Formation of retinal direction-selective circuitry initiated by starburst amacrine cell homotypic contact
- Duke University School of Medicine, United States
- Vanderbilt University School of Medicine, United States
Abstract
A common strategy by which developing neurons locate their synaptic partners is through projections to circuit-specific neuropil sublayers. Once established, sublayers serve as a substrate for selective synapse formation, but how sublayers arise during neurodevelopment remains unknown. Here, we identify the earliest events that initiate formation of the direction-selective circuit in the inner plexiform layer of mouse retina. We demonstrate that radially migrating newborn starburst amacrine cells establish homotypic contacts on arrival at the inner retina. These contacts, mediated by the cell-surface protein MEGF10, trigger neuropil innervation resulting in generation of two sublayers comprising starburst-cell dendrites. This dendritic scaffold then recruits projections from circuit partners. Abolishing MEGF10-mediated contacts profoundly delays and ultimately disrupts sublayer formation, leading to broader direction tuning and weaker direction-selectivity in retinal ganglion cells. Our findings reveal a mechanism by which differentiating neurons transition from migratory to mature morphology, and highlight this mechanism’s importance in forming circuit-specific sublayers.
https://doi.org/10.7554/eLife.34241.001eLife digest
Our experience of the world relies on circuits spanning the sense organs and the brain that process information received through our senses. These circuits are made up of many different types of nerve cells that form connections with each other while the brain is developing. For these circuits to be set up properly, nerve cells have to be selective about how they connect with each other. However, researchers know little about how exactly nerve cells form the right connections, or about which genes and proteins are involved.
One of the better understood circuits in the body is known as the ‘direction-selective circuit’. Found in the retina at the back of the eye of all backboned animals, this circuit’s task is to detect the direction that objects are moving. In the case of mice, scientists have identified all of the cells that make up the circuit, and know how they are all supposed to be connected together. This is a useful starting point for researchers to look in more detail at how nerve cells make the right connections during development to set up a working circuit.
Ray et al. looked at how the direction-selective circuit forms in the retinas of young mice by genetically engineering cells to carry fluorescent proteins, or staining them with chemicals. This allowed the cells to be examined under a microscope at different points in their development. It turns out that one type of cell, known as the ‘starburst amacrine cell’ because of its firework-like shape, coordinates the formation of the whole direction-selective circuit. First, starburst cells branch out and touch each other. Next, they build a scaffold for the circuit with their branch-like extensions. Finally, other cell types follow this scaffold to form connections and complete the circuit.
Ray et al. identified a protein called MEGF10 on the surface of starburst cells that tells the cells when they have made contact with each other. When starburst cells had MEGF10 taken away, or were prevented from contacting each other, they did not build a scaffold properly, and the circuit was less effective at detecting movement.
It is possible that cells in other brain circuits use a similar method to form connections. Understanding more about how nerve cells form circuits will help researchers to work out what goes wrong in developmental disorders that affect vision, memory and learning. This knowledge would be helpful for designing new treatments for these conditions.
https://doi.org/10.7554/eLife.34241.002Introduction
In the developing nervous system, neurons form selective synapses to generate circuits comprised of cell-type-specific connections. This selectivity is important for circuit function because it ensures connectivity between neurons specialized for particular information-processing tasks. Despite its importance, basic questions about selective synapse formation remain unanswered. For example, we do not know how cell types fated to form synapses coordinate their growth to establish contact with each other. This is a significant cell biological challenge, because the neurons that comprise a single circuit are often born at disparate times and physical locations.
In many tissues, notably the insect and vertebrate visual systems, synaptic specificity is facilitated by laminar specificity, the phenomenon whereby circuit partners project their axons and dendrites to narrow strata within a laminated neuropil (Sanes and Zipursky, 2010). The inner plexiform layer (IPL) of the vertebrate retina comprises at least 10 distinct sublayers built from the axons and dendrites of different amacrine, bipolar, and retinal ganglion cell (RGC) types (Baier, 2013). By projecting to the same IPL sublayer, circuit partners can be assured of encountering each other. The developmental events that create sublayers and guide circuit partners to converge upon them are therefore essential for establishment of retinal circuitry. At later developmental stages, when rudimentary IPL sublayers have already formed, neurons rely on molecular cues localized to those sublayers for guidance to the appropriate IPL strata (Duan et al., 2014; Matsuoka et al., 2011; Sun et al., 2013; Yamagata and Sanes, 2008; Visser et al., 2015). However, a crucial question remains unresolved: How do sublayers form in the first place? Understanding the mechanisms that initiate creation of sublayers will provide significant insight into the earliest step in circuit formation.
To learn how members of a single circuit create layers and converge upon them to achieve synapse specificity, we studied the direction-selective (DS) circuit of mouse retina (Figure 1A). This circuit reports the direction of image motion to the brain through the spiking activity of distinct DS ganglion cell (DSGC) types that are tuned to prefer stimuli moving in particular directions (Demb, 2007; Vaney et al., 2012). The DS circuit comprises a limited number of well-described cell types amenable to genetic marking and manipulation (Kay et al., 2011; Huberman et al., 2009; Duan et al., 2014): (1) DSGCs; (2) GABAergic/cholinergic interneurons called starburst amacrine cells (SACs); and (3) four subtypes of glutamatergic bipolar cells (Chen et al., 2014; Duan et al., 2014; Greene et al., 2016; Kim et al., 2014). These DS-circuit cell types project to two IPL sublayers, ON and OFF, named for the light response profiles of the neurons that project to them. ON-OFF DSGCs (ooDSGCs) send dendrites to both sublayers, while SACs and bipolar cells project to one or the other, depending on their subtype (Figure 1A). Several molecular perturbations have been described that influence ON vs. OFF laminar targeting in the mouse DS circuit (Sun et al., 2013; Duan et al., 2014), but in these cases, IPL sublayers still form in the right place; errors are limited to choosing the wrong DS sublayer. Thus, neither the establishment of the DS circuit sublayers nor their positioning in the appropriate IPL region depends on molecules that have been studied to date.
Figure 1 with 2 supplements see all
Initial formation of DS circuit IPL sublayers.
(A) Schematic of mature direction-selective (DS) circuit and its cell types, depicted in cross-section. SACs (green) and bipolar cells (blue) project to one of two IPL sublayers (OFF, ON). OFF SACs reside in inner nuclear layer (INL); ON SACs reside in ganglion cell layer (GCL). ooDSGCs (purple) send dendrites to both DS circuit sublayers. ONL, outer nuclear layer. (B) SAC IPL sublayer formation assessed in Megf10lacZ mice. All SACs are double-positive for anti-Sox2 (purple) and anti-βgal (green). Progenitors in outer neuroblast layer (ONBL) also express Sox2. SAC IPL sublayers (arrowheads) begin to appear by P0, and are fully apparent by P1. (C) Sparse labeling of neonatal SACs in ChatCre mice. Individual SACs have laminar-specific projections by P1 (arrows). tdT, tdTomato. (D,E) ooDSGCs (labeled by Hb9-GFP) project diffusely in the IPL at P1-P2, whereas SAC arbors are stratified (arrowheads). (D) retinal cross-sections. Vertical white bar denotes IPL width. E: Fluorescence intensity plots of SAC and ooDSGC dendrite staining across IPL, from representative images (P2 image in D; P1 image in Figure 1—figure supplement 2). ON and OFF strata (asterisks) are clear for SACs but not for ooDSGC dendrites. Scale bars: 25 µm.
Here, we seek to understand the earliest events leading to formation of the DS circuit IPL sublayers. Two lines of evidence suggest that SACs may take the lead in assembling this circuit. First, SACs are among the first cells to stratify the IPL: Even though other neurons innervate it contemporaneously, SACs are precocious in restricting their arbors into sublayers (Stacy and Wong, 2003; Kay and Sanes, 2013). Second, in mutant mice that entirely lack RGCs or bipolar cells, SAC IPL projections are largely normal, indicating SACs can form sublayers in the absence of their circuit partners (Moshiri et al., 2008; Green et al., 2003). Thus, we set out to test the hypothesis that SACs orchestrate assembly of the DS circuit sublayers. We find evidence supporting this hypothesis, and we identify a surprising cellular mechanism initiating SAC lamination: Rather than immediately innervating the IPL, newborn SACs first produce a transient homotypic arbor network outside the IPL. These early homotypic contacts serve as a cue promoting SAC dendrite development and circuit integration upon conclusion of their radial migration to the inner retina. When deprived of homotypic contacts, SAC IPL innervation – and consequent sublayer formation – is impaired. We identify the SAC cell-surface protein MEGF10 as the molecular mediator of IPL innervation upon homotypic contact. In the absence of MEGF10, SACs persist in growing arbors outside the IPL, delaying IPL innervation. This in turn delays formation of the DS circuit sublayers and leads to SAC sublaminar targeting errors that persist to adulthood. We further show that impaired SAC sublayer formation has consequences for laminar targeting of their circuit partners: While partnering remains intact, lamination is disrupted, leading to spatial inhomogeneity in the DS circuit network. Finally, we show that these MEGF10-dependent anatomical changes both broaden and weaken direction tuning across the population of ooDSGCs. These results demonstrate that SACs orchestrate DS circuit assembly, first by initiating sublayer formation via homotypic contact, and then by using their laminated dendrites as a scaffold that guides projections of their circuit partners.
