodrigues et aZ., 1988), insulin-like growth factor I (IGF-I), (Ocrant et al, 1989; Waldbillig et al., 1988; Zick et ah, 1987), and insulin-like growth factor II (IGFII) (Ocrant et al., 1989) have been found in neural retina of rodents and other mammals. Also of relevance to the present study are reports of receptors for insulin (Bassas et ah, 1989) and IGF-I (Bassnett and Beebe, 1990; Bassas et al, 1989) in embryonic chicken neural retina. In a previous investigation, we used ethylene/vinyl acetate copolymer (EVAc) implants to administer bFGF intraocularly following retinectomy in the stage 22-24 chicken embryo (Park and Hollenberg, 1989). Our results showed that bFGF induced retinal regeneration in this animal model in a dose-dependent manner within 7 days. The neural retina that regenerated in bFGFtreated eyes displayed reversed polarity and in all cases the RPE was no longer present. These two observations led us to hypothesize that the retinal regeneration process was initiated by transdifferentiation of the RPE.
PARKANDHOLLENBERG
Retinal
Accordingly, the present investigation was undertaken to determine the cellular source and fate of the regenerate during bFGF-induced retinal regeneration and to determine if aFGF and other growth factors reported to be present in neural and ocular tissue can induce retinal regeneration as well. MATERIALS
AND METHODS
Embryos
Fertilized White Leghorn chicken eggs (Glen Fenelon Farms, Toronto, ON) were incubated in a humidified atmosphere (56%) at 38°C. Embryos were staged according to the criteria of Hamburger and Hamilton (1951).
Regeneration
by Growth
Factors
323
Histological Procedures
Enucleation was carried out in the fixative solution (2.5% glutaraldehyde in 0.1 M phosphate buffer, pH 7.4). Eyes were bisected along the vertical meridian, immersed in fresh fixative for 2-4 hr, dehydrated in ethanol, and embedded in glycol methacrylate (JB-4 Embedding Kit, Polysciences, Inc., Warrington, PA). Serial sections (2 pm) were cut parallel to the vertical meridian and stained with alkaline toluidine blue for light microscopy. RESULTS
Embryos Treated with bFGF
Examination by light microscopy of eyes fixed immediately after insertion of EVAc implants containing bFGF revealed that removal of the neural retina was Preparation of EVAc implants containing bovine complete in all cases. The RPE remaining within the eye serum albumin (BSA, 60 pug; Sigma Chemical Co., St. Louis, MO) and individual growth factors [bFGF (100 appeared undamaged and intact (Fig. la). Within 24 hr implants, the ng, R & D Systems, Inc., Minneapolis, MN), aFGF (1 after insertion of the bFGF-containing entire RPE had given rise to a germinative neuroepitheng-2 pg, R & D Systems, Inc., Minneapolis, MN), insulin (1 ng-1 yg, Sigma Chemical Co.), IGF-I or -11 (1 ng-1 pug, lium several cells thick (Fig. lb). The following series of changes accompanied this process. IniCollaborative Research, Inc., Bedford, MA), nerve morphological tially (0 hr), the RPE in the central region of bFGFgrowth factor-0 (NGF-P, 1 ng-1 pg, Boehringer-Mannheim Canada, Dorval, PQ), or transforming growth fac- treated eyes appeared as a simple, low columnar epithelium (Fig. lc) and was morphologically indistinguishtor-& (TGF-P,, l-500 ng, R & D Systems, Inc., Bedford, able from that of the corresponding controls. Mitotic MA)] or BSA alone (60 pg) was carried out as described previously (Park and Hollenberg, 1989). Implants were figures were rarely observed in the RPE layer at this stage and rod-shaped pigment granules were confined hydrated in Hanks’ balanced salt solution for 10 min to the basal cytoplasm of the cells. Two layers of pigprior to use. mented columnar cells were present at the periphery The effect of neurotrophic molecules, such as brainand mitotic figures were occasionally seen in this rederived neurotrophic factor (BDNF), neurotrophin-3 (NT-3), and ciliary neurotrophic factor (CNTF) (for re- gion. By 8 hr, RPE cells throughout the eye had acquired an elongated shape and there was a marked vitread view, see Thoenen, 1991), was not examined in this study shift of pigment granules to the apical cytoplasm of the as a commercial source was not readily available. cells (Fig. Id). The entire RPE layer subsequently underwent proliferation, forming two to three layers of Surgical Techniques neuroblasts after 16 hr (Fig. le) and three to four layers Complete surgical removal of the neural retina (retinby 24 hr after implant insertion (Fig. If). Throughout ectomy) was performed in ova on the right eyes of stage the proliferative phase, pigment granules were observed 22-24 chicken embryos according to the procedure of almost exclusively in the innermost layer of cells borCoulombre and Coulombre (1965). The RPE was left in- dering the