Expression and role of retinoic acid receptor alpha in lens regeneration


*Author to whom all correspondence should be addressed.


The role of retinoids in eye development has been well studied. Retinoids and their receptors regulate gene expression and morphogenesis of the eye. In this study, a highly specific antagonist of retinoic acid receptor (RAR)-α was used in an attempt to study its function in lens regeneration. It was found that this antagonist inhibited lens regeneration and lens fiber differentiation. It was also shown that RAR-α is expressed in the lens during the process of regeneration. These results indicate that different RAR might have unique as well as redundant effects and patterns of expression in the regenerating lens.


The process of lens regeneration in adults occurs in only a few species of salamanders. Upon removal of the lens, these animals regenerate a complete, new lens through the transdifferentiation of the pigment epithelial cells (PEC) from the dorsal iris. Such an event might hold clues for the process of differentiation and genetic reprogramming, as well as to why other animals are not able to regenerate the lens (Tsonis et al. 2000, Tsonis 2000a, 2000b).

In the past few years, research into the molecular biology of this phenomenon has revealed the participation of several genes, but the mechanism that could be applied in other animals is still elusive (Del Rio-Tsonis et al. 1995, 1997, 1998, 1999). Our previous investigation has led us to believe that retinoic acid and its receptors might be involved in lens regeneration (Tsonis et al. 2000). While we found that exogenous retinoids did not affect the process of lens regeneration, inhibition of retinoid production or inhibition of their receptors showed some interesting results. Specifically, we were able to show that lens regeneration was affected by an antagonist to retinoic acid receptors (RAR). In most cases, lens regeneration was inhibited or retarded by the antagonist, but in a few spectacular cases the lens was regenerated from sites other than the normal dorsal site. We continued this research to more clearly delineate the components of such an effect. Since knockout experiments are impossible in newts, inhibiting the action of proteins by specific antagonists is an excellent method for our purpose. In a previous study we had used an antagonist that could inhibit the action of all RAR. In this study, we used an antagonist specific to RAR-α to determine the specific roles of RAR-α in the process of lens regeneration. We report here that RAR-α is expressed in the eye tissues of the newt and that its specific inhibition affects lens regeneration.

Materials and Methods

Animal operations

Newts (Notophthalmus viridescens) were lentectomized under general anesthesia by using 1% 3-aminobenzoic acid ethyl ester (Sigma, St Louis, MO, USA). Eyes were collected at different stages of lens regeneration for either histological staining or for expression studies.

Treatment with the retinoic acid receptor alpha antagonist

In our previous study we had established that the method of administration that produced the best results for inhibition or ectopic lens regeneration was to absorb the antagonist (20 mg/mL) on AGI-X2 formate beads (100–200 mesh; Bio-Rad, Hercules, CA, USA). Such treatment resulted in inhibition of lens regeneration from the dorsal iris and in ectopic lens regeneration from other sites. We therefore have selected this method and concentration of administration to examine and compare the effects of a specific antagonist to RAR-α. The beads were soaked in the solution and one of them was inserted in the eye cavity immediately upon lentectomy. The antagonist, AGN 194301 (Allergan, Irvine, CA, USA) was dissolved in dimethylsulfoxide (DMSO). Control animals received a bead soaked in DMSO only. AGN 194301 is a specific antagonist to RAR-α

. This compound is particularly effective in inhibiting retinoic acid-induced gene transcription at RAR-α (Teng et al. 1997).

In situ hybridization

Expression of RAR-α during lens regeneration was studied by in situ hybridization as described in detail in previous publications (Del Rio-Tsonis et al. 1997, 1998). The newt RAR-α probe was a gift from Dr J. P. Brockes (University College London, UK; Ragsdale et al. 1992).

Results and Discussion

Lens regeneration in the adult newt begins with proliferation and dedifferentiation of dorsal iris PEC. By dedifferentiation we mean the loss of characteristics that define PEC, such as pigmentation. Dedifferentiation initiates molecular events, such as re-entering the cell cycle, which are necessary for cell proliferation and the subsequent regeneration of the lens. At approximately 10 days post-lentectomy, a lens vesicle is formed from the depigmented dorsal PEC. Approximately 12–16 days post-lentectomy, the internal layer of the lens vesicle thickens and synthesis of crystallins begins. This marks the beginning of primary lens fiber differentiation. During days 15–19, proliferation and depigmentation of PEC slows down. In the internal layer, the lens fiber complex is formed, and in the margin of the external layers, non-dividing secondary lens fibers appear. By 18–20 days, the PEC have stopped proliferating and the lens fibers continue to accumulate crystallins. Lens regeneration is considered complete by day 25–30 (Tsonis 1999, 2000a, 2000b). Lens regeneration therefore is possible by transdifferentiation, which is the transformation of one cell type to another (in this case PEC to lens cells). The processes of dedifferentiation and transdifferentiation have been proven beyond any doubt in lens regeneration. These processes can also be observed when single PEC cells are placed in culture (Eguchi et al. 1974; Tsonis et al. 2001). As the PEC proliferate, they become depigmented and then transdifferentiate to lens cells. Therefore, while in many other regenerative tissues stem cells may play a role, such a possibility is very unlikely in lens regeneration.