Results
Timing of DS circuit IPL sublayer formation
To explore how the DS circuit creates its IPL sublayers, we began by determining when the sublayers first emerge in mouse. This analysis focused on SACs and ooDSGCs because bipolar cells develop later (Morgan et al., 2006). Previous estimates of layer emergence vary widely (Stacy and Wong, 2003; Sun et al., 2013) due to the lack of adequate markers to study dendrite development in neonatal SACs. We therefore assembled a suite of mouse lines and antibody markers for this purpose, enabling anatomical studies of the full SAC population as well as individual cells (Figure 1B–C; Figure 1—figure supplement 1; Figure 2—figure supplement 1). These markers revealed that SAC dendrites form two continuous well-defined laminae by P1 (Figure 1B,E). Some dendrites were stratified already at P0, even though the P0 IPL neuropil is less than one-cell diameter wide (Figure 1B; Figure 1—figure supplement 1). Further supporting this timeline, individual P1 SACs made lamina-specific projections (Figure 1C): 96% of OFF SACs in the inner nuclear layer (INL), and 99% of ON SACs in the ganglion cell layer (GCL), stratified within the expected IPL sublayer (n = 49/51 OFF; 78/79 ON; four mice). By contrast, ooDSGCs projected rudimentary and unstratified dendrites at P1 (n = 18 cells, three mice, none were stratified; Figure 1E; Figure 1—figure supplement 2; also see Peng et al., 2017). Even at P2, only 30% of ooDSGCs co-fasciculated with SAC arbors; the rest projected diffusely within the IPL (n = 23 cells, two mice; Figure 1D,E; Figure 1—figure supplement 2). These results indicate that SACs form IPL sublayers at P0-P1, and are joined later by their synaptic partners.
Early SAC projections target neighboring SAC somata
To gain insight into how SACs form their sublayers, we next investigated the cell-cell interactions that immediately precede SAC dendrite stratification. Because SACs stratify early – before any other cell type investigated to date (Figure 1; Kay and Sanes, 2013; Stacy and Wong, 2003) – they are unlikely to form strata by following pre-existing laminar cues. Instead, we hypothesized that SACs create their sublayers by engaging in homotypic interactions. To test this idea, we examined embryonic retina to determine if and when SACs establish homotypic contact. SACs exit the cell cycle at the apical retinal surface and migrate radially through the outer neuroblast layer (ONBL). They next arrive at the inner neuroblast layer (INBL), where postmitotic neurons reside (Hinds and Hinds, 1978); Figure 2A,B). Then they begin to innervate the nascent IPL, which begins to appear in some retinal regions at E16 (Figure 2A). To reveal SAC morphology throughout these steps, the early SAC marker Isl1 (Galli-Resta et al., 1997) was used to drive Cre-dependent expression a membrane-targeted GFP (mGFP) reporter (Isl1mG mice). We also examined the orientation of SAC dendrite projections using antibodies to internexin, a marker of SAC primary dendrites (Figure 2—figure supplement 1). Staining was performed at E16, when SACs at all stages of their early development could be discerned (Figure 2A–D).
Figure 2 with 3 supplements see all
Newborn SACs contact each other via a network of soma layer arbors.
(A,B) Isl1 labels SACs and RGCs in embryonic retina. A, immunostaining; B, mGFP driven by Isl1Cre (Isl1mG). Arrows, newborn SACs migrating apico-basally through ONBL. INBL SACs and RGCs predominantly reside in indicated regions. IPL neuropil (asterisks) exists in discontinuous patches at this age. NFL, nerve fiber layer containing RGC axons. Blue, nuclear counterstain. (B,C) Migrating SACs in ONBL (arrows) have multipolar morphology. They are far from other SACs and do not contact them. (D) Morphology of Sox2+Isl1+ SACs (large arrows) upon arrival at INBL. SACs contact each other outside the IPL (small arrow, connecting arbor). Their migratory morphology and distance from IPL (asterisks) indicate they have not yet innervated IPL (also see Figure 2—figure supplement 2). (E) Internexin immunostaining reveals polarization of SAC primary dendrites. SACs project toward IPL at P2. E16 INBL SACs often project tangentially within INBL, towards neighboring SAC somata (arrows). (F) Quantification of primary dendrite orientation. Left, polar histogram (raw counts) of primary dendrite angles (absolute values). 0˚ was defined as perpendicular towards IPL. Note that E16 and P1 are displayed on different scales; each plot is scaled to size of largest bin, aiding legibility in E16 plot where bin sizes are more evenly distributed. Right, fraction of dendrites oriented towards IPL, ONBL, or tangential quadrants of the polar plot (denoted by dashed lines, labels). Sample size of scored dendrites is indicated. (G) A network of arbors (arrows) connects somata of INBL SACs at E16. The network remains prominent in INL at P0 but is mostly gone by P2. Arrowhead, IPL-directed projection. (H,I) An individual P1 OFF SAC labeled by ChatmG, imaged en face to show its arbor morphology at IPL and INL levels. Full SAC population is revealed using Megf10:βgal. INL arbors (I) make selective contacts with SAC neighbors (purple): GFP+ arbor tips terminate on SAC somata (orange arrow) or SAC arbors (white arrows). Right panel (I) Higher magnification view of touching arbors. Images are Z-projections of confocal stacks encompassing each arbor’s volume (H, 2.0 µm; I, 3.5 µm). Projections are shown for illustration but all contacts were verified across stack volume – see Figure 2—figure supplement 3 for details. (J,K) Individual P1 OFF (K) and ON (L) SACs labeled by ChatmG (green) in cross-section. Purple, full SAC population. Some SACs are bi-laminar with arbors that contact neighboring somata (arrows, left panels); others project only to IPL (right panels). (L) Frequency of soma layer projections across development, determined from single ChatmG cells as in J,K. Error bars, standard error. Sample sizes, see Methods. (M) Schematic of newborn SAC morphology based on B-L. Soma-layer homotypic contacts are established upon completion of migration and are mostly eliminated by P3. Scale bars: 25 µm (A,B); 10 µm (all others).
Since mature SACs contact each other in the IPL, we expected that the onset of SAC homotypic contact would occur around the time of their earliest IPL projections. Surprisingly, however, this analysis revealed that SACs begin to contact each other within the INBL cell body layer upon the conclusion of their radial migration. Migrating SACs rarely interacted, but on arrival at the INBL, SAC arbors were observed touching the soma or primary dendrite of neighboring SACs (Figure 2A–D). The majority of INBL SACs engaged in these soma-layer contacts, such that a GFP+ arbor network connected them (Figure 2G). Analysis of primary dendrite orientation indicated that soma-layer contacts likely arose due to projections targeted within this layer: Unlike mature SACs, which exclusively project their primary dendrites toward the IPL, many E16 SACs projected tangentially through the INBL – that is, toward neighboring somata (Figure 2E,F). We even noted cases where SACs appeared to project directly towards each other (Figure 2E). These observations suggest that post-migratory SACs initiate contact with each other by generating an arbor network in the INBL cell body layer.
Many E16 SACs also innervate the nascent IPL, raising the question of whether the soma- or IPL-layer projection establishes the first homotypic contact. We concluded that soma-layer SAC contact precedes IPL innervation, for three reasons. First, soma contacts were found in retinal regions where the IPL had not yet emerged (Figure 2—figure supplement 2). Second, soma contacts were observed among cells that still showed migratory morphological features, such as apical and/or basal processes (Deans et al., 2011; Hinds and Hinds, 1978), and did not yet project into the IPL (Figure 2D; Figure 2—figure supplement 2). Third, SAC dendrite polarization in the tangential plane was highly transient: By P1, the vast majority of SAC primary dendrites were oriented toward the IPL (Figure 2E,F). These three observations suggest that INBL SACs transiently seek out homotypic soma contact before shifting to target the IPL.
We next sought to determine how long the soma-layer SAC arbor network persists. To this end, we examined SAC anatomy at early postnatal ages using Isl1mG and ChatmG (Figure 1—figure supplement 1) mice. At P0-1, although SAC arbors within the soma layers no longer express internexin (Figure 2—figure supplement 1), the arbor network remained remarkably prominent (Figure 2G). Most OFF SACs assumed a bi-laminar morphology, with one set of arbors in the IPL and another set targeting neighboring SACs in the INL (Figure 2H–J,L; Figure 2—figure supplements 2–3). INL contacts were highly SAC-selective: 88.8% of branches terminated homotypically (n = 122 arbor tips from 22 cells), significantly greater than the contact rate expected by chance (Figure 2—figure supplement 3). By P2-3, however, this dense INL network was mostly gone (Figure 2G,L; Figure 2—figure supplement 2). ON SACs also made soma layer projections between P0 and P3 that contacted neighboring SAC somata (Figure 2K,L; Figure 2—figure supplement 3). Together, these observations demonstrate that both ON and OFF SACs make transient soma-layer homotypic contacts that arise prior to IPL dendrite elaboration, and are disassembled at P2-3 after SAC sublayers have formed (Figure 2M).