vitreous cavity; mitotic figures were also contact. EVAc implants were inserted into the eyes of em- fined to this region. bryos immediately after retinectomy. Embryos which Subsequent differentiation of the germinative neuroreceived implants containing bFGF plus BSA, or BSA epithelium into neural retina followed the pattern seen alone (controls), were decapitated immediately after (0 during retinal histogenesis. The order of appearance of hr) and at various intervals (2,4,8, and 16 hr; 1,2,3,4,5, the various retinal cell types was identical to that of 6, ‘7, 10, 13, and 15 days) after surgery and eyes were normal development but the retina differentiated with removed for histological processing as described below. reversed polarity. The first layers to form, lying closest Embryos treated with other growth factors and controls to the choroid, were the ganglion and nerve fiber layers. were decapitated 7 days postoperatively and eyes were These were apparent on the third postoperative day processed for histological evaluation. (Fig. 2a). The inner plexiform layer differentiated next, Preparation
of Implants
324
DEVELOPMENTAL
BIOLOGY
VOLUME
148,1991
a
b FIG. 1. Early stages of bFGF-induced retinal regeneration in the chicken embryo. Eyes are shown at different times following insertion of EVAc implants containing 100 ng bFGF. Asterisks indicate EVAc implants. (a) Whole eye fixed immediately (0 hr) after implant insertion. The neural retina has been completely removed, leaving the RPE intact. (b) Whole eye 24 hr after implant insertion. The entire eye is lined with a germinative neuroepithelium several cells thick. (c) 0 hr. The RPE appears as a simple, low columnar epithelium with pigment granules evident in the bases of the cells. (d) 8 hr. Note that the RPE cells have become elongated and pigment granules (arrows) have shifted to the apices of the cells. (e) 16 hr. Proliferation of the RPE has occurred, forming a layer two or three cells in thickness. (f) 24 hr. The regenerate is three or four cells thick. Mitotic figures (arrows) are evident in its innermost cells. L, lens. Scale bar = 250 pm in (a, b); 20 pm in (c-f).
PARK AND HOLLENBERG
Retinal
Regeneration
by Growth
FIG. 2. Differentiation of the regenerating neural retina. Sections of chicken embryo eyes are implants containing 100 ng bFGF. (a) 3 days. At this stage, the ganglion cell (GCL) and nerve (arrows) can be seen in the innermost cells of the regenerate. (b) 5 days. Formation of the inner inner nuclear (INL), outer nuclear (ONL), and outer plexiform (OPL) layers have formed. (d) distinguishable, with the exception of the nerve fiber layer which is absent. The photoreceptor droplets at their distal ends. Scale bar = 25 Frn in (a-c); 27 pm in (d).
making its appearance on the fifth day (Fig. 2b). Formation of the outer plexiform layer and the photoreceptors began on the sixth day. On the seventh postoperative day, the outer plexiform layer could be readily discerned and the developing inner segments of the photoreceptor cells appeared as small spherical buds protruding through the external limiting membrane (Fig. 2~). By this time, the neural retina had achieved a stage of differentiation appropriate to the stage of the embryo in which it was growing. Thus, differentiation of the neural retina during bFGF-induced retinal regener-
Factors
325
shown at various times after insertion of EVAc fiber (NFL) layers have formed. Mitotic figures plexiform layer (IPL) is evident. (c) 7 days. The 15 days. The various retinal laminae are readily cells appear well-differentiated and display oil
ation occurred more rapidly than during normal development. Small areas of degeneration were evident for the first time in the central area of the neural retina on the 10th postoperative day. These foci of degeneration were more numerous and extensive in eyes examined 13 and 15 days postoperatively (Figs. 3a and 3b). In these patches, pyknotic nuclei were abundant in the inner and outer nuclear layers and, with the exception of the ganglion cell and inner plexiform layers, the retinal laminae could no longer be distinguished. However, neural retina flank-
326
DEVELOPMENTAL BIOLOGY
VOLUME 148,199l
FIG. 3. Degeneration of the regenerated neural retina. Section of a chicken embryo eye at 13 days after treatment with 100 ng bFGF. (a) Whole eye. Patches of degenerating neural retina (arrows) can be seen. Note that the RPE, pecten, and optic nerve are absent. (EVAc implant not present in section shown.) (b) Patch of degenerating neural retina shown in (a), at higher magnification. Ganglion cell and inner plexiform layers remain intact. Other retinal layers are disorganized and some cells display pyknotic nuclei. L, Lens; SC, scleral cartilage. Scale bar = 800 pm in (a); 50 pm in (b).
