Sema6B, Sema6C, and Sema6D Expression and Function during Mammalian Retinal Development

In the vertebrate retina, the formation of neural circuits within discrete laminae is critical for the establishment of retinal visual function. Precise formation of retinal circuits requires the coordinated actions of adhesive and repulsive molecules, including repulsive transmembrane semaphorins (Sema6A, Sema5A, and Sema5B). These semaphorins signal through different Plexin A (PlexA) receptors, thereby regulating distinct aspects of retinal circuit assembly. Here, we investigate the physiological roles of three Class 6 transmembrane semaphorins (Sema6B, Sema6C, and Sema6D), previously identified as PlexA receptor ligands in non-retinal tissues, in mammalian retinal development. We performed expression analysis and also phenotypic analyses of mice that carry null mutations in each of genes encoding these proteins using a broad range of inner and outer retinal markers. We find that these Class 6 semaphorins are uniquely expressed throughout postnatal retinal development in specific domains and cell types of the developing retina. However, we do not observe defects in stereotypical lamina-specific neurite stratification of retinal neuron subtypes in Sema6B−/− or Sema6C−/−; Sema6D−/− retinas. These findings indicate these Class 6 transmembrane semaphorins are unlikely to serve as major PlexA receptor ligands for the assembly of murine retinal circuit laminar organization.


Introduction
Nervous system function relies in part on precise patterns of synaptic connectivity established during development. In the vertebrate retina, external visual information is processed by distinct subtypes of retinal neurons that elaborate and establish synaptic connections within discrete laminae: the outer plexiform layer (OPL) and sublaminae of the inner plexiform layer (IPL). Within the IPL, two parallel neural circuit pathways respond to either an increment (ON pathway) or a decrement (OFF pathway) in illumination, and these are organized within separate sublaminae to allow for segregated processing of distinct visual stimuli. However, the molecular mechanisms that govern specific neuronal subtype targeting to retinal laminae in the inner and outer retina remain poorly understood.
Accumulating evidence shows that both adhesive and repulsive molecules direct neurite targeting to laminae in the inner and outer retina in vivo. Cell adhesion molecules, including sidekicks and Dscams, disturb lamina-specific neurite arborization within the IPL in chick and mouse [1,2,3]. In addition, multiple transmembrane semaphorins, including Sema6A, Sema5A, and Sema5B, regulate neurite targeting of many cell types in the inner and outer murine retina [4,5,6]. Specifically, Sema6A signals through PlexA4 to regulate select neuronal subtype targeting to specific sublaminae within the IPL [6]. Sema6A-PlexA4 signaling also controls horizontal cell neurite arborization, neurite targeting to the OPL, and rod ribbon synapse formation in the outer retina [5]. Conversely, Sema5A and Sema5B signal through PlexA1 and PlexA3 to prevent neurites of multiple inner retinal neuron subtypes from aberrantly misprojecting into the outer retina, thereby constraining these neurites within the inner retina [4]. Taken together, these findings demonstrate that the PlexA receptors play critical roles in the formation of retinal circuits. However, it is not known whether PlexA receptor ligands previously identified in non-retinal tissues also contribute to the retinal circuit formation.
The class 6 transmembrane semaphorins Sema6B, Sema6C, and Sema6D also signal through the PlexA receptors. For example, Sema6B binds to PlexA4, inducing inhibitory responses in hippocampal and sympathetic neurons that are PlexA4dependent in vitro [7,8]. In contrast, Sema6C and Sema6D bind to PlexA1, regulating specific aspects of neuronal and cardiac development [9,10]. These findings raise the question as to whether or not Sema6B-PlexA4 signaling, or Sema6C-and/or Sema6D-PlexA1 signaling, control retinal development. Therefore, we here investigate whether Sema6B, Sema6C, and Sema6D regulate retinal laminar organization and circuit formation.

In situ Hybridization
In situ hybridization was performed on fresh frozen retina sections (20 mm thickness) as described previously [6]. The digoxigenin-labeled antisense riboprobes specific for Sema6B, Sema6C, and Sema6D used in this study were previously described [8,9].