Homotypic contact is required for SAC IPL innervation and dendrite lamination
SAC homotypic contacts arise at a time when they could serve as a cue for IPL innervation and sublayer formation. To test this idea, we developed a genetic strategy to prevent SACs from contacting each other in vivo. Ptf1a encodes a transcription factor required for progenitor cells to assume an amacrine fate (Fujitani et al., 2006; Nakhai et al., 2007); Figure 3—figure supplement 1). We crossed conditional Ptf1aflox mutant mice (Krah et al., 2015) to a Cre line (Six3-Cre; Furuta et al., 2000) that drives widespread recombination in central retina but spares some progenitors from Cre activity in peripheral retina (Figure 3A; Figure 3—figure supplement 1). In Six3-Cre; Ptf1aflox/flox mice (abbreviated Ptf1a-cKO), only these spared Cre– progenitors were capable of giving rise to SACs, indicating that any SACs produced in these mutants are wild-type at the Ptf1a locus (Figure 3C). Therefore, the Ptf1a-cKO mutant creates a situation where otherwise-normal SACs are present at significantly lower density than in wild-type retina (Figure 3B,C). In P1-2 mutants, some SACs were effectively segregated from their neighbors – these were termed ‘solitary’ SACs – while others had neighbors sufficiently nearby that they touched (Figure 3B–F; Figure 3—figure supplement 2).
Figure 3 with 2 supplements see all
SAC homotypic contact is required for IPL sublayer formation.
(A) Top: Schematic illustrating Six3-Cre expression pattern in retinal cross-section. Bottom: En-face view of Six3-Cre recombination in peripheral retina, revealed using GFP Cre reporter. Asterisks, Cre– regions. (B) Reduced SAC density in Ptf1a-cKO retina. SACs (labeled by Sox2 and Megf10lacZ) are completely eliminated from Ptf1a-cKO central retina; some remain in peripheral retina (boxed regions, right panels). Top, littermate control (Ptf1a+/+). (C) En-face view of SACs in peripheral retina of Ptf1a-cKO and littermate control. Green, GFP Cre reporter. Control SACs were either Cre+ or Cre–. Mutant SACs were Cre– (arrows), indicating that they derive only from cell lineages that maintain Ptf1a function. (D–F) SAC IPL laminar targeting in Ptf1a-cKO (E,F) and littermate control (D). Ptf1a-cKO SACs close enough to touch (E) form IPL strata (blue arrowheads), similar to control SACs (D). Solitary SACs (F) are not polarized toward IPL; they have extensive INL-directed arbors (white arrows) and rudimentary IPL-directed arbors (orange arrows). Some solitary SACs entirely fail to innervate IPL (F, left cell) and resemble migrating E16 SACs (Figure 2B,C); others innervate IPL with minimally branched, non-stratified arbors (F, right cell). (G) Quantification of SAC dendrite phenotypes at P1-2. Left, frequency of soma layer innervation. *p=0.0350; **p=0.0081; ns, p=0.7516. Center, frequency of IPL innervation failure (e.g. F, left). ***p=4.0×10−7; ns, p=0.3723. Right, frequency of cells that send arbors into IPL but fail to stratify (e.g. F, right). *p=0.0110; ***p<1.0×10−7. Dots, individual animals. Error bars, S.E.M. p-values, Tukey’s post-hoc test. Sample sizes, see Methods. (H) Summary of Ptf1a-cKO SAC phenotype. Touching SACs are similar to wild-type SACs (Figure 2M); they are able to innervate the IPL and form sublayers. Solitary SACs remain multipolar, similar to migrating SACs, and fail to innervate the IPL. Soma layer arbors are more elaborate than in wild-type or touching SACs. Scale bars: 25 µm (A,C); 200 µm (B, left), 50 µm (B, right), 10 µm (D–F).
Comparing solitary to touching SACs in Ptf1a-cKO retinas revealed a role for homotypic contacts in promoting IPL innervation and sublayer formation. At P1-2, touching SACs projected normally to the IPL, similar to SACs from Ptf1a+ littermates (Figure 3D,E,G). This suggests that any changes in retinal cell type composition caused by loss of Ptf1a (Figure 3—figure supplement 1) are not by themselves sufficient to perturb SAC sublayer formation. By contrast, solitary SACs largely failed to innervate the IPL (Figure 3F,G). This was not caused by abnormal migration: Solitary SACs were properly positioned at the IPL border, but sent only rudimentary arbors into it (Figure 3F; Figure 3—figure supplement 2). Solitary SACs were also more likely to project processes into the soma layers (Figure 3G), and when they did so, the projections were typically more elaborate than those observed in wild-type retina (Figure 3D,F; Figure 3—figure supplement 2). Thus, solitary SACs overgrew arbors directed toward neighboring somata instead of growing IPL dendrites. Both types of projection errors were also seen at P15, indicating that early errors persist to retinal maturity (Figure 3—figure supplement 2). Misprojecting SACs were still closely apposed to numerous other amacrine cells, and their arbors were intermingled in the IPL, strongly suggesting that generic amacrine interactions are not sufficient to ensure normal dendrite targeting (Figure 3—figure supplement 2). Instead, homotypic interactions are specifically required for IPL innervation and sublayer formation.
Requirement for MEGF10 in SAC IPL innervation and sublayer formation
To understand how SACs initiate IPL innervation upon homotypic contact, we next sought to identify the molecular cues that SACs use to recognize that contact has occurred. The cell-surface protein MEGF10 (Figure 4A) is a strong candidate to mediate homotypic recognition in this context, for four reasons. First, it is selectively expressed by SACs during the perinatal period (Figure 1B; Figure 1—figure supplement 1). Second, the onset of its expression coincides with onset of SAC homotypic contact at the conclusion of radial migration (Figure 4B). Third, MEGF10 protein is present on soma-layer SAC arbors, making it available to transduce signals arising on these arbors (Figure 4C). Finally, MEGF10 mediates SAC-SAC interactions in a separate context – during formation of the orderly ‘mosaic’ among SAC cell bodies across the retina (Kay et al., 2012). Thus, we tested whether MEGF10 also mediates SAC-SAC recognition to initiate IPL innervation. If so, SACs from mice lacking Megf10 gene function should have phenotypes similar to solitary Ptf1a-cKO SACs – that is, reduced IPL innervation and increased arborization in cell body layers.
Figure 4
MEGF10 is expressed by SACs during early homotypic contact.
(A) Schematic of MEGF10 protein. TM, transmembrane domain. (B) Left, MEGF10 immunostaining at E16 reveals onset of protein expression at conclusion of radial migration. INBL SACs express MEGF10, but SACs migrating through ONBL do not. Arrow, INBL SAC with migratory morphology suggesting it is newly arrived. Right: Schematic illustrating timing of Megf10 expression onset in SACs (also see Kay et al., 2012). (C) SAC arbors in the INL (arrows) express MEGF10 protein. IPL dendrites are also labeled (arrowheads). Scale bars: 25 µm (B); 10 µm (C).
To test this prediction, we examined SAC anatomy in Megf10 null mutants (Kay et al., 2012) and littermate controls at P0-1, when sublayers are first forming. We found a striking effect on sublayer formation: Both ON and OFF strata were absent or severely disrupted in mutants (Figure 5A). The cause of sublayer absence was investigated using pan-SAC labeling (Figure 5A,B) and single-cell analysis (Figure 5C; Figure 6D). These studies revealed a severe deficit in IPL dendrite arborization: Most Megf10–/– SACs made only rudimentary, unstratified IPL projections at P0-1 (n = 1/15 OFF SACs were stratified). Other amacrine cell types showed normal dendritic morphology in Megf10 mutants (Figure 5—figure supplement 1), indicating that the phenotype was specific to SACs. Loss of IPL innervation was not due to aberrant SAC radial migration, because, at P0, mutant SACs had reached the inner retina in normal numbers (wild-type, 2600 ± 287 SACs/mm2; mutant, 3153 ± 145 SACs/mm2; p=0.144, 2-tailed t-test; n = 3 each group), and were positioned adjacent to the IPL, similar to littermate controls (Figure 5A). Furthermore, most mutant SACs sent at least some arbors into the IPL at P0-1 (Figure 5A,C; Figure 6D), suggesting that they migrated to a location from which IPL innervation was feasible. However, the mutant SAC arbors that reached the IPL appeared undifferentiated, with a lack of space-filling branches (Figure 5A,C). As a result, not only did their arbors enclose a significantly smaller IPL territory, but they also failed to sample as much of their enclosed territory as control SACs (Figure 5C; also compare to control cell in Figure 2H). By P3 some ON SAC IPL innervation was evident, but OFF SAC arbors remained largely confined to the soma layer; those that did reach the IPL remained undifferentiated (Figure 5B; Figure 6A,D). These observations indicate that deletion of MEGF10 causes an IPL innervation phenotype strongly reminiscent of Ptf1a-cKO solitary SACs: Both manipulations profoundly impair SAC dendrite arborization within the IPL, preventing timely sublayer formation.
Figure 5 with 1 supplement see all
Megf10 is required for initial formation of SAC IPL sublayers.
(A) SAC sublayers are absent from P0-1 Megf10 mutant IPL. Antibodies to Sox2 and βgal reveal SACs in retinal cross-sections. Littermate control, Megf10lacZ/+. Vertical white lines denote IPL location. Arrowheads, SAC IPL strata. Arrows, exuberant arbor growth in mutant INL and GCL. Note that mutant somata abut the IPL at P0, indicating their radial migration was similar to controls. By P1 OFF somata have moved apically. (B) At P3, SAC IPL sublayers remain disrupted in Megf10 mutants. Mutant OFF SACs mostly fail to ramify arbors in IPL; instead they arborize in INL (also see Figure 6). INL projections are absent from controls. Some ON SACs are stratified in mutants (arrowhead) but have not yet formed a continuous restricted sublayer as is seen in controls. (C) Individual SACs show IPL innervation deficits in Megf10 mutants. En-face whole-mount view of single P0 OFF SAC IPL arbors. Mutant SAC IPL dendrites appear undifferentiated, with less branching (arrow). Their arbors cover smaller retinal territories than SACs from wild-type (wt) littermate controls (quantified at right, mean ± s. e. m.). Images are Z-projections of slices encompassing full IPL arbor volume. ***p(on)<1.0×10−7, p(off)=9.38×10−5; one-way ANOVA/Tukey’s post-hoc test. Sample size, see Methods. Scale bars: 25 µm (A,B); 10 µm (C).