ing these regions was intact and had achieved an advanced stage of differentiation by Day 15 (Fig. 2d). Well-differentiated photoreceptor cells, displaying oil droplets and both inner and outer segments, were evident. The inner and outer nuclear layers and inner and outer plexiform layers were also readily distinguishable but the nerve fiber and ganglion cell layers were attenuated (Fig. 2d). The subsequent fate of the neural retina could not be determined because all surgically manipulated embryos died prior to hatching. There was no evidence of formation of the optic nerve or pecten in any of the bFGF-treated eyes examined. Also, neural retina was not observed regenerating from the extreme periphery of the eye cup. However, at the margins, cells began to be added at 8 hr postoperatively, forming a sparsely pigmented, or often unpigmented, simple, columnar epithelium, which by the seventh day after surgery completely covered the vitreal surface of the lens.
Embryos Treated with aFGF A significant increase in the incidence of retinal regeneration compared to that of controls was found in embryos treated with 100 ng of aFGF (Table 1). Administration of aFGF at a dose of 1 or 2 pg per implant resulted in retinal regeneration in all embryos examined. In these cases, the regenerated neural retina lined almost the entire eye, displayed reversed polarity, and had achieved a stage of differentiation appropriate to the stage of development of the embryo (Figs. 4a and 4b). An additional loop of neural retina, which was continuous with the RPE at the margins of the eye cup and extended into the vitreous cavity, was observed in 3 of 19 embryos treated with 1 pg aFGF and in 2 of 11 embryos treated with 2 pg aFGF (Fig. 4~). The neural retina comprising these loops displayed normal polarity and was apposed to the RPE over part of its length (Figs. 4d and 4e). Its stage of differentiation was not as advanced as that of the neural retina lining the eye cup.
PARKANDHOLLENBERG TABLE
Retinal
1
INCIDENCEOFRETINALREGENERATIONINGROWTH FACTOR-TREATEDANDCONTROLCHICKENEMBRYOS Treatment aFGF Control 1w 10 ng 100 ng 1 P8 2 Pg TGF-/3, Control 1 wz 10 ng 100 ng 500 ng Insulin Control 1 ng 10 ng 100 ng 1 llg IGF-I Control 1w 10 ng 100 ng 1 Pg IGF-II Control 1w 10 ng 100 ng 1 P!z NGF-P Control 1 ng 10 ng 100 ng 1 fig
N
Retinal
regeneration
7 5 8 13 19 11
o/7 o/5 O/8 7/13 19/19 ll/ll
(0) (0) (0) (54)* (loo)** (loo)***
6 3 7 16 13
O/6 o/3 o/7 O/16 o/13
(0) (0) (0) (0) (0)
6 9 10 12 3
O/6 o/9 o/10 o/12 o/3
(0) (0) (0) (0) (0)
11 14 11 16 2
o/11 o/14 o/11 O/16 o/2
(0) (0) (0) (0) (0)
6 10 11 16 2
O/6 o/10 o/11 O/16 o/2
(0) (0) (0) (0) (0)
7 6 14 12 8
o/7 O/6 o/14 o/12 O/8
(0) (0) (0) (0) (0)
Note. Proportions indicate the number of embryos in which retinal regeneration was observed in the total number examined (N); the percentage of eyes showing regeneration is given in parentheses. The significance of differences in the incidence of retinal regeneration in growth factor-treated versus control embryos was tested using a onetailed Fisher’s exact test. * P = 0.02. ** P = 0.00002. *** P = 0.00003.
