May 1, 2009;
Defining retinal progenitor cell competence in Xenopus laevis by clonal analysis.
Extrinsic cues and intrinsic competence act in concert for cell fate determination in the developing vertebrate retina
. However, what controls competence and how precise is the control are largely unknown. We studied the regulation of competence by examining the order in which individual retinal progenitor cells (RPCs) generate daughters. Experiments were performed in Xenopus laevis, whose full complement of retinal cells is formed in 2 days. We lineage-labeled RPCs at the optic vesicle stage. Subsequently we administered a cell cycle marker, 5-bromodeoxyuridine (BrdU) at early, middle or late periods of retinogenesis. Under these conditions, and in this animal, BrdU is not cleared by the time of analysis, allowing cumulative labeling. All retinal cell types were generated throughout nearly the entire retinogenesis period. When we examined the order that individual RPCs generated daughters, we discovered a regular and consistent sequence according to phenotype: RGC
, Ho, CPr, RPr, Am, BP, MG. The precision of the order between the clones supports a model in which RPCs proceed through stepwise changes in competence to make each cell type, and do so unidirectionally. Because every cell type can be generated simultaneously within the same retinal environment, the change in RPC competence is likely to be autonomous.
[+] show captions
References [+] :
Fig. 1. GFP+ retinal sections of stage 41 Xenopus tadpoles. Immunolabeled for GFP (red), BrdU (green) and rhodopsin (blue). Sclerad is up, vitread is down. Each cell type has a readily recognized morphology. Am, amacrine cell; BP, bipolar cell; CPr, cone cell; GCL, ganglion cell layer; Ho, horizontal cell; INL, inner nuclear layer; IPL, inner plexiform layer; MG, Müller glia; ONL, outer nuclear layer; OPL, outer plexiform layer; RPr, rod cell; RGC, retinal ganglion cell.
Fig. 2. Retinal sections of stage 41 Xenopus containing clusters of GFP-labeled cells. Sections were immunolabeled as in Fig. 1. (A,B,C,E,G,I) Monochrome images showing GFP labeling only. (A) BrdU injected at stage 30. This cluster has at least one member in each of the retinal layers, and they are tightly radially aligned. (B,C) BrdU injected at stages 33/34 (B) and 30 (C). Clusters contain radially aligned cells in the ONL and INL, except C, where one cell exhibits some tangential dispersion. (D-I) Pairs of images from the same clone; E,G,I show only the GFP label allowing identification of cell phenotype. The BrdU labeling of each cell is indicated by the color of its identifier: yellow, BrdU+; blue, BrdU–. (D,E) BrdU injected at stages 33/34. In Clone #146, the RPr were post-mitotic at the time of BrdU administration, but the BP and Am were not yet post-mitotic. (F,G) BrdU injected at stage 31. In Clone #90, an RGC and an RPr are BrdU– whereas two BP and an Am are BrdU+. (H,I) BrdU injected at stages 33-36. In Clone #153, all members are BrdU–.
Fig. 3. Features of GFP-labeled clusters suggest that members are progeny from individually labeled retinal progenitor cells. (A) Bar chart of the number of GFP-labeled clusters per eye. Almost as many eyes did not have GFP-labeled cells as had them. The proportion with more than one cluster decreased dramatically. (B) Bar chart of our data (black) obtained using DNA lipofection on the size of GFP lineage-labeled cell clusters compared with data (white) produced by intracellular injection of HRP (Holt et al., 1988). All cell cluster parameters for the two studies are similar, indicating that both successfully label cell clones.
Fig. 4. Mitotic landmarks of retinal development in Xenopus laevis and the pattern of genesis of the seven major cell types at the population level. (A) The timing of gfp transfection and BrdU injection according to three different scales: (1) hours post-fertilization at 22–24°C; (2) embryonic stage according to Nieuwkoop and Faber (Nieuwkoop and Faber, 1967); and (3) mitotic index of retinal cell population [modified from Holt et al. (Holt et al., 1988)]. (B) The frequency of BrdU+ and BrdU– cells of each phenotype for three stages of retinogenesis. Every cell type was generated at early and middle periods, and all but RGC, Ho and RPr were produced at the late period.
Fig. 5. Representation of 51 clones containing heterogeneous BrdU labeling, plus five each of randomly selected clones with all BrdU+ and all BrdU– profiles. When the retinal phenotypes (columns) are ordered as shown, we see a smooth transition from white (BrdU+) to black (BrdU–) that sweeps from the top right to the bottom left. Empty box, no cells present; white circle, all cells BrdU+; black circle, all cells BrdU–; half-black half-white circle, some cells BrdU–, some BrdU+. Arrows indicate the clones that do not fit the overall sequence. Asterisks indicate the clones shown in Fig. 2D-G.
Fig. 6. A matrix of pairwise comparisons of clones where one phenotype (reference) is 100% BrdU–, but the others (comparisons) are 100% BrdU+. This analysis reveals that the order of retinal cell birth is highly regular. Black boxes indicate the number of clones of the reference cell phenotype, i.e. the number of clones that exclusively contain BrdU– cells of this phenotype. White boxes indicate the number of clones of the comparison cell phenotype, i.e. the number of comparison clones that exclusively contain BrdU+ cells for a given reference phenotype. For example, of the 26 clones for horizontal cells (all Ho BrdU–), there are three cone clones (all CPr BrdU+).
Subsets of retinal progenitors display temporally regulated and distinct biases in the fates of their progeny.