Figure 6 with 1 supplement see all
Exuberant soma-layer innervation by Megf10 mutant SACs.
(A,B) Individual SACs seen in cross-section (A) and whole-mount en face view (B) of ChatmG tissue at indicated ages. Megf10 mutant SACs have more extensive branches in INL than littermate Megf10+/– SACs. Images in B are Z-projections of confocal slices encompassing each cell’s INL arbors. (C) En face images of βgal staining to show complete SAC INL network at P1. INL projections (arrows) are present at this age in controls (top); however, they are much more prominent in mutants (bottom), consistent with single-cell anatomy (B). Images are Z-projections of confocal slices encompassing volume of 2.0 µm (bottom) or 2.4 µm (top). (D) Higher-magnification view of individual OFF SACs labeled as in A. Genotypes and ages are indicated. Arrows, arbors in INL. Dashed lines denote borders of IPL. Mutant IPL projections (arrowheads) fail to arborize or stratify. See Figure 6—figure supplement 1 for ON SACs. (E) Frequency of soma layer projections across development in mutants (–/–) and littermate controls (+/–), determined from single ChatmG cells as in D. Wild-type (WT) data replotted from Figure 2L to show that +/– controls resemble WT. Error bars, standard error. Sample size, see Materials and methods. (F) Internexin immunoreactivity reveals orientation of SAC primary dendrites (arrows) at P1. Right: Example of mutant SACs projecting primary dendrites in tangential plane, within soma layer. Control primary dendrites were almost exclusively oriented towards IPL (left). (G) Summary of Megf10–/– phenotype. After initial contact at E16, mutant SACs do not immediately innervate the IPL, instead overgrowing arbors in cell body layers. This leads to delayed sublayer formation and persistent soma-layer projections at P3. Scale bars: 25 µm (A,C); 10 µm (B,D,F).
In contrast to their underinnervation of the IPL, Megf10 mutant SACs arborized exuberantly in the soma layers (Figure 6A). Both ON and OFF SACs were affected (Figure 6D,E; Figure 6—figure supplement 1), but the OFF SAC phenotype was particularly striking: Starting at P1, the mutant INL network became much more elaborate than the control network of any age (Figure 6A,C). INL arbor density increased in mutants from P0 to P1 and remained high at P3; by contrast, control SACs largely eliminated their INL projections over the same period (Figure 5A,B; Figure 6A,E). To understand how mutant SACs generate a denser and more persistent soma-layer network, we assessed single SAC morphology (Figure 6A,D). From this analysis, we determined that one reason for the denser mutant network, particularly at P2-3, was that a larger number of mutant cells projected to the soma layers (Figure 6E). However, this reason was not sufficient to explain the denser mutant INL network at P1 (Figure 6C), because at that age the number of cells projecting to the INL was similar in mutants and littermate controls (Figure 6E). Therefore, to account for this increase in INL arbor density, we surmised that individual mutant SAC must, on average, overinnervate INL. Supporting this conclusion, we found that mutant SACs frequently had more extensive INL arbors than littermate control SACs (Figure 6B). Further, mutant SACs continued to grow primary dendrites tangentially within soma layers at P1, when the vast majority of control SACs only targeted the IPL (Figure 6F; also see Figure 2F). These observations indicate that mutant SACs continue to expand their soma layer arbor network at P1. Thus, as with solitary Ptf1a-cKO SACs, soma layer projections were both more frequent and more exuberant for Megf10–/– SACs.
Together, these data suggest that MEGF10 governs a developmental transition from soma-layer to IPL-layer dendrite growth (Figure 6G): Whereas control SACs have only a brief period of soma-layer growth, switching to IPL ramification around P0, Megf10 mutant SACs do not make this transition and instead persist in soma-layer innervation. As a result of this failed transition, many individual mutant SACs ramify extensively in the INL but underinnervate the IPL, causing the dendrite targeting phenotypes that were observed at the population level (Figure 6A–C). We conclude that, because MEGF10 regulates IPL innervation in this way, MEGF10 is required for initial formation of SAC IPL sublayers.
SAC dendrite targeting requires transcellular MEGF10 signaling
Given the similar phenotypes of Megf10 mutant and solitary Ptf1a-cKO SACs, we hypothesized that MEGF10 is the molecular cue that triggers IPL innervation upon SAC-SAC contact. A key prediction of this model is that SACs should require MEGF10 signals from their neighbors to target their dendrites properly. To test this prediction, we generated a conditional Megf10flox allele and used it to create a situation where Megf10+ SACs were surrounded by Megf10– mutant cells. This was accomplished via the same Six3-Cre strategy that we employed in our Ptf1a-cKO studies (Figure 3A–C). In central retina of Six3-Cre; Megf10flox/lacZ (Six3-Megf10-cKO) animals, the vast majority of cells expressed a Cre-dependent GFP reporter, indicating that they lacked Megf10 function (Figure 7A). Accordingly, SACs projected exuberantly to the INL and sublayer formation was disrupted, as in null mutants (Figure 7B; Figure 7—figure supplement 1).
Figure 7 with 2 supplements see all
Megf10 mediates transcellular SAC signals for dendrite development.
(A,B) En-face images of INL in Six3-Megf10-cKO retinas stained for GFP Cre reporter (A) and βgal SAC marker (B). Reporter expression indicates loss of MEGF10 (see Figure 7—figure supplement 1). In central retina (top row), most SAC are mutant, and project extensive INL dendrites (B, arrows; compare to C). In peripheral retina (bottom row), some cells escape Cre (asterisks) and retain MEGF10 but still make ectopic INL projections. Purple, Cre reporter; green, βgal. (C) Littermate control retina imaged as in B; SACs rarely project INL dendrites at P2. (D) Quantification of P2 INL projection phenotypes illustrated in A-C. Six3-Megf10-cKO (cKO) SACs that escape Cre (M10+) make projection errors at similar rate as surrounding mutant cells from the same tissue (M10–). (E–G) Chat-Megf10-cKO phenotype. Morphology of single SACs, revealed by ChatmG in cross-sections. Anti-MEGF10 (M10) distinguished two classes of cKO SACs (orange arrows): Those that express MEGF10 (F) are anatomically similar to littermate control SACs (E). Those lacking MEGF10 (G) arborize extensively in INL (yellow arrowheads) but minimally in IPL (blue arrowheads). Vertical line, IPL. (H) SAC soma-layer projection frequency at P3. Sparse M10 deletion (blue, –) phenocopied germline null (red). Chat-Megf10-cKO cells that retained M10 (blue, +) resembled controls (flox/+). (I) Schematic of MEGF10 proteins used for co-immunoprecipitation (IP). Intracellular domain was deleted (ΔICD) and replaced with epitope tags (Flag or GFP). Ex, extracellular; TM, transmembrane. (J) Co-IP from lysates of HEK 293 T cells transfected with indicated constructs (I). Western blot with antibodies to GFP (green) and Flag (red). IP with anti-GFP, but not rabbit IgG control, pulled down both MEGF10-ΔICD constructs (2nd lane from right, orange text). IP with anti-Flag gave similar result (Figure 7—figure supplement 2). GFP alone did not co-IP with M10-Flag. Ladder molecular weights (kDa) at left. Full blots in Figure 7—figure supplement 2. (K) Model of MEGF10 function in early SAC dendrite development. Left, soma-layer contact between neighboring SACs initiates MEGF10-mediated signaling in each cell. This signal inhibits soma-layer dendrite growth and promotes arborization in IPL (see green cell, right panel). Purple, phenotype of SACs that fail to receive MEGF10 signals, either because neighbors do not have MEGF10, or because the cell itself lacks MEGF10 as a receptor. These SACs project exuberant soma-layer arbors and fail to ramify dendrites in IPL. Error bars, 95% confidence interval. Sample sizes, see Methods. Scale bars: 25 µm (A), 10 µm (B–G).
In peripheral retina, some SACs escaped Cre activity, leading to absence of the GFP reporter and continued MEGF10 protein expression (Figure 7A,B; Figure 7—figure supplement 1). Our model predicts that these cells should have mutant dendrite phenotypes despite retaining MEGF10. To test this prediction, we imaged βgal-stained OFF SACs from Six3-Megf10-cKO and littermate control mice at P2. This age was chosen because wild-type and null mutant mice showed a large difference in SAC INL projection frequency (Figure 6E). In littermate controls, we found that βgal+ SACs rarely projected to the INL (Figure 7C,D); therefore, they behaved like control SACs from earlier experiments (Figure 6E). By contrast, Megf10+ SACs surrounded by mutant SACs in Six3-Megf10-cKO retina showed a high rate of INL projections, nearly identical to their Megf10– neighbors (Figure 7B,D; Figure 7—figure supplement 1). Thus, when Megf10+ SACs are deprived of MEGF10 signal from adjacent SACs, they make exuberant soma-layer projections. This finding implicates MEGF10 as a transcellular signal that controls SAC dendrite targeting (Figure 7K).