Embryos Treated with Other Growth Factors
No evidence of retinal regeneration was observed in eyes of embryos treated with TGF-/3,, insulin, IGF-I, IGF-II, or NGF-P at any of the doses tested (Table 1). Further, no morphological differences were noted between these growth factor-treated eyes and those of controls, with the exception of eyes treated with the
Regeneration
by Growth
Factors
327
highest dose of TGF-&. Administration of high doses of TGF-0, had an inhibitory effect on RPE cell proliferation. An intact, single layer of heavily pigmented RPE cells was not observed lining the vitreous cavity of eyes treated with 500 ng TGF-& for 7 days, as was the case in control eyes or those treated with other growth factors or lower doses of TGF-0,. Instead, the tissue surrounding the EVAc implant consisted of closely packed mesenchymal cells; interspersed among these were clusters of pigment-containing cells (Figs. 5a and 5b). Control Embryos
No evidence of neural retina was found in eyes of control embryos up to 15 days after retinectomy and insertion of EVAc implants containing BSA alone. In eyes examined immediately after insertion of the implants (0 hr) (Figs. 6a and 6b), The RPE consisted of a single, intact layer of low columnar epithelial cells, except at the margins of the eye cup, where two layers of these cells were evident. A few rod-shaped pigment granules could be distinguished in the basal cytoplasm of the RPE cells at this stage. By the second postoperative day, RPE cells in the central region of the eye were heavily pigmented and had assumed a cuboidal shape. At the periphery, cells added from the margins of the eye cup as a result of mitosis in this region formed a pigmented, simple, cuboidal epithelium which eventually covered the vitreal surface of the lens. By 15 days postoperatively, a single layer of heavily pigmented, cuboidal cells completely lined the vitreous cavity (Figs. 6c and 6d). No evidence of transdifferentiation of RPE cells was seen in any of the control eyes examined. DISCUSSION
The present study has demonstrated that the new neural retina which regenerates following retinectomy and intraocular administration of bFGF in the stage 22-24 chicken embryo arises by transdifferentiation of the RPE. In addition, our examination of the effects of several other growth factors in this animal model has shown that aFGF also induces retinal regeneration, while TGF-& appears to have an inhibitory effect on RPE cell proliferation. In our investigation, bFGF-induced retinal regeneration occurred exclusively by transdifferentiation of the entire RPE layer. The morphological changes observed in RPE cells following treatment with bFGF were identical to those described previously by Coulombre and Coulombre (1965) in patches of RPE that transdifferentiated into neural retina following retinectomy and insertion of fragments of the excised neural retina. The latter investigators also reported addition of normally oriented neural retina from the margins of the eye cup.
328
DEVELOPMENTALBIOLOGY
V0~~~~148,1991
PARKANDHOLLENBERG
Retinal
Re.qeneratir*n
by Growth Factors
329
FIG. 5. Effect of TGF-P, on chicken embryo eye development after ret .inectomy. Section of an eye from an embryo 7 days after treatment with 500 ng TGF-0,. Asterisk indicates EVAc implant. (a) Whole eye. The EVAc implant is surrounded by mesenchymal tissue. An intact, single layer of RPE cells is absent. (b) Area indicated by arrow in (a) shown :it higher magnification. Densely packed mesenchymal cells and clusters of pigment-containing cells are evident. L, lens. Scale bar = 300 pm in 1 (a); 30 pm in (b).
However, in our study with bFGF, we did not observe this second mode of forming new neural retina. Tissue arising from the cup margins consisted of a single layer of cells which often, but not always, contained pigment granules. One explanation for this discrepancy is that there may have been other growth factors or molecules released from the fragments of neural retina used to induce retinal regeneration in the Coulombres’ experiment. Perhaps one or a combination of these molecules stimulated proliferation and differentiation of those cells at the margin of the eye cup that are responsible for the growth of the neural retina during normal eye development. In bFGF-treated eyes, retinal regeneration was not accompanied by formation of the optic nerve or pecten, and the ganglion and nerve fiber layers of the regenerated neural retina appeared attenuated at later stages. Failure of the optic nerve to develop in bFGF-treated eyes may reflect the fact that the appropriate molecular signals required for axonal guidance are no longer present in sufficient quantity at the time at which the reti-
nal ganglion cells (RGCs) of the regenerate begin to differentiate. Possibly, since the RGCs failed to innervate their target, the optic tectum, they were deprived of target-derived factor(s) required for their survival and this resulted in axon loss and RGC death. During normal development of the chicken embryo, large numbers of RGCs and their axons are produced and subsequently eliminated (Provis and Penfold, 1988). Experiments performed on chicken embryos have shown that early ablation of the tectum greatly enhances this normally occurring cell death and it has been postulated that trophic substances supplied by the tectum are necessary for RGC survival (Hughes and LaVelle, 1975; Hughes and McLoon, 1979). In our study, degeneration of the other laminae of the regenerated neural retina was also evident at later stages of development. This may have been due to the fact that the RPE was no longer present and the pecten failed to form in eyes in which retinal regeneration had occurred. Since the chicken neural retina is avascular, it is dependent on the RPE for the delivery of metabolites from the choriocapillaris. The pecten
FIG. 4. Effect of aFGF on chicken embryo eye development after retinectomy. Sections of eyes from embryos 7 days after treatment with 2 fig aFGF. (EVAc implants not present in sections shown.) (a) Whole eye. The regenerated neural retina lines the eye cup. (b) High power view of the regenerated neural retina shown in (a). Note that the polarity of the regenerate is reversed: developing photoreceptors (arrows) protrude into the vitreous cavity. (c) Whole eye. In this eye, a loop of neural retina which extends into the vitreous cavity is present in addition to the neural retina lining the eye cup. (d) High power view of normally oriented neural retina from loop shown in (c). (e) High power view of anterior region of eye shown in (c). The loop of neural retina, which appears to have arisen at the margins of the eye cup, displays normal polarity and is apposed to the RPE over part of its length. L, lens; SC, scleral cartilage; NR, neural retina; RPE, retinal pigment epithelium; ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GCL, ganglion cell layer; NFL, nerve fiber layer. Scale bar = 500 wrn in (a) and (c); 20 pm in (b); 25 pm in (d); 120 pm in (e).
