Afadin sorts different retinal neuron types into accurate cellular layers
- Department of Ophthalmology, University of California, San Francisco, United States
- Department of Ophthalmology, Stanford University School of Medicine, United States
- Department of Cell Biology and Human Anatomy, School of Medicine, University of California, Davis, United States
- College of Optometry, University of Houston, United States
- Viterbi Family Department of Ophthalmology, University of California, San Diego, United States
- Department of Physiology and Kavli Institute for Fundamental Neuroscience, University of California, San Francisco, United States
Figures
Figure 1
Retina-specific Afadin conditional mutants disrupt the cellular layer organization.
(A–H) Postnatal time course of mouse retinal cryosections in Afadin control (AfadinF/F) and Afadin knockout (AfadincKO). The Six3Cre-Afadin knockout (E–H) displays disruption of the tri-neuronal layer organization. As a result, it leads to a fused singular outer layer (fused INL/ONL) compared to control retinae (A–D), which contain three distinct nuclear laminae separated by two plexiform layers. Rosettes (F–H, asterisks) in the fused INL/ONL are visible from as early as P2 to adulthood and are devoid of cell bodies and primarily contain neurites (see Figure 2G, K and L) (F–H). The IPL is retained in AfadincKO but contains columnar-like structures of displaced neurons (F, arrow). In adult AfadincKO mice (H), there is significant shrinkage of the fused INL/ONL (see Figure 4E–H). Scale bars (A–H): 100 μm. (I–T) Afadin conditional knockout results in the mislocalization of major retinal cell types. In cross-section view (I–R), the control retina displays stereotypical lamination of three major cell types: bipolar cells (Chx10), retinal ganglion cells (RBPMS), and amacrine cells (AP2a) (I–M). In control retinae, RGCs and BCs stayed in the GCL and INL strictly, with very little displacement. ACs have about 13.2 ± 0.4% displacement (S). In contrast, AfadincKO showed aberrant localization of three major cell types (N–R). RGCs, ACs, and BCs display 33.9 ± 0.4%, 42.0 ± 3.7%, and 37.6 ± 2.4%, respectively (S). Across three replicates, there was no significant difference between cell counts across the three cell types (T). Unpaired two-sided Student’s t-tests; n.s., not significant; ****p<0.0001; ***p<0.001. Data presented as mean percentage mislocalized ± SEM. Mislocalization and cell density quantification were obtained from P14 mice. n=3 mice in each condition. Scale bars (I–T): 100 μm. (U) Generation of AfadincKO mice. Six3Cre transgenic mouse crossed with Afadin conditional knockout mouse. Exon 2 is flanked by LoxP sites, enabling Cre-mediated deletion, resulting in a frameshift and premature stop codon.
-
Figure 1—source data 1
Cell type quantifications.
- https://cdn.elifesciences.org/articles/105575/elife-105575-fig1-data1-v1.xlsx
Figure 2 with 2 supplements
Synaptic rosettes persist despite lateral displacement of cell types.
(A–H) Wholemount retinal cross-sections display lateral displacement of major cell types in AfadincKO. In a wholemount section view, AfadincKO (E–H) displays lateral displacement of cell types within unexpected laminae, which are absent in control (A–D). Notably, BCs (Chx10) are seen in the GCL (E), and RGCs are visible in the fused INL/ONL (G, H). ACs are found throughout all laminae in AfadincKO. The IPL in the AfadincKO contains regularly interspaced clusters of RGCs, BCs, and ACs, which form vertical bridging columns (F); IPL for the control retina is not shown. The rosettes are visible, with the neurites of RBPMS +RGCs projecting inward toward the rosette center (H). Scale bars (A–H): 50 μm. (I–L) Rosettes in the fused INL/ONL have characteristics of an ectopic IPL. Wholemount section (I) of an outer region of the fused INL/ONL showing starburst amacrine cell (SAC) processes labeled by VAChT forming a central rosette structure upon which BC processes colocalize. In a 60x magnification of the dashed region in (I), the spoke-like processes of rod bipolar cells stained with PKCα are visible (J). RGC dendrites also co-cluster in the rosette structure projecting centrally (L, RBPMS). A subset of these RGCs is Cartpt-positive, which labels ooDSGCs, indicating an ectopic IPL circuit composed of DSGCs is retained at the histological level (K). Scale bar (I): 300 μm; Scale bar (J): 50 μm; Scale bars (K, L): 30 μm.
Figure 2—figure supplement 1
Canonical synaptic pairs persist in AfadincKO despite mislocalization (related to Figure 2).
(A–F) Canonical amacrine cell (AC) and retinal ganglion cell (RGC) synaptic pairs in the inner plexiform layer (A–C) continue to co-fasciculate in AfadincKO (D–F), despite ectopic localization of both subtypes in the outer nuclear layer (ONL). These pairings include ON/OFF direction-selective RGCs (Cartpt) and SACs (VAChT) (left), glutamatergic ACs (VGlut3) and S3-IPL-targeting RGCs (Kv4.2) (center), and dopaminergic ACs (TH) and intrinsically photosensitive RGCs (ipRGCs) (OPN4) (right). Scale bars (A–F): 50 μm. (G, H) In the wholemount view, both αRGCs (Spp1) (G) and melanopsin-positive ipRGCs (H) were retained near the rosette structures in the fused INL/ONL. Additionally, TdTomato-positive RGCs from AAV-Retrograde injection into the SC are shown (H). Scale bars (G, H): 50 μm. (I) A diagram illustrating the major findings and wholemount sectioning procedure is shown (I).