Next, we investigated how SACs receive this MEGF10 signal from their neighbors. Given that MEGF10 can function as a receptor in other contexts (Chung et al., 2013; Kay et al., 2012), we speculated that MEGF10 might act as its own receptor. In support of this idea, co-immunoprecipitation experiments using intracellularly truncated Megf10 constructs showed that MEGF10 can interact with itself through its extracellular domain (Figure 7I,J; Figure 7—figure supplement 2). Thus, MEGF10 appears biochemically capable of acting as both ligand and receptor.
If MEGF10 is indeed a receptor in this context, SACs should require it to detect contact with MEGF10-expressing homotypic neighbors. To test this prediction, we asked whether removal of Megf10 from a single SAC, during the period of soma-layer homotypic contact, would impair its IPL innervation despite normal MEGF10 expression by surrounding cells. We used ChatCre to achieve sparse recombination in SACs of neonatal mice, as in the anatomy experiments described above (Figure 2H–K; Figure 6D). In Chat-Megf10-cKO animals, MEGF10 immunostaining was used to identify SACs that lost MEGF10 protein prior to P3 – that is, during the period when soma-layer arbors are present (Figure 7F,G). MEGF10– cells constituted a small minority of SACs at P3, meaning that they were generally surrounded by MEGF10+ neighbors (Figure 7—figure supplement 1). In this context, MEGF10– SACs produced more exuberant soma-layer arbors than neighboring MEGF10+ cells, while sending only minimal arbors into the IPL (Figure 7E–H). Thus, single MEGF10– SACs had phenotypes similar to SACs from mice entirely lacking Megf10 (Figure 7G,H; compare to Figure 6D). By contrast, adjacent MEGF10+ cells in the same Chat-Megf10-cKO retinas were indistinguishable from littermate control SACs (Figure 7E,F,H). Therefore, when Megf10 is lost during dendro-somatic contact (but not after; see below), SACs make projection errors typical of neurons deprived of homotypic interactions, and they do so even if their neighbors express MEGF10 and are developing normally. Together, these experiments support the conclusion that MEGF10 is a receptor through which SACs detect each other to terminate soma-layer growth and initiate IPL innervation (Figure 7K).
SAC errors persist to adulthood in Megf10 mutants
We next asked whether neonatal MEGF10-mediated interactions influence the anatomy of SAC IPL sublayers at maturity. We found that SAC sublayers eventually formed (by P5; Figure 8H), and were present in the mature Megf10–/– retina, but they were marred by numerous errors. Sporadically, and at apparently arbitrary retinal locations, two kinds of local laminar disruptions were apparent. First, there were discontinuities in the ON and OFF strata, such that mutant SACs did not completely innervate their sublaminae (Figure 8A–C). These discontinuities diminished retinal coverage within each mutant sublayer by ~15% (OFF decrease, 15.0 ± 0.9%; ON decrease, 13.7 ± 4.0%; mean ±SD; n = 9 fields of view/2 mice per genotype). Innervation gaps were not observed for other amacrine cells, indicating that SACs were selectively affected (Figure 8—figure supplement 1). Examination of single SACs revealed that while dendritic patterning substantially recovered between P1 and adulthood, SAC arbor territories remained significantly smaller in mutants (Figure 8D). These phenotypes suggest that mutant SACs never fully made up for their initial IPL innervation deficit, thereby contributing to gaps in the dendritic plexus.
Figure 8 with 1 supplement see all
SAC IPL errors persist to maturity in Megf10 mutants.
(A,B) SAC IPL phenotype in mature (two-week-old) retina, cross-section view. Blue, soma counterstain. Control IPL has two continuous SAC dendrite bands (blue arrowheads). Mutant IPL has sporadic ectopic SAC arbors (yellow arrowheads) or laminar gaps (white arrows). (C) En-face views of SAC dendrites, stained with anti-ChAT, in adult retinal whole-mounts. The same fields of view are shown at two different Z-stack planes, corresponding to OFF and ON SAC sublayers. SAC dendrite plexus is uniform in littermate controls, but has holes (arrows) and large gaps (asterisks) in mutants. Note that errors are not spatially correlated between OFF and ON sublayers. Images are Z-projections of 5 slices encompassing 2.0 µm in Z. (D) Single SAC labeling in adult (3 month old) mice, via ChatCre-dependent viral fluorescent protein expression. Megf10–/– SACs have relatively normal morphology but are significantly smaller than wild-type (WT) control cells (***p=4.6×10−6, two-tailed t-test). Sample size, see Materials and methods. (E) En-face images reveal extent of ectopic mutant SAC dendrite network. Same fields of view as C, but at different Z-planes depicting OFF IPL (at IPL-INL border). In mutant but not control, SAC dendrite fascicles (arrows) are evident at this IPL level. Images are Z-projections of 3 (left) or 2 (right) slices spaced 0.4 µm in Z. (F,G) Dendrite targeting of individual OFF SACs assessed across development. Fraction of mutant SACs projecting into ectopic network does not change over time (F). However, location of ectopic arbors shifts from INL to IPL (G), starting at P5, when cells projected to either or both locations (G,H). In F, P0-3 data are replotted from Figure 6E, with both control groups combined. Sample sizes, see Materials and methods. (H) Example P5 SACs from Megf10 mutant and littermate control. A mutant SAC projects to three different locations: (1) correct IPL sublayer (blue arrowhead); (2) ectopic IPL sublayer (yellow arrowhead); (3) ectopic INL arbor aggregate (orange arrowhead). ON SACs also make ectopic IPL projections (white arrowhead). Control cells are monostratified in IPL (left). Note that SAC sublayers have formed by P5 in mutants. Error bars, 95% confidence intervals. Scale bars: 50 µm (C–E); 25 µm (A,B,H).
The second type of SAC error in mature Megf10–/– IPL was dendrite mistargeting to ectopic IPL strata (Figure 8A,B,E). Both ON and OFF SACs were affected; in each case, ectopic arbors were mostly found in IPL regions inappropriately close to the soma layers (Figure 8A,B). En-face images of mutant IPL revealed that ectopic OFF arbors formed a patchy but extensive fascicle network connecting many of the cells (Figure 8E,F; 78.5 ± 3.5% of SACs participated in the network, mean ±95% CI). This IPL network was morphologically similar to the ectopic INL network observed in mutants at earlier ages (Figure 6C), raising the possibility that the early network gives rise to the adult network by shifting location from the INL to the IPL. Supporting this view, we found that a soma layer-to-IPL transition occurs at P5, when mutant SACs began projecting to ectopic IPL locations in addition to the soma layers (Figure 8G,H; Figure 8—figure supplement 1). This transition occurred without a significant change in the number of mutant SACs projecting into the ectopic network (Figure 8F; Figure 8—figure supplement 1), suggesting that the same cells continued to participate in the network but simply altered their anatomy to target the IPL. Thus, early exuberant soma-layer projections appear to give rise to adult IPL ectopias, starting between P3 and P5.
Together, these two adult mutant phenotypes demonstrate that DS circuit sublayer formation is delayed and imperfect in the absence of MEGF10. While other mechanisms appear to partially compensate for MEGF10 in generating the sublayers, such mechanisms are not sufficient to prevent persistence of innervation gaps and laminar targeting errors. Thus, MEGF10 is essential for normal formation of the mature SAC IPL projection.
Next, we sought to directly test the idea that MEGF10 is required early – at the time of initial SAC homotypic contact – to ensure normal SAC IPL lamination at maturity. To this end, we used Megf10flox mice to delete MEGF10 at different times. Deletion prior to the onset of homotypic contact, using the Six3-Cre line, fully phenocopied Megf10–/– adult IPL errors (Figure 9A), suggesting a requirement for MEGF10 at the time of contact. To remove MEGF10 from SACs that had already established homotypic contact, we used ChatCre. In this line, the number of SACs expressing Cre gradually increases over the first postnatal days to encompass the full SAC population (Xu et al., 2016). Therefore, Chat-Megf10-cKO mice can be used both for early, sparse MEGF10 deletion (Figure 7F–H) and for later, broad MEGF10 deletion. MEGF10 immunostaining revealed that this late, broad deletion occurs between P3 and P5 (Figure 7—figure supplement 1), such that MEGF10 expression is largely preserved during the period when homotypic soma-layer contacts exist (Figure 2L), but is eliminated shortly thereafter. In this ChatCre-mediated deletion regime, SAC laminar targeting and gap errors were exceedingly rare (Figure 9A). These experiments therefore define a time window for MEGF10 function (Figure 9C): Adult IPL targeting phenotypes require absence of MEGF10 during the soma-layer projection phase of SAC development – that is, prior to P3. Any additional activity of MEGF10 after P3 is dispensable for the adult IPL phenotype. These findings strongly support a model whereby the functions of MEGF10 during early homotypic contact – that is, promoting IPL innervation and terminating soma-layer arbor growth – are necessary for development of normal SAC IPL innervation at maturity.
Figure 9 with 1 supplement see all
Temporal requirements for MEGF10 in SAC IPL stratification and mosaic spacing.