330
DEVELOPMENTALBIOLOGY V0~~~~~148,1991
FIG. 6. Sections of eyes from control embryos treated with BSA (60 rg) alone. Asterisks indicate EVAc implants. (a) Whole eye immediately after (0 hr) implant insertion. No neural retina is present within the eye. (b) Area indicated by arrow in (a) shown at higher magnification. The RPE remaining within the eye has the appearance of a sparsely pigmented, simple, low columnar epithelium. (c) Whole eye 15 days after implant insertion. No evidence of retinal regeneration can be seen. (d) Area indicated by arrow in (c) shown at higher magnification. The RPE lining the vitreous cavity appears as a heavily pigmented, cuboidal epithelium. L, lens; SC, scleral cartilage. Scale bar = 250 pm in (a); 20 pm in (b) and (d); 600 pm in (c).
is also thought to have a nutritive function (Romanoff, 1960). In the absence of these structures, the regenerated neural retina may have been deprived of a normal supply of nutrients and thus underwent degeneration. Of the other growth factors we examined, aFGF also induced retinal regeneration in a close-dependent manner. However, a lo-fold greater quantity of aFGF was required to elicit the same response observed with bFGF. This observation is in keeping with reports that aFGF is lo- to lOO-fold less potent than bFGF (Walicke and Baird, 1989). While we have not yet examined the early events of aFGF-induced retinal regeneration, the fact that the RPE was absent in aFGF-treated eyes suggests that this process also involved transdifferentia-
tion of the RPE. In a few eyes treated with 1 or 2 pg aFGF, regeneration of a normally oriented neural retina occurred from the cup margins, in addition to regeneration of neural retina of reversed polarity. A study of the sequence of morphological changes that occur following retinectomy and administration of higher doses of aFGF in this animal model will be required to determine if this effect of aFGF is dose dependent and to confirm the cellular source of the two types of regenerates. Perhaps in the previous experiments of Coulombre and Coulombre (1965), bFGF and/or aFGF released by the implanted neural retina fragment induced transdifferentiation of patches of the RPE into neural retina with reversed polarity while, in addition, aFGF elicited
PARKANDHOLLENBERG
Retinal Regeneration by Grmth
regeneration of neural retina with normal polarity from the margins. It would be of interest to determine whether the normally oriented neural retina apposed to the RPE which was seen in some aFGF-treated eyes is capable of long-term survival. In contrast to our findings with FGFs, none of the other growth factors examined in the present study elicited retinal regeneration in the stage 22-24 chicken embryo. With the exception of TGF-&, treatment with these growth factors did not result in any observable morphological changes in the RPE or other ocular tissues. High doses of TGF-&, however, appeared to have an inhibitory effect on the growth of RPE cells after retinectomy while stimulating local accumulation of mesenchymal cells. These findings are in accordance with reports that TGF-P is a potent inhibitor of the proliferation of many cell types in vitro, particularly of epithelial cells, and is also known to enhance proliferation of cells of mesenchymal origin @porn et ah, 1987; Lyons and Moses, 1990). Further, this growth factor has been shown to have antimitotic effects on cultured human fetal retinal cells (Kimchi et al, 1988). Another interpretation is that TGF-/3, induced transdifferentiation of RPE cells into mesenchymal cells. Further studies on the early morphological changes that occur following administration of TGF-& are required to ascertain the fate of the RPE and the origin of the mesenchymal cells. Little is known about the regulation of retinal differentiation during the early stages of eye development in vertebrates. However, on the basis of the findings presented here, one may speculate that FGFs may play a role in this process. It is possible that during normal ocular development, those retinal progenitors that are able to respond to FGFs do so by differentiating into neural retina while those that are unable to respond form RPE. Which pathway of differentiation the precursor cells follow may depend on whether sufficient concentrations of these growth factors are present in the local microenvironment and, assuming that the action of FGFs is direct, on whether the precursor cells express FGF receptors. The actual source of FGFs and their concentration levels within the eye during development is unknown. It is possible that these molecules could be produced by the presumptive neural retina and/or RPE as their presence in these tissues has been demonstrated by several investigators (Mascarelli et al., 1987; Noji et al., 1990; Schweigerer et ah, 1987). Studies have also identified FGF binding sites in the embryonic rodent retina (Fayein et al., 1990; Jeanny et al., 1987) but their pattern of distribution during retinal development in the chicken embryo has not been determined. Although the outer layer of the optic cup normally differentiates as RPE, its cells have the capacity to form neural retina under certain conditions. There is evi-
