Semaphorin 3f and post-embryonic regulation of retinal progenitors

Neural progenitors produce specific cell types that form the circuits of the nervous system. Extrinsic signals regulate both progenitor proliferation and the production of specific neuron types. Where progenitors reside within a progenitor niche determines to which of these signals they are exposed, and thus likely has important consequences on the progeny they produce. Little is known, however, of the signals that determine progenitor location within the niche. Here we show that a member of the Class III family of secreted Semaphorins, Semaphorin3fa (Sema3fa), is required for the orderly arrangement of progenitors with a niche present in the periphery of the larval and adult retina of zebrafish, the ciliary marginal zone (CMZ). We used whole mount and fluorescent slide in situ hybridization to show that CMZ progenitors express mRNAs for various Sema3 receptors, including for nrp2a, nrp2band plxna1. We investigated the function and gene expression of progenitors in wild type fish and in a previously characterized CRISPR/Cas9-generated sema3fa mutant allele (sema3faca304). We found that mutant juvenile fish had a reduced eye size, implicating Sema3fa in the ongoing production of retinal cells by the CMZ. We used EdU labeling, and PCNA and pHH3 immunolabelling to show that larval mutant CMZ progenitors show altered cell cycle parameters. We also performed whole mount in situ hybridization and immunohistochemistry in retinal sections with various mRNA and protein markers of different domains of the CMZ. We found that the spatial organization of functionally distinct progenitors is disrupted. An EdU pulse-chase experiment revealed that the generation of retinal cell types in the appropriate proportions and numbers, as determined by laminar location of cells, was disrupted in the mutant fish. Our data support a model whereby Sema3fa secreted by CMZ progenitors reduces adhesive interactions, which allows for smooth progression of progenitors through the niche, ensuring progenitors receive the correct recipe of extrinsic signals to secure the proper generation of new retinal circuits.

The dataset includes the measured values that formed the averaged data points in the graphs provided in the figures of the manuscript. Each point in the graphs comes from a separate animal (n) and the data is pooled between N=2-4 independent data sets. When areas were measured this was done on retinal sections, either plastic sections of whole mount in situ hybridization samples, or cryostat sections. For cryostat sections, samples were counter-stained with Hoescht to label nuclei. These sections came from the central retina that contained a lens. Two-four images/eye/larva of the central retina, with a lens, were collected. To normalize area measurements we measured the area of the CMZ, as defined by the peripheral edge of the inner plexiform and outer plexiform layers evident either in plastic sections, or as defined by the lack of Hoescht label in cryostat sections.

For the RT-qPCR data the values are from independent data sets, performed on mRNA isolated from 30 surgically isolated eyes. Three technical replicates were performed/data set for each primer set.

Methods
Methods

Animals: Larvae were used from the respective in-crosses of wild type (WT) Tupfel Long Fin (TL) fish, sema3faca304/ca304 zebrafish, and for the latter a corresponding WT line generated from the same clutch that gave rise to the specific sema3faca304/ca304 fish line (Halabi et al., 2021). For whole mount in situ hybridization, pigment synthesis was inhibited by adding 1-phenyl-2-thiourea (PTU) to the E3 solution at 12-24 hpf.

Eye Size Measurements: Live sema3faca304 and WT siblings were imaged from a dorsal orientation at 10 days post fertilization (dpf). Separate sets of one-month old fish were imaged from a lateral orientation. The areas of the pigmented eyes of larvae (dorsal view) and juveniles (lateral view) were measured in a masked fashion and normalized to the nose to swim bladder length and the size of the head (measured to the back of the gills), respectively.

Sema3fa Overexpression: We engineered a sema3fa heat shock construct, hsp70:sema3fa:p3E-MTpA in pdestTol2CG2 with myl7:egfp as the transgenesis marker (Kwan et al., 2007). The heat shock construct was injected into one cell-stage zebrafish embryos alongside transposase mRNA. Larvae (5 dpf) with eGFP positive heart expression were heat-shocked, and two days later processed for EdU labeling. Sema3fa overexpressing cells were identified by immunolabelling retinal cryostat sections for Myc and EdU cells by Click-iT. Numbers of EdU-positive cells (3-4 sections/larva) were counted in dorsal CMZ that showed no (Myc-) or considerable Myc (Myc+) immunolabelling.

Cdh2 analysis

Embryos injected with hsp70:sema3fa:p3E-MTpA were heat shocked at 52 hpf and 24 hours later embryos were fixed.Cryostat sections (12 µm) of 72 hpf WT and sema3faca304 retinas, or heat-shocked Sema3fa overexpression larvae, were immunostained with a polyclonal antibody against Cdh2. The area of the bright Cdh2 immunopositive label was represented as a percentage of the area of the CMZ, and the percentages for several sections averaged for each individual embryo.

Mutant CMZ analysis: Larvae were collected at 72 hours post-fertilization (hpf) and 7 days post-fertilization (dpf). Cell counts and gene expression domain areas were measured in plastic or frozen cryostat retinal sections and then normalized to the area of the CMZ. Frozen sections were counterstained with Hoescht to label cell nuclei. The CMZ domain was defined by the anatomical landmarks of the edges of the inner and outer plexiform layers. CMZ measurements were made in 2-3 sections through the central retina of a single eye for each larva, and data was pooled for 2-4 independent experiments. Each point in the graphs is from a single larva. Analysis of the CMZ in wild type and sema3fa mutant larve included:

1) Nearest Neighbour Analysis: Hoechst-labelled nuclei within the CMZ of retinal sections of 7 dpf fish, We used the Delaunay Triangulation plugin available in Fiji (Schindelin et al., 2012) to determine distances between the centres of nuclei of neighbouring Hoescht-labelled nuclei of retinal sections of 7 dpf fish.

2) In Situ Hybridization: We performed either whole mount in situ hybridization (ISH) or fluorescent ISH on retinal sections for 72 hpf and 7 dpf fish. Probes included: sema3fa, nrp2a, nrp2b, plxna1a, plxna1b, plxna2, plxna4, atoh7, neurod4 and vsx2, plxna3; and probes generated by PCR (col15a1b, cdkn1c F, ccnd1, hes6, bmp4, nrp1a, nrp1b) of cDNA generated from zebrafish embryos. Whole mount preparations were plastic sectioned (7 µm). In sections the areas of ISH signal were measured and normalized to the CMZ area. CMZ measurements were made in 2-3 sections through the central retina of a single eye for each larva, and data was pooled for 2-4 independent experiments.

3) Immunohistochemistry: Immunohistochemistry was performed on 12 µm cryostat cut frozen retinal sections. Antibodies included those against Phosphohistone H3, Proliferating Cell Nuclear Antigen, atypical PKC, Crumbs2a, c-Myc, and ß-catenin. PCNA area was measured in sections and normalized to the CMZ area, pHH3 cells were counted within the CMZ, and the distances between neighbouring PCNA cells were measured.

4) Progenitor Proliferation: We used Click-iT chemistry to label proliferating cells in the CMZ of retinal cryosections. Larvae at the indicated ages were bathed in Edu for 1,2,4,6 or 12 hours at 28°C and then processed either immediately (progenitor analysis) or after 3-14 days (radial cohort analysis). For the radial cohort analysis, we restricted our analysis to those cell cohorts where the participating sections were fully intact and cells could be counted across all three layers.