Article Figures & Data
Figures
- Figure 1.
Developmental defects and microphthalmia in Gsk3-deficient retina with aberrant nuclear translocation of β-catenin, a key effector of the Wnt canonical pathway. A, IHC of E12.5 retinas from Gsk3αf/fβf/f mice expressing or not α-Cre using a pan-GSK3 antibody (green) shows efficient deletion at the periphery where the Cre expression has been previously reported (delimited by dashed line). Scale bar, 100 μm. B, H&E staining on methacrylate sections at E12.5, E14.5, and P2 reveals large retinal morphogenesis defects in Gsk3αf/fβf/f;α-Cre with blood invasion into the eyeball (showed by white arrow). L, Lens; NR, neural retina. Scale bars: E12.5 and E14.5, 100 μm; P2, 500 μm; E14.5, right, magnified image, 50 μm.
- Figure 2.
One allele of either Gsk3α or Gsk3β is sufficient for the development of a functional retina. A, Immunoblot analysis of protein extracts from 2-month-old animals with a different combination of Gsk3α and Gsk3β floxed alleles (Gsk3αf/fβ+/+, Gsk3α+/+βf/f, Gsk3αf/+βf/f, or Gsk3αf/fβf/+) with or without Cre recombinase using anti-pan GSK3 antibody (recognizing both isoforms) reveals decreased expression of GSK3α or GSK3β (arrowheads). α-Tubulin is used as a loading control. B, IHC on 2-month-old retinal sections from control and Gsk3αf/+βf/f; α-Cre retinas with or without Cre recombinase using anti-GSK3β antibody (red) showing ubiquitous Gsk3β expression in all retinal layers, whereas its expression is lost in the Cre-expressing retina. C, Expression of only one Gsk3 allele (Gsk3α) is sufficient for proper photoreceptor development. IHC using anti-rhodopsin (Rho; red) and anti-Cone arrestin (Arr3; red) antibodies to label rod and cone photoreceptors, respectively. D, Expression of only one Gsk3 allele (Gsk3α) is sufficient for proper interneuron development. IHC using anti-Calretinin (Calr; green) and anti-calbindin (Calb; red) antibodies to label horizontal and amacrine cells, respectively. onl, outer nuclear layer; inl, inner nuclear layer; gcl, ganglion cell layer. Scale bar, 20 μm. E, F, ERG recording in 2-month-old Gsk3αf/+βf/f;α-Cre animals and littermate controls. Photopic (cones; E) and scotopic (rods; F) responses in Gsk3αf/+βf/f;α-Cre animals are similar to those in controls. The mean ± SEM intensity response curves of a- and b-wave responses were averaged from eight biological replicates of each genotype.
- Figure 3.
Gradual loss of Gsk3α and/or Gsk3β leads to an increased number of Brn3a-positive retinal ganglion cells displaced in the INL of adult retina. A, Brn3a (red) and NF68 (green) IHC on 2-month-old Gsk3αf/+βf/f; α-Cre mouse retinas reveals the presence of supernumerary dRGCs (arrows) in the INL of Gsk3αf/+βf/f; α-Cre compared with littermate controls. Top panel represents control retinas, middle panel represents a peripheral retinal area, and bottom panel represents a more central area. Scale bar, 20 μm. B, dRGCs send their axons into the optic nerve. Visualization of dRGCs after 3D reconstruction of 2-month-old flat mounted retinas of control and Gsk3αf/+βf/f; α-Cre animals following retrograde labeling with rhodamine-dextran applied onto the optic nerve. inl: inner nuclear layer, gcl: ganglion cell layer. C, Gradual loss of Gsk3α and Gsk3β alleles (Gsk3αf/fβ+/+, Gsk3α+/+βf/f, Gsk3αf/+βf/f or Gsk3αf/fβf/+) leads to a gradual increase of Brn3a-positive RGCs located to the INL, with the highest number observed in Gsk3αf/+βf/f; α-Cre animals. Left histograms represent counting of the total number of Brn3a-positive cells per section located in the GCL (left) or in the INL (middle). Right histogram represents the percentage of the dRGCs among the total number of Brn3a-positive cells per section for each combination. Mean ± SEM values are presented from four biological replicates. A nonparametric Mann–Whitney U test was applied. *p ≤ 0.05; ns, non significant. D, Brn3a (red) and Rbpms (green) IHC on 2-month-old mouse retinas reveal the coexpression of these two RGC markers (dRGCs, arrows) in the INL of both Gsk3αf/+βf/f; α-Cre dRGCs and in littermate controls. Scale bar, 20 μm. E, Flat-mounted retinas from Gsk3αf/+βf/f; α-Cre and littermate controls labeled with anti-Rbpms antibody demonstrated the large number of Rbpms-positive dRGCs in the INL of Gsk3αf/+βf/f; α-Cre mice. Scale bar, 50 μm. See Extended Data Figure 3-1 for Islet1 and Rbpms colocalization used to complete dRGC characterization and the repartition of the dRGCs in the retina. See Extended Data Figure 3-2, showing that Brn3a-positive dRGCs are not positive for amacrine or horizontal cell markers.