Factors
331
dence that the ability of these cells to respond to differentiation signals, such as growth factors, may be influenced by the cells’ physical configuration. Early studies on the chicken embryo have indicated that during formation of the optic cup, close apposition of the inner layer (presumptive neural retina) to the outer layer (presumptive RPE) represses the tendency of the latter to form neural retina. If contact between these two layers is prevented by insertion of a thread at the optic vesicle stage, the outer layer differentiates as neural retina (Orts-Llorca and Genis-Galvez, 1960). Even after the cells of the outer layer of the optic cup overtly display the morphological characteristics of RPE cells, they are capable of transdifferentiating into neural retina if the developing neural retina is removed and a fragment of the excised tissue is reinserted into the eye (Coulombre and Coulombre, 1965,197O) or if FGFs are administered intraocularly, as was shown in the present investigation. Thus, separation of the two layers appears to be necessary for the initiation of RPE transdifferentiation although this action is in itself an insufficient stimulus for this process. This was shown in the studies of Coulombre and Coulombre (1965) as well as in our previous study (Park and Hollenberg, 1989): retinal regeneration in the stage 22-23 chicken embryo is not observed after complete retinectomy. Other experiments have shown that neural retina develops where the cells of the eye primordium form aggregates and RPE forms where they are spread into a single layer and are surrounded by mesenchyme (for review, see Lopashov and Stroeva, 1964). For example, Lopashov (1945,196O) observed that optic vesicles from amphibian embryos explanted into saline contracted into a dense aggregate of cells which differentiated only as neural retina. However, when the anterior neural plate was wrapped in ectoderm and surrounded by mesenchyme and explanted into saline, it gave rise to vesicles composed of a single layer of RPE. Dorris (1938) individually cultured pieces of the outer and the inner wall of the optic cup from 70- to 80-hr-old chicken embryos on the surface of plasma clots and observed that both types of cells spread as a thin layer and difierentiated into RPE. However, when explants of the outer wall of the optic cup were cultured in saline, the cells formed an aggregate which sometimes gave rise to neural retina. More recent studies have shown that RPE explants, but not dissociated RPE cells, from early chicken embryos undergo transdifferentiation into retinal progenitors when cultured in the presence of bFGF (T. Reh, personal communication). This would indicate that the ability of RPE cells to transdifferentiate in response to bFGF is governed by the physical configuration of the cells. An interesting observation in this regard is that the binding of FGF to several different non-
332
DEVELOPMENTAL BIOLOGY
transformed cell lines in culture decreases as the cell density increases and that this reduction in FGF binding is due to a reduction in FGF receptor number (Veomett et aZ., 1989). Perhaps contact between the neural retina and the RPE following invagination of the optic vesicle in some way alters the physical configuration of the RPE cells such that FGF receptor expression is inhibited. It is possible that if this contact between the RPE and the neural retina is prevented, for example, by introduction of a thread at the optic vesicle stage, as in the experiment of Orts-Llorca and Genis-Galvez (1960), or is disrupted by retinectomy, as in the present study, RPE cells express receptors for FGFs and respond to these growth factors by differentiating as neural retina. In summary, the present investigation has demonstrated that the state of differentiation of RPE cells can be altered in vivo by certain growth factors. Specifically, this work shows that at stages 22-24 of development embryonic chicken RPE cells will transdifferentiate into neural retina in the presence of certain concentrations of bFGF. It is likely that aFGF-induced retinal regeneration occurs from the RPE by a similar process and that, in addition, aFGF may stimulate cells at the margins of the optic cup to produce neural retina. As with bFGF, these responses to aFGF appear to be highly concentration dependent. The results presented here raise the distinct possibility that variations in the local production of FGFs and their receptors in the eye during development may, in part, regulate the pathway of differentiation of RPE and neural retina precursors. The expert technical assistance of Aina Tilups is gratefully acknowledged. This research was supported by a grant to M.J.H. from the RP Eye Research Foundation of Canada.