Sema6B, Sema6C, and Sema6D mRNA Expression in the Developing Mouse Retina
To investigate whether Sema6B, Sema6C, and Sema6D regulate retinal development, we first analyzed mRNA expression of Sema6B, Sema6C, and Sema6D during postnatal retinal development by in situ hybridization (Fig. 1). We performed in situ hybridization at the postnatal ages of P7 and P14, developmental time points when inner and outer retinal circuits are established. We found that Sema6B, Sema6C, and Sema6D are all expressed at these stages of postnatal retinal development. Specifically, Sema6B is expressed most prominently in the ganglion cell layer (GCL) and inner nuclear layer (INL) at P7 and P14 (Fig. 1A, 1G). At both ages, a majority of retinal neurons in the GCL and inner INL express Sema6B (Fig. 1A, 1G), whereas Sema6B expression was either low or absent in the outer nuclear layer (ONL) at these postnatal ages. On the other hand, Sema6C was expressed broadly in the INL and sparsely in the GCL at P7 and P14 (Fig. 1C, 1I). A majority of retinal neurons in the INL strongly express Sema6C, but only a subset of neurons in the GCL appear to express Sema6C at both ages. Sema6D expression was strong in neuronal subtypes that reside in the GCL and INL at P7 and P14 (Fig. 1E, 1K). The retinal cell types that express Sema6D robustly in the GCL at these ages are retinal ganglion cells and/or displaced amacrine cells, while those expressing Sema6D strongly in the INL close to the OPL are likely horizontal cells, given their morphological features and horizontal process staining (Fig. 1E, 1K). At P7 and P14, Sema6D was also expressed in a subset of neurons in the inner INL ( Fig. 1E, 1K). We used the corresponding sense probes as references for detecting background staining and did not observe specific signals with these probes at P7 and P14 (Fig. 1B, 1D, 1F, 1H, 1J, 1L).
These expression patterns of Sema6B, Sema6C, and Sema6D during postnatal retinal development suggest a role for these semaphorins in regulating retinal development in vivo. To determine whether these class 6 semaphorins regulate retinal development in vivo, we analyzed mice that harbour null mutations in genes encoding Sema6B (Sema6B 2/2 ), or both Sema6C and Sema6D (Sema6C 2/2 ; Sema6D 2/2 ) [8,12]. For each genotype we analyzed eyes from 4 animals. We examined neurite elaboration in distinct retinal subtypes in the IPL and OPL of these mutant mice by immunohistochemistry using a range of retinal cell-type markers. Each of these markers labels neurites belonging to specific subtypes of RGCs, amacrine and bipolar cells within the IPL or photoreceptors, bipolar and horizontal cells within the OPL.
Mü ller Glia Morphology and Process Extension are Normal in Sema6B 2/2 and Sema6C 2/2 ; Sema6D 2/2 Retinas To assess whether morphology and process extension of Müller glia cells are affected in Sema6B 2/2 and Sema6C 2/2 ; Sema6D 2/2 retinas, we used antibodies directed against glutamine synthetase and vimentin to visualize cell bodies and processes of Müller glia cells. Müller glia cell bodies are found in the INL, and their cell processes span across the retina and form the outer and inner limiting membranes in wild-type retinas (Fig. 6A, 6D) [23]. We observed that the morphology of Müller glia cells in adult retinas from Sema6B 2/2 and Sema6C 2/2 ; Sema6D 2/2 mice does not apparently differ from that in wild-type retinas (Fig. 6).

Discussion
We show here that the formation of neural circuitry in the inner and outer retina of Sema6B 2/2 and Sema6C 2/2 ; Sema6D 2/2 mice is comparable to that observed in WT mice. Previous studies show that these class 6 transmembrane semaphorins signal through the PlexA receptors to regulate neural and non-neural development [7,8,9,10]. We recently found that Sema6A and several PlexA receptors play critical roles in regulating the development of retinal circuits [4,5,6], and thus we hypothesized that these class 6 transmembrane semaphorins might function through PlexA receptors to regulate retinal development.
Our results show that these class 6 transmembrane semaphorins are all expressed in specific and unique retinal cell types during postnatal development. To investigate the possibility that these transmembrane proteins regulate retinal development, we ana- lyzed Sema6B 2/2 and Sema6C 2/2 ; Sema6D 2/2 mutant retinas by immunohistochemistry using a range of markers that label distinct retinal neuron subtypes in the inner and outer retina. However, our extensive phenotypic analyses of these mutant retinas did not reveal clear deficits in stereotypical neurite stratification of distinct subtypes in either the inner or outer retina. Although we used many immunological markers that label different subtypes of retinal cells, our analysis did not include several additional retinal neuron subtypes present in the retina. We do not exclude the possibility that more specific retinal neuronal or glial subtypes, which were not visualized by the markers used here, show defects in neurite elaboration or sublaminar targeting. It is also possible that penetrance of retinal phenotypes, if any, in these mutant mice is not high enough to be revealed by the number of the mutant mice we examined (n = 4 mice/genotype analyzed). We also did not assess the cell number and cell migration of each retinal subtype, nor did we investigate mosaic patterns of cell body distribution of these subtypes in the horizontal retinal plane. Finally, our analysis did not include a detailed analysis of retinal neuronal or glial process morphology in the plane of individual laminae. Thus, our present study does not exclude the possibility of phenotypes in Sema6B 2/2 and Sema6C 2/2 ; Sema6D 2/2 retinas that affect these aspects of retinal organization. Nevertheless, given the important roles played by PlexA receptors in sublaminar targeting of multiple retinal neuron subtypes [4,5,6], our work shows that these class 6 semaphorins do not serve major functions in retinal sublaminar targeting. Since the transmembrane semaphorins Sema5A and Sema5B together constrain neurites from inner retinal neuron subtypes to the IPL [4], it is possible that class6 semaphorins are functionally redundant with respect to regulation of retinal development. It was shown that Sema6A and Sema6B function additively to regulate murine hippocampal mossy fiber projections in vivo [8], and thus Sema6B might also cooperate with Sema6A to control specific events of retinal development, such as horizontal cell OPL neurite stratification.
We expect that a variety of membrane-bound and secreted guidance cues will be investigated further to identify the complete spectrum of molecules required for the establishment of laminar organization of the vertebrate retina. Recent studies show that secreted guidance cues, including netrins and slits, direct laminaspecific photoreceptor and RGC axon targeting in the fly and zebrafish visual system [24,25]. It will be of interest to determine how membrane-bound and secreted cues cooperate to build the laminar structure of vertebrate retinal circuits, advancing our understanding of how segregated processing of visual information in the retina comes to be established during development.