(A) SAC IPL errors (yellow arrowheads) induced by early deletion of Megf10 in Six3-Megf10-cKO mice, but not deletion between P3-5 in Chat-Megf10-cKO mice (see Figure 7—figure supplement 1 for MEGF10 expression in these strains). Blue arrowheads, SAC sublayers. (B) Mosaic spacing phenotype measured at P17 using Voronoi domain regularity index (top graph). Dashed line, index for simulated random SAC arrays. In both Six3 and Chat conditional mutants, SAC positioning is less regular than in controls (ChatCre; Megf10flox/+). Megf10–/– and simulation data from Kay et al. (2012). ns, p=0.6438; **p=0.0023; ***p=2.1×10−6; ****p<1.0×10−6 (one-way ANOVA/Tukey’s post-hoc test). Bottom graph, regularity effects are not due to changes in SAC cell density across genotypes. One-way ANOVA, F(3, 15)=0.6063; p=0.6210. (C) Summary of results from all conditional Megf10 mutant studies (A, B; Figure 7). Loss of MEGF10 while soma-layer arbors are present (i.e. prior to P3) disrupts both SAC mosaic patterning and IPL laminar targeting; this conclusion is based on phenotypes of: (1) germline mutant; (2) Six3-Megf10-cKO mutant; (3) occasional cells in Chat-Megf10-cKO mice that lost MEGF10 prior to P3 (Figure 7F–H). When MEGF10 is lost after P3, as is the case for most Chat-Megf10-cKO SACs (Figure 7—figure supplement 1), only mosaic is disrupted (A,B). Thus, MEGF10 acts at distinct, albeit partially overlapping times, to influence these two distinct SAC developmental events. Error bars, S.E.M. Scale bar, 25 µm.
Mosaic spacing errors do not account for SAC IPL phenotype in Megf10 mutants
In addition to laminar targeting errors, Megf10 mutants also show disruptions in the mosaic spacing of SAC cell bodies across the retina: Instead of a regular, uniform distribution, mutant SAC positioning is random (Kay et al., 2012). We considered the possibility that SAC IPL errors might arise due to MEGF10 effects on soma spacing. Two lines of evidence suggest that this is not the case. First, the two phenotypes were not well correlated at the individual SAC level: Regardless of the severity of their mosaic spacing defects, SACs made IPL targeting errors at a constant rate (Figure 9—figure supplement 1). This finding suggests that disturbed cell positioning does not influence the probability of making an IPL error. Second, using our Megf10flox allele, we were able to dissociate the IPL and mosaic phenotypes: Deletion of MEGF10 after P3 in Chat-Megf10-cKO mice caused mosaic patterning deficits, but IPL projections were largely normal (Figure 9A,B). This finding demonstrates that IPL laminar perturbations are not an inevitable consequence of altered soma positioning. Altogether, these experiments support the notion that altered SAC position makes at best a minor contribution to IPL phenotypes; instead, delayed IPL innervation and exuberant soma-layer arborization are likely the major sources of perturbed SAC projections at maturity.
SAC IPL errors induce laminar targeting errors by their DS circuit partners
Next, we asked whether MEGF10, and its effects on SAC sublayer formation, are important for assembly of the broader DS circuit. To this end, we tested the impact of SAC IPL stratification errors on laminar targeting by their circuit partners. First, we examined ooDSGC IPL projections using the Hblx9-GFP (referred to as Hb9-GFP; Figure 10) and Drd4-GFP (Figure 10—figure supplement 1) transgenic lines, which label ooDSGC subtypes with different preferred directions (Trenholm et al., 2011; Huberman et al., 2009). In littermate control mice (n = 9), ooDSGC dendrites were tightly and selectively associated with SAC arbors, as shown previously (Vaney and Pow, 2000). This association was maintained in Megf10 mutants: Both normal and ectopic SAC IPL arbors reliably recruited ectopic ooDSGC projections (Figure 10A–C; Figure 10—figure supplement 1; n = 240 ectopias from five mutants,>97% contained ooDSGC arbors). Further, when SAC gaps were present in the mutant IPL, ooDSGC dendrites typically grew around the gap edges and failed to enter them (Figure 10D; Figure 10—figure supplement 1; n = 325 gaps from five mutants, >95% devoid of ooDSGC arbors). Thus, SACs provide both permissive cues required for ooDSGC IPL innervation, and also attractive cues sufficient to recruit ooDSGCs to the wrong IPL sublayer.
Figure 10 with 1 supplement see all
SACs guide IPL sublayer choice by ooDSGCs.
(A,B) SACs (ChAT, purple) and ooDSGCs (Hb9-GFP, green) labeled in cross-sections. Right panels, fluorescence profile plots across IPL showing position of ON and OFF sublayers. In both controls (A) and Megf10 mutants (B) ooDSGC dendrites strictly co-localize with SAC arbors. Arrowhead (B), ectopic sublayer. (C) En-face view of OFF IPL near INL border. Same fields of view as Figure 8E. In Megf10 mutants (bottom), the ectopic SAC network is extensively innervated by ooDSGC dendrites (arrowheads). Control retina (top) lacks DS circuit arborization at this IPL level. (D) En-face view of ON (top) and OFF (bottom) SAC IPL sublayers. In Megf10 mutants, ooDSGC dendrites (green) fail to enter IPL regions (asterisks) that are not innervated by SACs (purple). Images in C,D are Z projections of confocal slices encompassing ≥1.2 µm (C) or 2–4 µm (D). All scale bars: 25 µm. Also see Figure 10—figure supplement 1 for phenotype of Drd4-GFP ooDSGCs.
Next we determined the impact of altered SAC lamination on the axons of bipolar cells that participate in the DS circuit (Figure 11A). We examined the four cell types (BC2, BC3a, BC5, and BC7) that make extensive monosynaptic connections with SACs and ooDSGCs (Duan et al., 2014; Ding et al., 2016; GreeneGreene et al., 2016; KimKim et al., 2014; Chen et al., 2014). Bipolar axons were marked with type-specific antibodies and mouse lines reported previously (Wässle et al., 2009; Duan et al., 2014), as well as a novel transgenic marker of BC5 (Gjd2-GFP; Figure 11—figure supplement 1). In wild-type retina, DS-circuit bipolar cells arborized in close contact with SAC dendrites; however, unlike ooDSGCs, they remained adjacent to SACs rather than overlapping them (Figure 11A–D; Figure 11—figure supplement 1). This arrangement was preserved in Megf10 mutants: Axons of all four bipolar cell types were recruited to ectopic IPL locations by mistargeted SAC arbors, where they stratified adjacent to SACs (Figure 11B–D,F; Figure 11—figure supplement 1). For example, BC5 and BC7 terminals always sandwiched SAC arbors, regardless of their IPL location – even when doing so required formation of a supernumerary BC axon field between the normal and ectopic SAC sublayers (Figure 11C,D). To quantify the mistargeting effect, we measured the position of BC5 and BC7 terminals adjacent to ON SAC ectopias. Their arbors were pushed farther apart by SAC arbor clumps (Figure 11C–E), which shifted BC7 terminals significantly toward the GCL by ~4 µm (69 ± 0.8% of IPL depth in control regions to 74 ± 1.9% in affected regions; mean ± S.E.M.; n = 21 control, n = 6 affected; 2-tailed t-test, p=0.0024). These observations indicate that DS-circuit bipolar cells, like ooDSGCs, respond to SAC attractive cues. However, in contrast to ooDSGCs, bipolar cell projections were minimally affected by SAC IPL gaps. While BC5 and BC7 terminals were slightly mispositioned in the absence of SAC arbors – they were closer together – innervation of gap regions was otherwise normal (Figure 11C–F). Thus, DS-circuit bipolar axons either do not require SAC-derived signals for IPL innervation, or the relevant signals are capable of acting over larger distances than the typical SAC IPL gap size (35–45 µm maximum diameter). Altogether, these analyses of DS circuit anatomy in Megf10 mutants support the notion that early-stratifying SACs form a scaffold that directs IPL laminar targeting of their circuit partners using multiple guidance strategies.
Figure 11 with 1 supplement see all
SACs guide IPL sublaminar targeting by DS circuit bipolar cells.
(A) Illustration depicting anatomy of four bipolar cell types known to participate in DS circuit. BC3a, BC5, and BC7 arborize axons in narrow sublayers directly adjacent to SAC strata. BC2 projects more broadly through OFF IPL, overlapping with SAC sublayer (see Figure 11—figure supplement 1). (B) BC3a IPL projections assessed with anti-HCN4. Blue arrowhead, normal BC3a stratum. Yellow arrowhead, ectopic stratum containing SAC dendrites (purple) and BC3a axons (green). (C,D) BC5 and BC7 IPL projections (blue arrowheads), labeled in Kcng4mG mice. C, images; D, representative fluorescence plots of Kcng4mG (green) and ChAT (purple) across IPL. In littermate controls, or normal regions of mutant IPL (C, arrow), BC5 and BC7 arborize in sublayers immediately adjacent to ON SAC layer, but do not enter it. In Megf10 mutants, ectopic SAC arbors displace BC5+7 terminals to new IPL locations, where they remain adjacent to SACs but non-overlapping (C, yellow arrowheads; D, center plots). Asterisk (D) ectopic BC arbors between normal and ectopic SAC strata. BC5/7 arbors can innervate gaps in the SAC stratum (C, center panel); in these cases, their terminals are abnormally close together (C, white arrowhead; D, right plot). Vertical bars in D: distance between BC5/7 strata. (E) Quantification of BC5-BC7 distance in normal IPL and in presence of SAC innervation gaps or ectopic arbor clumps. *p=0.0219; **p=0.0012; ns, p=0.3965 (Tukey’s post-hoc test). Sample sizes, see Methods. Error bars, S.E.M. (F) Summary of Megf10–/– bipolar cell phenotypes. BC3a, BC5, and BC7 are illustrated here (see Figure 11—figure supplement 1 for BC2). Each of these cell types can either make errors (left cell in each pair) or project normally (right cell). All three cell types show recruitment to ectopic IPL locations. BC5 and BC7 terminals innervating SAC gaps colonize the sublayer normally occupied by SACs. Error bars, 25 µm.