Figure 2—figure supplement 2
Representative AfadincKO wholemount.
Representative wholemount displaying rosettes in the fused INL/ONL (A): wholemount retina labeled with RBPMS and VAChT aids in the quantification of the number of rosettes per retina. Quantification (B) of rosettes across four adult AfadincKO mice: 160±13 rosettes across four adult AfadincKO mice. Data presented as mean rosette count ± SD.
-
Figure 2—figure supplement 2—source data 1
Rosette count.
- https://cdn.elifesciences.org/articles/105575/elife-105575-fig2-figsupp2-data1-v1.xlsx
Figure 3
A significant fraction of displaced RGCs project to the central targets in the SC.
(A) SC sections labeled with bilateral retina injections. CTB-488 and CTB-555 dye were injected into the left and right eye, respectively, of both AfadinF/F and AfadincKO mice. An ectopic CTB-488 patch was found close to the midline of the left SC. Scale bar: 500 um. (B) Illustration showing stereotaxic protocol. AAV (Retro)-TdTomato was injected unilaterally into the right SC of adult mice. The retrograde virus was uptaken by RGC terminals in the SC and selectively labeled RGC somata in the retina. (C–E) Displaced RGCs in AfadincKO send axons to the SC. Retinal cross-sections of AfadinF/F mice show retrograde-AAV labeled RGCs restricted to the GCL layer (C). In AfadincKO sections, RGCs co-labeled by RBPMS and TdTomato are mislocalized into the fused INL/ONL (dashed circle) and are close to an IPL column (D). TdTomato-positive RGCs (arrows) are also mislocalized to a rosette structure in the fused INL/ONL (E). Scale bar: 50 um. (F) Quantifications of AAVretro-labeled RGC somata. 2.7 ± 2.4% of RGCs co-labeled with RBPMS and anti-TdTomato in control mice displayed mislocalization beyond GCL, versus 42.2 ± 8.8% of RGCs in AfadincKO. Unpaired two-sided Student’s t-tests; ****p<0.0001.
-
Figure 3—source data 1
AAV retrograde data.
- https://cdn.elifesciences.org/articles/105575/elife-105575-fig3-data1-v1.xlsx
Figure 4
Afadin mutants lose photoreceptor-mediated visual function in adults.
(A, B) Retinal cryosection displaying thinning and rod photoreceptor loss. Retinal cryosection taken from the central retina of AfadincKO (B) displays loss of recoverin-positive rod photoreceptors; the photoreceptor layer is maintained in control (A). Yellow arrow: displays remnant recoverin-positive patch in the peripheral retina. Scale bars: 500 um. (C–J) Postnatal time course of AfadincKO displaying progressive photoreceptor loss. Recoverin-positive photoreceptors display aberrant lamination in AfadincKO (G–J) but not in AfadinF/F (C–F). By early adulthood, AfadincKO have a near-complete loss of photoreceptors (J). Dashed red lines indicate the transition from ONL to OPL. Solid red lines indicate the transition from fused INL/ONL to IPL. Scale bars: 50 um. (K–M) Representative ERG traces shown for one right eye of both AfadinF/F (K) and AfadincKO (L) mouse when shown a dark-adapted intensity program of 30 cd.s/m2 white stimulus. In the control mouse, the a-wave amplitude was –273.3 uV, and the b-wave amplitude was 718.6 uV. In the mutant, the a-wave amplitude was –14.33 uV, and the b-wave amplitude was 40.95 uV. The ERG responses were quantified in (M). Average a-wave and b-wave responses for AfadinF/F and AfadincKO mice after dark-adapting overnight. Under scotopic conditions using 30 cd·s/m2 flash of white light, the average a-wave response was 235.6±49.0 uV for control and 17.0±4.3 uV for AfadincKO mice; the average b-wave response was 564.4±121.2 uV for control and 89.4±48.3 uV for AfadincKO mice. Under photopic conditions using 10 cd·s/m2 flash of white light, the average a-wave response was 8.2±6.7 uV for control and 8.3±6.3 uV for AfadincKO mice; the average b-wave response was 142.7±84.2 uV for control and 50.4±35.4 uV for AfadincKO mice. Unpaired two-sided Student’s t-tests; ns, not significant; ****p<0.0001; *p<0.1. Data presented as mean wave response (uV) ± SEM between right and left eyes across 6 different mice. ERG quantification was obtained from adult mice. n=6 mice per condition.
-
Figure 4—source data 1
ERG quantification.
- https://cdn.elifesciences.org/articles/105575/elife-105575-fig4-data1-v1.xlsx
-
Figure 4—source data 2
Raw electrophysiology ERG data exported from Diagnosys machine.
- https://cdn.elifesciences.org/articles/105575/elife-105575-fig4-data2-v1.pdf
Additional files
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Afadin sorts different retinal neuron types into accurate cellular layers
eLife 14:RP105575.
https://doi.org/10.7554/eLife.105575.2