VOLUME 1481991
DABAGHIAN, N. V. (1959). Regulative properties of the eye in the embryos of Ascipenseridae. Dokl. Akad Nauk SSSR 125,938-940. DORRIS, F. (1938). Differentiation of the chick eye in vitro. J. Exp. Zool. 78,385-415. FAYEIN, N. A., COURTOIS, Y., and JEANNY, J. C. (1990). Ontogeny of basic fibroblast growth factor binding sites in mouse ocular tissues. Exp.
Cell Res. 188, 75-88.
HAMBURGER, V., and HAMILTON, H. L. (1951). A series of normal stages in the development of the chick embryo. J. Morphol. S&49-92. HUGHES, W. F., and LAVELLE, A. (1975). The effects of early tectal lesions on development in the retinal ganglion cell layer of chick embryos. J. Comp. Neural. 163,265-284. HUGHES, W. F., and MCLOON, S. C. (1979). Ganglion cell death during normal retinal development in the chick: Comparisons with cell death induced by early target field destruction. Exp. Neural. 66, 587-601. JACQUEMIN, E., HALLEY, C. ALTEIRO, J., LAURENT, M., COURTOIS,Y., and JEANNY, J. C. (1990). Localization of acidic fibroblast growth factor (aFGF) mRNA in mouse and bovine retina by in situ hybridization. Neurosci. Lett. 116, 23-28. JEANNY, J.-C., FAYEIN, N., MOENNER, M., CHEVALLIER, B., BARRITAULT, D., and COURTOIS,Y. (1987). Specific fixation of bovine brain and retinal acidic and basic fibroblast growth factors to mouse embryonic eye basement membranes. Exp. Eye Res. 171,63-75. KIMCHI, A., WANG, X.-F., WEIBERG, R. A., CHEIFETZ, S., and MASSAGUI?, J. (1988). Absence of TGF-/3 receptor and growth inhibitory responses in retinoblastoma cells. Science 240,196-199. LOPASHOV, G. V. (1945). 0 nekotorykh prosteishikh organizatsii zaehatkov glaz i mozga u zarodyshei khvostatykh amfibii (Some simple processes of organization of eye and rudiments in urodele embryos). Dokl. Akad. Nauk SSSR 48,634-637. LOPASHOV, G. V. (1949). 0 roli razlichnykh protessov v vosstanovlenii glaz u amfibii (The role of various processes in the restoration of the eyes in amphibians). Dokl. Akad. Nuuk SSSR 69E, 865-868. LOPASHOV, G. V. (1960). Mekhaizmy razvitiya zachatkov glaz v embriogeneze pozvonchnykh (The mechanisms of development of eye rudiments in vertebrate embryogenesis). USSR Academy of Sciences Press, Moscow. LOPASHOV, G. V., and SOLOGUB, A. A. (1972). Artificial metaplasia of pigmented epithelium into retina in tadpoles and adult frogs. J. Embryo1
REFERENCES BAIRD, A., ESCH, F., GOSPODAROWICZ,D., and GUILLEMIN, R. (1985). Retina- and eye-derived endothelial cell growth factors: Partial molecular characterization and identity with acidic and basic fibroblast growth factors. Biochemistry 24,7855-7860. BARRITAULT, D., ARRUTI, C., and COURTOIS,Y. (1981). Is there a ubiquitous growth factor in the eye? Proliferation induced in different cell types by eye-derived growth factors. Diflerentiation 18.29-42. BASSAS, L., GIRBAU, M., LESNIAK, M. A., ROTH, J., and DE PABLO, F. (1989). Development of receptors for insulin and insulin-like growth factor-I in head and brain of chick embryos: Autoradiographic localization. Endocrinology 125,2320-2327. BASSNETT, S., and BEEBE, D. C. (1990). Localization of insulin-like growth factor-I binding sites in the embryonic chick eye. Invest. OphthalmoL
Visual
Sci 31,1637-1643.
COULOMBRE, J. L., and COULOMBRE, A. J. (1965). Regeneration of neural retina from the pigmented epithelium in the chick embryo. Dev. Biol.