- Figure 4.
dRGCs are produced in the same differentiation wave as oRGC located in the GCL. A, EdU-positive cells (green) and Brn3a-positive cells (red) were found both in the GCL and the INL of 30-day-old Gsk3αf/+βf/f; α-Cre animals after a single injection of EdU at E12.5. B, Percentage of EdU- and Brn3a-positive cells located either in the GCL or in the INL among total number of Brn3a-positive cells. Mean ± SEM values are presented from three to four biological replicates. A nonparametric Mann–Whitney U test was applied. ns, non significant. C, Brn3a (red) and NF68 (green) immunostaining on P0 mouse retinas revealed that a large number of dRGCs were already present in Gsk3αf/+βf/f; α-Cre mice, but were fewer in littermate controls (white arrows). D, Left stacked histogram represents the counting of the total number of Brn3a-positive cells per section located in the GCL (white bars) and the INL (black bars) of Gsk3αf/+βf/f; α-Cre retinas. Right histogram represents the percentage of the dRGCs among the total number of Brn3a-positive cells per section. Mean ± SEM values are presented from six biological replicates. A nonparametric Mann–Whitney U test was applied. **p ≤ 0.01. inl, inner nuclear layer; gcl, ganglion cell layer. Scale bar, 20 μm.
- Figure 5.
Lack of Gsk3β results in RGC projections into the ipsilateral medial terminal nucleus. A, All panels are light sheet fluorescence microscopy of solvent-cleared adult brain from control, Gsk3αf/fβ+/+; α-Cre, Gsk3α+/+βf/f; α-Cre, and Gsk3αf/+βf/f; α-Cre animals after intravitreal injection of CTB coupled to either Alexa Fluor-555 or Alexa Fluor-647. Ipsilateral projections of RGCs into the MTN were observed in the absence of Gsk3β expression. NOT, Nucleus of optic tract; LGN, lateral geniculate nucleus; vLGN, ventral LGN; IGL, intergeniculate leaflet; OPT, olivary pretectal nucleus; MTN, medial terminal nucleus; dMTN, dorsal MTN; vMTN, ventral MTN; OT, optic tract; SCN, suprachiasmatic nucleus; ON, optic nerve; SC, superior colliculus. Scale bar, 1 mm. *Ipsilateral MTN. B, Quantification of the signal intensity ratio between ipsilateral and contralateral MTN in controls and Gsk3 mutants (including Gsk3αf/fβ+/+; α-Cre, Gsk3α+/+βf/f; α-Cre, and Gsk3αf/+βf/f; α-Cre). A nonparametric Mann–Whitney U test was applied. ns, non significant. **p ≤ 0.01. See Extended Data Figure 5-1 for costaining of the CTB-positive cells with Brn3a and NF68.
- Figure 6.
