Efficiency and morphological impact of Sox9 deletion on the adult mouse retina. (A) Hematoxylin-eosin stained histological sections of retinas from control and Sox9Δ/Δ mice. Mutant retinas exhibit variability in phenotype severity, with some showing mild morphological defects and a few displaying an extreme phenotype characterized by the loss of the outer nuclear layer (ONL). (B) Double immunofluorescence for SOX8 (green) and SOX9 (red) in retinal sections from adult control and Sox9Δ/Δ mice. The severity of phenotypes correlates with the extent of Sox9 inactivation. Regions completely lacking SOX8+ cells are indicated by arrowheads. DAPI (blue) was used to counterstain nuclei. (C) Quantification of the percentage of SOX8+ cells co-expressing SOX9 in control and Sox9Δ/Δ retinas with mild and extreme phenotypes. (D) Quantification of SOX8+ cells per 100 μm of inner nuclear layer (INL) length in control and Sox9Δ/Δ retinas. The black scale bar represents 50 μm in A, and the white scale bar represents 100 μm in B. GCL, Ganglion cell layer; INL, inner nuclear layer; RPE, retinal pigmented epithelium.

Immunodetection of retinal cell markers in adult retinas from control and Sox9Δ/Δ mice. (A) Immunofluorescence analysis of cone photoreceptor cells in retinal whole-mounts (upper images) and histological sections (lower images) using double staining for OPN1SW (S opsin) and OPN1LW (M opsin). (B) Immunofluorescence analysis of rod photoreceptor cells in retinal whole-mounts (upper images) and histological sections (lower images) using Rhodopsin staining. (C) Immunofluorescence analysis of ganglion cells in retinal sections using BRN3A staining. (D) Immunofluorescence analysis of amacrine and horizontal cells in retinal sections using double staining for PAX6 and AP2α. DAPI (blue) was used to counterstain nuclei. The scale bars in A and B represent 1 mm for the top row and 50 µm for the bottom row; the scale bars in C and D represent 90 µm.

Assessment of retinal damage and apoptosis in Sox9-deficient mice. (A) TUNEL staining of adult retinas from control and Sox9Δ/Δ mice with mild and extreme phenotypes. A large number of TUNEL-positive photoreceptor cells is evident in mutant mice with mild phenotypes, indicating extensive apoptotic events affecting the outer nuclear layer. (B) Immunofluorescence staining of GFAP in adult retinas from control and Sox9Δ/Δ mice with mild and extreme phenotypes. Müller glial cell activation is observed in Sox9-deficient retinas, with GFAP expression extending across the entire thickness of the retina in extreme phenotypes, suggesting progressive gliosis. The scale bars represent 50 μm.

Analysis of the fate of SOX9-expressing cells through lineage tracing experiments. (A- B) Immunofluorescence for SOX9 in limbal (A) and corneal (B) sections of adult control mice. DAPI (blue) was used to counterstain nuclei. (C) Schematic of the genetic lineage tracing strategy. Tamoxifen (TX) was administered to label SOX9-expressing cells and their progeny. (D-F) Analysis of EYFP expression in the limbus (D-E) and cornea (F) of Sox9-EYFP adult mice. At 10 days after TX administration (10 DATX), discrete EYFP-positive clones are visible in the limbal region (D). By 20 DATX, these clones expand towards the peripheral cornea, with some cells reaching the outer epithelial layers at the limbus-cornea border (arrowhead in E). In the central cornea, EYFP-positive cells extend along the entire epithelium by 20 DATX (arrowhead in F). (G) Short-term whole mount X-gal staining of Sox9-LacZ eyes. The first LacZ-positive cells scattered across the limbal and corneal surface appeared at 3 DATX (arrows). (H) Long-term whole mount X-gal staining of Sox9-LacZ eyes. At 45 DATX, LacZ-positive stripes emerge from the limbus and extend into the peripheral cornea (arrow). At 60 DATX, circumferential clones are observed in the limbus (arrowheads). At 90 DATX, three clone types are visible: circumferential clones in the limbus (arrowheads), stripes from the limbus reaching the central cornea (arrow), and stripes without a base in the limbus (asterisk). At 365 DATX, the first two clone types are mainly observed, though in reduced numbers and larger sizes (see arrowheads and arrow as in the 90 DATX picture), with some stripes without a base in the limbus (asterisk). Scale bars in E and F represent 50 µm; scale bars in G and H represent 500 µm.

Single-cell RNA-sequencing (scRNA-seq) analysis of Sox9 expression in the mouse limbal and corneal epithelium. (A) Uniform manifold approximation and projection (UMAP) visualization of scRNA-seq data of isolated limbal epithelial cells from Altshuler et al. (2022). (B) Density plot of Sox9 expression. (C) Violin plot of Sox9 expression levels across the different cell clusters identified in A. (D) Pseudotime partition-based graph abstraction (PAGA) graph of the trajectory of limbal and corneal cells. (E) Sox9 expression in the trajectory of cell differentiation described in D. (F) Violin plot of Sox9 expression according to cells exhibiting high and low expression levels. (G) Volcano plot of differentially expressed genes (DEG) between cells exhibiting high and low Sox9 expression levels. (H) Gene ontology analysis of DEG identified in G. CB, corneal basal; CJB, conjuntiva basal; CJS, conjuntiva suprabasal; CS1, corneal suprabasal 1; CS2, corneal suprabasal 2; ILB, inner limbal basal; LS, limbal suprabasal; Mit 1, mitiotic 1; Mit 2, mitotic 2; OLB, outer limbal basal.

Impact of Sox9 deletion on limbal stem cell differentiation and clonogenic capacity. (A) Hematoxylin-eosin stained histological sections of the limbus and cornea from control and Sox9Δ/Δ mice. (B) Immunofluorescence staining for SOX9 (red) and P63 (green) in the limbus and cornea from control and Sox9Δ/Δ mice. (C) Schematic of the generation of Sox9Δ/Δ-LacZ mice. Tamoxifen (TX) was administered for ten days to label Sox9-deleted cells and their progeny. (D) X-gal staining of whole eyes from control and Sox9Δ/Δ-LacZ mice at 98 days after TX administration. Control corneas exhibited numerous LacZ+ patches and elongated stripes originating from the limbus. In contrast minimal X-gal staining was observed in Sox9Δ/Δ-LacZ corneas, indicating impaired clonogenic capacity in Sox9-deleted cells. (E) Quantification of X-gal-stained surface areas in control and Sox9Δ/Δ-LacZ mice. Scale bar in A represents 100 µm; Scale bar in B represents 50 µm; Scale bar in D represents 1 mm for the top row and 500 µm for the bottom row.