12, 79-92.
COULOMBRE, J. L., and COULOMBRE, A. J. (1970). Influence of mouse neural retina on regeneration of chick neural retina from chick embryonic pigmented epithelium. Nature 228,559-560.
Exp.
Morphol.
28,521-546.
LOPASHOV, G. V., and STROEVA, 0. G. (1964). “Development of the Eye: Experimental Studies.” Israel Program for Scientific Translations, Jerusalem. LYONS, R. M., and MOSES, H. L. (1990). Transforming growth factors and the regulation of cell proliferation. Eur. J. Biochem. 18’7, 467473. MASCARELLI, F., RAULAIS, D., COUNIS, M. F., and COURTOIS,Y. (1987). Characterization of acidic and basic fibroblast growth factors in brain, retina and vitreous of the chick embryo. B&hem. Biophys. Res. Commun.
146,478-486.
MERCOLA, M., and STILES, C. D. (1988). Growth factor superfamilies and mammalian embryogenesis. Develoment 102,451-460. NOJI, S., MATSUO, T., KOYAMA, E., YAMAAI, T., NOHNO, T., MATSUO, N., and TANIGUCHI, S. (1990). Expression pattern of acidic and basic fibroblast growth factor genes in adult rat eyes. B&hem. Biophys. Res. Commun.
168,343-349.
OCRANT, I., VALENTINO, K. L., KING, M. G., WIMPY, T. H., ROSENFELD, R. G., and BASKIN, D. G. (1989). Localization and structural characterization of insulin-like growth factor receptors in mammalian retina. Endocrinology 125,2407-2413. ORTS-LLORCA, F., and GJ?.NIS-GALVEZ,J. M. (1960). Experimental production of retinal septa in the chick embryo: Differentiation of pigment epithelium into neural retina. Acta. Anat. 42,31-70.
PARK AND HOLLENBERG
Retinal
PARK, C. M. and HOLLENBERG, M. J. (1989). Basic fibroblast growth factor induces retinal regeneration in viva. Dev. Biol. 134,201-205. PROVIS, J. M., and PENFOLD, P. L. (1988). Cell death and the elimination of retinal axons during development. Prog. Neurobiol. 31,331347. REYER, R. W. (1977). The amphibian eye: Development and regeneration. In “Handbook of Sensory Physiology” (Crescitelli, F., Ed.), Vol. VII/5, pp. 309-390. Springer-Verlag, Berlin. RODRIGUES, M., WALDBILLIG, R. J., RAJAGOPALAN, S., HACKETT, J., LEROITH, D., and CHADER, G. J. (1988). Retinal insulin receptors: Localization using a polyclonal anti-insulin receptor antibody. Brain
Res. 443,389-394.
ROMANOFF, A. L. (1960). The Avian Embryo. Macmillan Co., New York. SCHWEIGERER, L., MALERSTEIN, B., NEUFELD, G., and GOSPODAROWICZ, D. (1987). Basic fibroblast growth factor is synthesized in cultured retinal pigment epithelial cells. Biochem. Biophys. Res. Cmnmun,. 143, 934-940. SPORN, M. B., ROBERTS, A. B., WAKEFIELD, L. M., and DE CROM-
Regeneration
by Growth
Factors
333
BRUGGHE, B. (1987). Some recent advances in the chemistry and biology of transforming growth factor-beta. J. Cell Biol. 105,10391045. STROEVA, 0. G. (1960). Experimental analysis of the eye morphogenesis in mammals. J. Evnbryol. Exp. Morphol. 8,349-368. THOENEN, H. (1991). The changing scene of neurotrophic factors. Trends Neurosci. 14,165-170. VEOME~, G., KUSCZYNSKI, C., KAZAKOFF, P., and RIZZINO, A. (1989). Cell density regulates the number of cell surface receptors for fibroblast growth factor. B&hem. Biophys. Res. Cwmmun. 159,694-700. WALDBILLIG, R. J., FLETCHER, R. T., SOMERS, R. L. and CHADER, G. J. (1988). IGF-I receptors in the bovine neural retina: Structure, kinase activity and comparison with retina insulin receptors. Exp. Eye
Res. 47,587-607.
WALICKE, P. A., and BAIRD, A. (1989). Trophic effects of fibroblast growth factor on neural tissue. Prog. Brain Res. 78,333-338. ZICK, Y., SPIEGEL, A. M., and SAGI-EISENBERG, R. (1987). Insulin-like growth factor I receptors in retinal rod outer segments. J. Biol. Chem.
262,10259-10264.