Whole-transcriptome meta-analysis suggests that dRGCs in the Gsk3αf/+βf/f; α-Cre retina are DS-RGCs. A, Volcano plot representation of differentially expressed genes between Gsk3αf/+βf/f; α-Cre and control retinas plotted on the x-axis (log2 scale). FDR-adjusted significance is plotted on the y-axis. Orange and blue dots indicate significantly upregulated and downregulated genes in Gsk3αf/+βf/f; α-Cre retinas, respectively. Vertical dashed lines represent FC = 1.5. Horizontal dashed line represents FDR = 0.05. B, Chord plot representation of DEGs related to GO (Gene Ontology) annotations belonging to either molecular function (MF) or biological process (BP). Overlaps in GO annotation among genes within each category are visualized. *Genes expressed in previously published purified RGCs (blue, slightly expressed genes in RGCs between 1 and 5 FPKM; red, highly expressed genes in RGCs >5 FPKM). C, qRT-PCR validation of selected DEGs identified by RNA-Seq analysis. Differential expression analysis by qRT-PCR of Cartpt, Th, Epha2, Cplx1, Chrna5, Chrna2, Chrna7, and Chrnb4 in Gsk3αf/+βf/f; α-Cre retinas at 2 months of age, relative to levels in littermate control retinas. All values are expressed as the mean ± SEM from three biological replicates. A nonparametric Mann–Whitney U test was applied, *p ≤ 0.05. D, IHC on 2-month-old mouse retinas reveals the presence of a subset of dRGCs (Rbpms-positive dRGCs, red) in Gsk3αf/+βf/f; α-Cre expressing either the transcription factor Tbr2 (gray) or Foxp2 (green). Arrows indicate Tbr2 and Rbpms-positive dRGCs; arrowheads represent Foxp2 and Rbpms-positive dRGCs. onl, outer nuclear layer; inl, inner nuclear layer; gcl, ganglion cell layer. Scale bar, 50 μm. E, The mean OMR indices (±SEM) are plotted as a function of spatial frequency for each genotype (n = 13 for Gsk3αf/+βf/f and 18 for Gsk3αf/+βf/f;α-Cre genotype). The baseline (1; dashed line) represents unspecific head movements and no response to the stimulus. OMR at 100% and 50% contrast in Gsk3αf/+βf/f; α-Cre mice (dashed line) and controls (black line). A Grubbs’ test was performed at 5% to remove outliers followed by two-way ANOVA: *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. Extended Data Figure 6-1 for the hierarchical clustering of the DEGs. See Extended Data Figure 6-2 for pathway analysis results.
Tables
- Table 1
List of primary and secondary antibodies used for immunohistochemistry and western blot
Antigen Host Supplier Catalog no. Dilution IHC WB Primary antibodies α-Tubulin Mouse Sigma-Aldrich T5168 1:200.000 GSK3α/β Mouse Thermo Fisher Scientific 44–610 1:250 1:1000 GSK3β Mouse BD 610201 1:250 Brn3a Mouse Santa Cruz Biotechnology sc-8429 1:200 Calbindin D-28k Rabbit Swant 300 1:100 Calretinin Mouse EMD Millipore MAB1568 1:1000 Cone arrestin Rabbit EMD Millipore AB15282 1:1000 Rhodopsin Mouse Abcam MAB5316 1:2000 Tbr2 Rat Thermo Fisher Scientific 14–4876 1 :300 Foxp2 Goat Santa Cruz Biotechnology sc-21069 1 :1000 Rbpms Rabbit PhosphoSolutions 1830-RBPMS 1 :400 CHAT Goat Millipore AB144P 1 :100 Secondary antibodies Alexa Fluor-555 anti-mouse IgG2A Goat Thermo Fisher Scientific A21127 1:1000 Alexa Fluor-555 anti-mouse IgG2B Goat Thermo Fisher Scientific A21147 1:1000 Alexa Fluor-488 anti-rabbit Donkey Thermo Fisher Scientific A21206 1:1000 Alexa Fluor-488 anti-mouse IgG1 Goat Thermo Fisher Scientific A21240 1:1000 Alexa Fluor-488 anti-rabbit Goat Thermo Fisher Scientific A21244 1:1000 HRP anti-mouse IgG Goat Sigma-Aldrich A4416 1:5000 Alexa Fluor-488 anti-goat Donkey Thermo Fisher Scientific A11055 1:1000 WB, Western blot; IHC, Immunohistochenistry.