Histological examination of the eyes 20 days after lentectomy indicated that in 30% of cases (seven out of 23) the lens was affected. In three of those the lens was completely absent and in the remaining four the lens was smaller (to varying degrees). Examples of such cases can be seen in Figure 1. Figure 1A shows a section through an eye with a regenerating lens 20 days post-lentectomy. This section is from a control eye (treated in DMSO solution) and shows a normally regenerated lens. No effects were seen in 31 control cases. The lens (fibers) is surrounded by the lens epithelium and is of normal size, filling the space between the dorsal and ventral iris. In Figure 1B–D we can observe the effects of the antagonist on lens differentiation and regeneration; in Figure 1B, a smaller lens has regenerated with apparent fiber differentiation in the interior. The bead can also be seen in this section. In Figure 1C and D, lens differentiation and regeneration has been completely inhibited. We can, however, see a small lens vesicle at the tip of the dorsal iris. In our previous study (Tsonis et al. 2000) we showed that an antagonist to all RAR was able to affect lens regeneration in more than 60% of cases. Also in that study, regeneration was observed from sites other than the dorsal iris. In the present study, treatment with a specific RAR-α antagonist resulted in fewer affected lenses. The differences in these effects should not be attributed to the concentration of the two different antagonists because the same method and concentration was used for administration. Rather, this might indicate that the different RAR might act in a redundant fashion or that different RAR can contribute to different stages of lens formation (or both). Redundancy in RAR is a well-established phenomenon (Kastner et al. 1994). Obviously, by using specific antagonists to RAR (as they become available) we will be able to delineate their exact function in lens regeneration. One obvious conclusion from the present study is that RAR-α is involved in the proliferation and differentiation process of the lens vesicle, but is not involved in the mode of regeneration as it pertains to the site of origin in the eye. Such effects could be mediated via the specific synthesis of retinoic acid along the dorsoventral axis in the eye (Manns & Fritzsch 1991; McCaffery et al. 1992, 1993; Wagner et al. 2000). Such spatial distribution patterns are known to affect eye morphogenesis as well as gene expression (Enwright & Grainer 2000).

Figure 1.

Effects of retinoic acid receptor (RAR)-α antagonist on lens regeneration. (A) Section through an eye with a normally regenerated lens 20 days after lentectomy. The lens (l) occupies the area between the dorsal iris (di; from where it was originated) and the ventral iris (vi). The lens fibers are surrounded by the lens epithelium (le). The neural retina (nr), ganglion cell layer of the neural retina (g) and cornea (c) are also denoted. (B–D) Sections through lentectomized eyes treated with the antagonist 20 days after lentectomy. Note the varying degrees of lens regeneration. In (B), a small lens has regenerated from the dorsal iris. This lens has differentiated fibers. The bead (b) is also visible in this section. In (C) and (D), however, there is only a small lens vesicle (arrowheads) without apparent lens fiber differentiation.

Expression of RAR-α was consistent with such effects. RAR-α was expressed weakly only in the ganglion layer of the retina in the intact eye. The intact lens was negative for RAR-α (Fig. 2A). However, as the PEC from the dorsal iris dedifferentiated and produced the lens vesicle, RAR-α transcripts were detected in the dedifferentiated cells and in the subsequent transdifferentiated lens epithelial cells (Fig. 2B). The expression in the epithelial cells was prominent even by day 20 post-lentectomy. However, expression in the differentiating lens fibers was also strong (Fig. 2C). In this respect, RAR-α shows similar patterns to the previously reported expression patterns for RAR-δ (Tsonis et al. 2000). Such patterns of expression of RAR are quite consistent with their roles. It is conceivable that blocking the function of RAR by specific antagonists would inhibit lens formation, as the differentiation of the lens fibers depends on the lens epithelium. Related to this, it has been shown before that retinoid receptors regulate crystallin synthesis in the lens (Gopal-Srivastava et al. 1998). The fact that we see inhibition of lens formation can then be explained by the blocking of RAR function. We believe that while the function of different RAR can be redundant, it can be complementary or unique as well. Extension of these studies with other specific antagonists of retinoid function might lead to a better understanding of the mechanisms of lens regeneration and its uniqueness in some urodele amphibia.

Figure 2.

Expression of retinoic acid receptor (RAR)-α in intact and regenerating lens. (A) Expression in the intact eye. Signal was detected only in the ganglion layer of the retina (arrowheads). (B) Expression in a regenerating lens vesicle 10 days after lentectomy. Expression is obvious in the dedifferentiated vesicle (arrowhead) from the dorsal iris (di). (C) Expression in a regenerated lens 20 days after lentectomy. Note expression in the lens epithelium (arrowhead) and the lens fibers. (D) Negative control, a section with a regenerated lens 20 days after lentectomy hybridized with the sense probe. Note the absence of signal.


This work was supported by a National Eye Institute (NEI) grant (EY10540) to P. A. T.