May 1, 2011;
The Retinal Homeobox (Rx) gene is necessary for retinal regeneration.
The Retinal Homeobox (Rx) gene is essential for vertebrate eye
development. Rx function is required for the specification and maintenance of retinal progenitor cells
(RPCs). Loss of Rx function leads to a lack of eye
development in a variety of species. Here we show that Rx function is also necessary during retinal regeneration. We performed a thorough characterization of retinal regeneration after partial retinal resection in pre-metamorphic Xenopus laevis. We show that after injury the wound is repopulated with retinal progenitor cells
(RPCs) that express Rx and other RPC marker genes. We used an shRNA-based approach to specifically silence Rx expression in vivo in tadpoles. We found that loss of Rx function results in impaired retinal regeneration, including defects in the cells that repopulate the wound and the RPE
at the wound site. We show that the regeneration defects can be rescued by provision of exogenous Rx. These results demonstrate for the first time that Rx, in addition to being essential during retinal development, also functions during retinal regeneration.
[+] show captions
Fig. 1. The retina is essentially regenerated 30 days after resection.
(A–D) The progress of regeneration was analyzed by hematoxylin and eosin staining. (A) The retina after resection of the nasal-dorsal quarter on day 1. The site of resection is evidenced by the disruption of the retinal lamination and RPE (red asterisks). (B) On post-resection day 3 the RPE has re-assembled around the site of resection (red arrow) and cells have begun to fill in the wound. (C) On post-resection day 13 the RPE has closed around the wound (red arrow) and RPCs have repopulated the wound. (D) On post-resection day 30 the lamination of the retina is completely restored and the resection site is no longer evident.
(E–J) Analysis of regeneration progress using markers of differentiated neural cell types. Immunolabeling for Islet-1 (E–G) and Rhodopsin (H–J) in control retinas (E, H) and regenerating retinas at 15 days (F, I) and 30 days post-resection (G, J). Control retinas shown in panels E and H are from sibling embryos to those shown in panels G and J, respectively. At 15 days post-resection, the putative RPCs are still present at the site of resection (F, I; red bracket). The putative RPCs are not immunoreactive to Islet 1 (F; red bracket) or Rhodopsin (I; red bracket) antibodies. At 30 days post-resection, the putative RPCs are absent from the nasal-dorsal quarter of the retina and complete retinal lamination is observed by immunoreactivity to Islet-1 (G) and Rhodopsin (J). Uninjured retinas lack putative RPCs in the nasal-dorsal quarter and show Islet-1 and Rhodopsin immunoreactivities (E, H). L — lens; G — ganglion cell layer, I — inner nuclear layer; and P — photoreceptor layer. Scale bar = 50 μm.
Fig. 2. The regenerating wound is populated by retinal progenitor cells and is organized similarly to the CMZ.
(A–C) In situ hybridization performed using retinal sections of embryos at 9 days post-resection. Cells filling the regenerating wound express pan-RPC markers Rx1A (A), Pax6 (B), and Sox2 (C). (D, E) Cells filling the regenerating wound are proliferating. Immunolabeling of regenerating retinas at 9 days post-resection with anti-BrdU antibody. The putative RPCs incorporate BrdU and are immunoreactive to the anti-BrdU antibody (E, red bracket). The nasal-dorsal quarter of an uninjured retina lacks proliferating RPCs (D). (F, G) In situ hybridization performed on sections of embryos at 9 days post-resection with riboprobes for Notch1 (F) or NeuroD (G). (H) Double in situ hybridization for Notch1 (blue) and NeuroD (red). Different subsets of the RPCs (red) express Notch1 and NeuroD. Notch is expressed closer to the center of the wound (H; blue brackets) than NeuroD (H; red brackets) confirming that the expression of these two markers begins in different subsets of the RPCs that repopulate the wound. (I, J) The cyclin-dependent kinase inhibitor Xic1 is expressed at the extreme periphery of the regenerating region. (I) In situ hybridization for Xic1 (red brackets) demonstrates expression at the periphery of the regenerating wound and not in the center (blue bracket). (J) Overlay of BrdU incorporation (fluorescent green) and Xic1 in situ hybridization from (I). Proliferating cells are largely in the center of the regenerating wound (blue bracket), with little overlap with cells expressing Xic1 (red brackets). (K) Left — Model of normal CMZ (adapted from Perron et al., 1998). Right — Model of the CMZ formed in the regenerating wound. Scale bar = 50 μm.
Fig. 3. Retinal regeneration is abnormal in Rx knockdown tadpoles.
(A–D) Histological staining of regenerating retinas of a control non-transgenic tadpole (A), a control shRNA transgenic tadpole (B), and Rx shRNA transgenic tadpoles scored at 9 days post-resection (C, D). Rx shRNA transgenic tadpoles display shorter and/or rounder RPCs that are sometimes disorganized (C) and incompletely re-formed, disorganized RPE (D). (E–H) Rx (E, F) and Pax6 (G, H) expression is markedly reduced in the cells that repopulate the wound in Rx shRNA transgenic tadpoles. In situ hybridization on retinal sections of regenerating retinas from Rx shRNA transgenic tadpoles (F, H) and control non-transgenic tadpoles (E, G). Rx expression is markedly reduced in the cells that repopulate the wound in Rx shRNA transgenic tadpoles, but is not reduced in the Rx expressing cells at the INL (F). Pax6 expression is also reduced in the cells repopulating the wound in Rx shRNA transgenic tadpoles, but not in the INL or GCL (H). (I, J) Expression of Sox2 is also markedly reduced in the cells that repopulate the wound (I, red bracket). (J) Overlay of panel I with BrdU incorporation visualized by immunofluorescence (fluorescent green color). Arrow indicates RPE at the wound site; bracket indicates RPCs at the wound site. Scale bar = 50 μm. (K) Number of RPCs in the wound sites of regenerating retinas from control nontransgenic or Rx shRNA transgenic tadpoles. Each dot represents the RPC count from a single regenerating retina. The horizontal bar represents the average of the 5 counts shown; the vertical bar represents standard deviation from the mean for each group.
The effects of Rx knockdown on regeneration can be rescued by mouse Rx. (A) Upper construct: schematic of the X. tropicalis Rx (tRx) genomic locus showing the relative positions of ultraconserved genomic elements UCE2 and 3 (red) within the tRx regulatory region (gray). The Rx coding region (CDS) is indicated (blue). Lower construct: transgene containing a 3 kb portion of the X. tropicalis Rx locus (tRx3000), UCE2, and a GFP expression cassette (green). (B–E) In situ hybridization using a GFP antisense riboprobe using sections of uninjured (B, C) or regenerating transgenic tadpoles (D, E). The tRx3000/GFP transgene is not expressed in the RPCs at the distal tip of the CMZ (B, red arrowhead) or RPCs at the center of the regenerating wound (D). Addition of UCE2 drives transgene expression in RPCs throughout the CMZ (C) and the regenerating wound (E). (F) Schematic of mRx rescue construct, containing X. tropicalis Rx transcriptional regulatory elements as shown in (A) and the mouse Rx coding region (green). (G, H) Hematoxylin and eosin staining of retinal sections from a non-transgenic tadpole (G) and a Rx shRNA+ rescue tadpole (H) at day 9 post-resection. (I) Quantification of regeneration impairment in Rx shRNA transgenic tadpoles relative to nontransgenic controls, control (CO) shRNA transgenic tadpoles, and tadpoles co-transgenic for mRx. Categories of phenotype severity are defined in Table 1. Scale bar = 50 μm.