Early SAC homotypic interactions impact DS circuit function
Finally, we investigated the extent to which developmental events controlled by MEGF10 affect DS circuit function. We sought to determine whether the anatomical perturbations caused by loss of MEGF10 – that is, SAC laminar targeting and mosaic spacing errors – alter direction coding by ooDSGCs. To do this, we recorded from wild-type and Megf10–/– retinas on a large-scale multielectrode array (Field et al., 2007; Yu et al., 2017). ooDSGCs were identified based on their responses to drifting gratings and moving bars (see Materials and methods), which unambiguously distinguished them from other recorded RGCs (Figure 12A). Because MEGF10 is not expressed in the adult DS circuit (Kay et al., 2012), we could be confident that any mutant physiological phenotypes reflect anatomical changes that arose during development.
Figure 12 with 1 supplement see all
Broader and weaker direction tuning of ooDSGCs in Megf10 mutants.
(A) Histograms of RGC direction selectivity indices, measured on a multielectrode array, for wild-type (WT, black) and Megf10–/– (red) retinas. Bimodal histograms fit with two-Gaussian mixture model distinguished DSGCs (filled bars) from non-DSGCs (open bars). (B) Spike rasters from representative WT and Megf10–/– posterior-preferring ooDSGCs in response to a bright bar moving along 12 directions (arrows). (C) Direction tuning curves from cells in B normalized to the maximum response (line: von Mises fit). Non-zero values at tails of mutant curve reflect increase in null-direction spikes (B, left- and right-most bins). (D,E) Cumulative distribution of tuning widths (D) and tuning strengths (E) for all ooDSGCs recorded from two retinas of each genotype (WT n = 80 cells; Megf10–/–n = 74 cells). Mutant ooDSGC population is tuned more broadly (D, right shift of red curve) and more weakly (E, left shift of red curve) than WT. Similar results were obtained when ON and OFF responses were considered separately (not shown). Mutant ooDSGCs also exhibit higher firing rate to null direction motion (E, inset). **p=0.005 (D), p=0.003 (E), paired KS-test. (F) Rasters and polar plot of a representative WT ooDSGC, highlighting preferred directions of ON (gray) and OFF (black) responses (arrows). Δϕ, angular difference between preferred directions of ON and OFF responses. (G) ON and OFF direction tuning curves for cell in F (line, von Mises fit). ON and OFF preferred directions (arrowheads) are well aligned in WT retina. (H) Cumulative distribution across all ooDSGCs of ON-OFF preferred direction difference (Δϕ). Same cells as in D,E. Rightward shift of mutant curve indicates larger ON-OFF misalignment. **p=0.004, paired KS test. For all panels, background light level was photopic (104 P*/M-cone/sec; contrast of moving bar was 60%). Error bars/bands, S.E.M. Also see Figure 12—figure supplement 1.
These experiments revealed that ooDSGCs with robust direction selectivity were present in both wild-type and Megf10–/– retinas (Figure 12A,B), and constituted a similar fraction of the RGC population in both strains (wild-type: 80/609, 13.1%; mutant: 74/551, 13.4%). Furthermore, loss of Megf10 did not alter the organization of ooDSGC preferred directions along cardinal axes (Oyster and Barlow, 1967), or the fraction of ooDSGCs preferring each direction (Figure 12—figure supplement 1). These results are consistent with the observation that mutant SACs remain paired with ooDSGC dendrites and bipolar cell axons even when normal lamination and arbor spacing are disrupted. They indicate that the qualitative functional properties of the circuit are still present.
However, a more careful examination of DS tuning properties in Megf10–/– retinas revealed clear quantitative differences in ooDSGC responses. Moving bars were used to measure the width and strength of direction tuning for each identified ooDSGC across the populations recorded on the electrode array (Figure 12C). Tuning width was measured as the circular standard deviation of the tuning curve, while tuning strength was measured as the normalized response difference to motion in the preferred and null directions (see Materials and methods). These experiments revealed systematic shifts toward broader (Figure 12D) and weaker (Figure 12E) direction tuning across the population of ooDSGCs in Megf10 mutant retinas. This was mainly due to higher null direction spiking among ooDSGCs in mutants (Figure 12B,C,E). Furthermore, these effects on tuning width and strength persisted across a broad range of stimulus contrasts (Figure 12—figure supplement 1). These results demonstrate that disruption of MEGF10-dependent developmental patterning degrades the precision and strength of ooDSGC direction tuning. They further suggest that perturbations to the anatomical regularly of the circuit across space (e.g. laminar uniformity and SAC spacing) may effectively introduce noise in the DS circuit that broadens and weakens direction tuning (see Discussion).
This idea led us to consider additional functional properties of ooDSGCs that might depend on the spatial regularity of the DS circuit, and therefore might be perturbed in Megf10 mutants. One such property is the generation of symmetric DS responses to stimuli that are darker or brighter than the background (Figure 12F,G). This ON-OFF symmetry allows the DS response to be largely insensitive to contrast reversals (Amthor and Grzywacz, 1993); it arises because ooDSGCs receive highly symmetric SAC inputs in both ON and OFF sublayers (Figure 1A). In Megf10 mutants, ON-OFF anatomical symmetry is disturbed, because ON and OFF SAC errors are not spatially correlated (Figure 8A–C). We hypothesized that this might lead to disparities in the direction tuning of individual cells’ ON and OFF responses. Indeed, Megf10–/– ooDSGCs exhibited greater separation (i.e. less coherence) between their ON and OFF preferred directions than wild-type ooDSGCs, across a broad range of contrasts (Figure 12H; Figure 12—figure supplement 1). These results support the idea that MEGF10 serves to establish a highly uniform and regular network of SAC dendrites (via controlling both the precise timing of INL lamination and through regularizing inter-SAC spacing), the net effect of which is to allow greater precision and coherence in the direction tuning of ooDSGCs.
Discussion
Neural circuits typically consist of multiple cell types born at different places and times, raising the question of how circuit partners manage to converge at a common site for selective synapse formation. Here, we describe a developmental strategy that the retinal DS circuit uses to solve this problem. We show that SACs coordinate amongst themselves to assemble a dendritic scaffold that subsequently recruits projections from their DS circuit partners. By identifying for the first time a genetic manipulation – loss of Megf10 – that causes SACs to misproject outside their two typical IPL layers, we uncover mechanisms by which SACs assemble this dendritic scaffold. Further, we use Megf10 mutants to examine the effects on DS circuit anatomy and function when the SAC scaffold is disrupted. We find that MEGF10 establishes DS circuit spatial homogeneity across the retina, both by controlling IPL innervation patterns and by positioning SAC cell bodies. In Megf10 mutants, disruptions in circuit homogeneity occur with minimal effects on radial SAC dendrite anatomy or synaptic partnering, making the phenotype unique among DS circuit developmental mutants. Finally, we find that this abnormal spatial pattern degrades DS circuit function by broadening the range of directions to which ooDSGCs will respond, and by weakening overall direction selectivity. These results provide new insight into general strategies for circuit development, as well as the specific mechanisms that ensure functional assembly of the DS circuit.
Homotypic recognition as a mechanism regulating dendrite differentiation
During radial migration, newborn central nervous system neurons have a multipolar morphology, but on arrival at their final position within the tissue they become highly polarized (Nadarajah et al., 2001; Tabata and Nakajima, 2003; Cooper, 2014; Chow et al., 2015; Krol et al., 2016; Hinds and Hinds, 1978). This morphological change enables elaboration of dendrites and integration into local circuitry. If dendrite differentiation begins early, migration is impaired (Hoshiba et al., 2016), suggesting that the transition from migratory to mature morphology must be highly regulated to ensure that neurons only differentiate once they arrive at their final position. The extracellular cues that signal arrival are poorly understood in most nervous system regions.
Here, we show that SACs use homotypic recognition, mediated by MEGF10, to initiate IPL-directed dendrite morphogenesis. When deprived of homotypic neighbors or MEGF10, SACs at the IPL retain a multipolar morphology (compare Figure 2C to Figures 3F and 6A) instead of polarizing arbors toward the IPL. This indicates that the transition from migratory to mature morphology is impaired in the absence of SAC homotypic recognition. We show that migrating SACs first establish homotypic contact upon arrival at the inner retina. At this stage, they are still multipolar (Figure 2D), but they orient primary dendrites tangentially within the INBL to ultimately contact their SAC neighbors. These contacts occur prior to IPL innervation, and are required for it to occur in a timely manner. SACs lacking neighbors or the molecular means to detect them (i.e. MEGF10) appear to persist in this multipolar soma-layer-targeting phase, causing over-innervation of the INL/GCL and delaying IPL innervation (Figure 6G). Thus, establishment of homotypic contact is a key checkpoint for the progression of SAC dendrite differentiation and IPL sublayer morphogenesis.