Gene name Primer forward Primer reverse Cartpt 5′-TAAAGTTTGCGTTCCCCTCAG-3′ 5′-CAACACCATTCGAGGCATTCT-3′ Th 5′-ACTATGCCTCTCGTATCCAGC-3′ 5′-CGGATGGTGTGAGGACTGTC-3′ Epha2 5′-GACCTCCCCATCTTCATTTGG-3′ 5′-GCGTACAGTGCCCTAGTCATA-3′ Cplx1 5′-GGTGATGAGGAAAAGGACCCC-3′ 5′-TCTTGGCGTACTTTGCTTTGC-3′ Chrna5 5′-CTTGAGTACCAACACTGTCCG-3′ 5′-CCAGTACTCCAAAGATGCCCT-3′ Chrna2 5′-CATTATCGTCTGCTTCCTGGG-3′ 5′-CTTGGAGCCAACATGAGGGA-3′ Chrna7 5′-CTGTAGCTGTCGGTCTTGAGA-3′ 5′-CAATGATATGCCGGTGATGGG-3′ Chrnb4 5′-AAACTGATCTGGCTACCTCCC-3′ 5′-GTAGAGAGTCCAGGAGATGCC-3′
Figure 3-1
dRGCs express the nuclear factor Islet-1. A, IHC on 2-month-old mouse retina reveals that most dRGCs (Rbpms-positive dRGCs, white arrows, red) in the INL of Gsk3αf/+βf/f; α-Cre and littermate controls were positive for Islet-1 (green), a marker expressed in the nuclei of ganglion cells, and of cholinergic amacrine cells, ON-bipolar cells, and subpopulations of horizontal cells. onl, Outer nuclear layer; inl, inner nuclear layer; gcl, ganglion cell layer. Scale bar, 50 μm. B, Counting on flat mount of Rbpms- or Brn3a- positive cells located in the INL at the dorsal, ventral, nasal, and temporal part of control and Gsk3αf/+βf/f;α-Cre retina. Histogram represents the number of Brn3a- or Rbpms-positive cells per field. Mean ± SEM values are presented from four biological replicates. Download Figure 3-1, TIF file.
Figure 3-2
Brn3a-positive cells located in the INL of Gsk3αf/+βf/f; α-Cre retina are dRGCs. Brn3a-positive RGCs located in the INL of Gsk3αf/+βf/f; α-Cre retina do not express markers of other INL neurons such as CHAT or calbindin (Calb). onl, Outer nuclear layer; inl, inner nuclear layer; gcl, ganglion cell layer. Arrowheads indicates Brn3a-positive dRGCs. Scale bar, 20 μm. Download Figure 3-2, TIF file.
Figure 5-1
Intravitreal injection of CTB labels dRGCs. After intravitreal injection of CTB coupled to an Alexa Fluor-555 (red) in Gsk3αf/+βf/f; α-Cre eye led to the labeling of Brn3a-positive (green) and NF68-positive (gray) cells located in the INL. Scale bars, 20 μm. Download Figure 5-1, TIF file.
Figure 6-1
Hierarchical clustering of the identified differentially expressed genes. Hierarchical clustering representing the 111 DEGs [abs(FC), ≥1.5; FDR, ≤0.05; FPKM, >1] between 2-month-old Gsk3αf/+βf/f; α-Cre retina and those of littermate controls were clustered by their z-score. Each column for each genotype corresponds to one sample. For both groups, triplicates were analyzed. Left, Downregulated genes; Right, upregulated genes. Download Figure 6-1, TIF file.
Figure 6-2
Identification of enriched pathways from DEGs identified in 2-month-old Gsk3αf/+βf/f; α-Cre retina. A, Gene ontology (GO) annotations of DEGs in Gsk3αf/+βf/f; α-Cre retinas compared with those in littermate controls. Top over-represented pathways for biological process (BP), molecular function (MF), KEGG (Kyoto Encyclopedia of Genes and Genomes), and TRRUST (transcriptional regulatory relationships unrevealed by sentence-based text mining) were identified by enrichment analysis using Metascape. B, Circular visualization for BP and MF of GO enrichment analysis. Downregulated genes (blue dots) and upregulated genes (red dots) within each GO pathway are plotted based on logFC. The z-score bars indicate whether an entire GO category is more likely to be increased or decreased based on the genes within it. Download Figure 6-2, TIF file.
In this issue