We propose that the function of this checkpoint is to ensure that SACs elaborate dendrites only when they have arrived adjacent to the IPL. The presence of other SACs that have already completed their migration is a reliable indicator of arrival in the proper location. Because soma-layer SAC contacts appear earliest, and because MEGF10 selectively influences IPL innervation during the period when they exist, we favor the notion that the key homotypic interactions occur through these arbors. However, we cannot exclude that IPL-based interactions also play a role. INL-directed arbors resembling those we describe can be discerned in many developing zebrafish amacrine cells (Godinho et al., 2005; Chow et al., 2015), raising the possibility that this mechanism applies across species and across other amacrine cell types. Because most neurons require a way to control when and where they differentiate, we anticipate that this homotypic contact strategy, or variations upon it, may have important roles in the differentiation of other CNS neurons at the completion of their radial migration.
MEGF10 as the signal mediating SAC homotypic recognition
We conclude that MEGF10 is the molecule responsible for homotypic recognition during SAC IPL innervation. Four key results support this conclusion. First, MEGF10 is expressed at the right time and place to assume this role: It is expressed selectively in SACs (Figure 1), upon conclusion of their radial migration, and in the soma-layer arbors that we propose mediate recognition (Figure 4). Second, Megf10 null mutant SACs phenocopy the dendrite polarization errors seen in solitary Ptf1a-cKO SACs, suggesting that homotypic recognition requires Megf10. Third, co-immunoprecipitation experiments indicate that MEGF10 interacts with itself via its extracellular domain, suggesting it could act as both ligand and receptor. While this biochemical interaction may take place in the cis configuration, the fourth line of evidence indicates that MEGF10 interacts in trans as well: Using a conditional-null Megf10 allele in vivo, we show that MEGF10 is required on the cell that sends homotypic signals as well as the cell receiving those signals. Loss of MEGF10 on either side leads to dendritic phenotypes resembling solitary SACs and Megf10 null mutants. Together, these data are consistent with a model whereby SAC-SAC contact initiates a transcellular MEGF10 homophilic interaction, in which MEGF10 serves as both receptor and ligand to trigger the switch from migratory to mature morphology (see model, Figure 7K).
This homophilic model of MEGF10 function is consistent with its role during establishment of mosaic cell body patterning (Kay et al., 2012). In that context, MEGF10 acts as ligand and receptor to mediate cell-cell repulsion, thereby spacing SAC somata evenly across the retina. Here we discover a second MEGF10 function in SAC IPL innervation. Because the two SAC phenotypes have different underlying cell biology (soma movement vs. dendrite dynamics), and separable temporal requirements for MEGF10 function (Figure 9), it seems unlikely that they reflect disruption of a single biological event. Instead, MEGF10 appears to act at distinct, albeit partially overlapping times, to control different aspects of SAC development, each of which are regulated by contact with homotypic neighbors (see model, Figure 9C).
Formation of SAC IPL sublayers
Our results shed light on the mechanisms controlling SAC dendrite lamination. While repulsion mediated by Sema6a and PlexinA2 prevents OFF SACs from straying to the ON sublayer (Sun et al., 2013), molecules required for formation of the SAC sublayers have not been identified. We show that SACs deprived of homotypic neighbors or MEGF10 initially fail to form IPL sublayers, and when they eventually do so, their strata are riddled with errors. Both the lack of sublayers at early stages and the dendritic mistargeting to inappropriate sublayers at maturity are novel SAC phenotypes; they implicate MEGF10 as a key player in forming SAC IPL sublayer-specific projections.
It is generally assumed that sublayer formation has two basic molecular requirements: 1) Attractive/adhesive molecules that mediate co-fasciculation of stratified arbors; and 2) repulsive cues that prevent straying of arbors into other sublayers (Lefebvre et al., 2015; Sanes and Yamagata, 2009). Our MEGF10 studies suggest an additional, earlier requirement for cell-cell interactions that occur prior to neuropil innervation. The purpose of this surprisingly early SAC-SAC interaction, we propose, is to ensure that SACs grow dendrites at the right time and place to co-fasciculate with their SAC neighbors. The molecular basis of this homotypic co-fasciculation – clearly another essential player in sublayer formation – remains to be determined. MEGF10 is probably not involved; the co-fasciculation system appears intact in Megf10 mutants given that sublayers do eventually form. Perhaps this system is part of the mechanism that compensates for loss of MEGF10 to ultimately generate the sublayers.
When IPL arborization is delayed by loss of Megf10, two SAC errors ensue. First, SACs generate mistargeted dendritic material that appears to persist as ectopic IPL sublayers. Second, SACs never completely innervate their sublayers, resulting in fragmented IPL strata. These two errors are caused by delays rather than an ongoing requirement for MEGF10 during later stages of arbor growth, as shown by conditional mutant experiments (Figure 9). Thus, our findings support the idea that timing is critical to the sequential lamination of the IPL: When SAC dendrites arrive in the IPL too late, they encounter a different cellular and molecular milieu that may not support the proper development of their arbors. In this view, the normal role of MEGF10 in DS circuit assembly is to instigate SAC dendrite outgrowth at the crucial time when laminar self-assembly can occur.
SACs may face an additional obstacle to overcoming their delayed IPL innervation in Megf10 mutants: abnormal soma positioning. While mosaic spacing errors do not account for the Megf10 mutant ectopic IPL phenotype, we cannot exclude the possibility that the placement of IPL arbor gaps might be at least partly explained by soma position. If SACs are struggling to make up for their delayed IPL innervation, it is plausible that increasing the distance between SACs (as happens sporadically due to random positioning) might further hinder the development of complete retinal coverage.
SACs as a scaffold for DS circuit assembly
Because of their early stratification, SAC dendrites have been proposed to act as a scaffold that guides assembly of the DS circuit (Stacy and Wong, 2003). A key prediction of this model is that laminar targeting of later-stratifying cell types should depend on the existence of this scaffold. We show using a SAC-specific manipulation – removal of Megf10 – that disruption of SAC stratification causes their bipolar and ooDSGC circuit partners to make corresponding projection errors. Based on the kinds of errors we observed, SACs appear to provide attractive, permissive, and possibly even repulsive arbor sorting cues to influence the laminar positioning of their circuit partners. This work thus constitutes the first critical test of the scaffolding model, and provides strong support for it. We find that SACs use homotypic interactions to initiate formation of their circuit sublayers, and then heterotypic interactions to recruit circuit partners to join them. SACs might achieve their scaffolding functions directly, by providing guidance cues to their partners; or they may do so indirectly, by patterning the IPL projections of an intermediary cell type that in turn guides later-arriving projections. Direct scaffolding may be mediated in part by Cadherins 8 and 9, which regulate interactions between SAC dendrites and DS circuit bipolar cell axons (Duan et al., 2014). Molecular mediators of ooDSGC-SAC dendrite interactions remain to be identified.
Evidence that the SAC scaffold can be repulsive – or at least can exclude bipolar arbors from certain IPL regions – came from our observations of BC axon anatomy. In wild-type retina, we were surprised to note how completely the BC3a, BC5, and BC7 axon terminals were excluded from the SAC territory – they contacted it but did not enter (Figure 11B–D; Figure 11—figure supplement 1). This behavior stands in stark contrast to the behavior of ooDSGC dendrites, which completely overlapped SACs (Figure 10A–B; Vaney and Pow, 2000). Moreover, in Megf10 mutants, the laminar distance between BC5 and BC7 terminals was reduced in the absence of SAC arbors, and increased in the presence of SAC ectopias, further suggesting the existence of local SAC-BC repulsion. The finding that SACs exclude bipolar circuit partners from their sublayers appears at first counterintuitive. But given that no bipolar cell type is exclusively devoted to the DS circuit (Wässle et al., 2009; Greene et al., 2016; Kim et al., 2014), a mechanism must exist to ensure that they can also contact non-DS partners. We speculate that SACs initially recruit their bipolar partners using long-range attractive cues, and then use contact-repulsion (or an equivalent arbor sorting mechanism) to displace bipolar arbors such that they remain in contact with the SAC layers but also innervate adjacent layers. This model is consistent with bipolar arbor phenotypes in Megf10 mutants, but will require further study.
Role of MEGF10 in the functional assembly of DS circuitry
We found that impairment of SAC interactions in the perinatal retina causes permanent functional DS circuit deficits. In Megf10 mutants, direction tuning of ooDSGCs becomes broader and weaker, and their ON/OFF preferred directions are less aligned. Direction tuning is degraded in large part because mutant ooDSGCs have aberrant spiking responses to null-direction stimuli. This suggests that impaired null-direction inhibition – which arises from SACs – is a key contributor to the phenotype. Broader ooDSGC tuning curves have been shown, in modeling studies, to degrade population-level coding of directional information, and the ability of downstream neurons to extract such information (Fiscella et al., 2015). Thus, the physiological phenotypes we identified are likely sufficient to impair the ability of mutant retina to appropriately relay visual information.
Dysfunctional DS circuit physiology in Megf10 mutants is almost certainly a consequence of its effects on development, because neurons do not express MEGF10 beyond the second postnatal week (Kay et al., 2012). Further, even though MEGF10 is expressed by Müller glia in adulthood, we have been unable to detect any changes in Müller glia anatomy or interactions with DS circuit synapses upon loss of Megf10 function (Wang et al., 2017); J.W. and J.N.K., unpublished observations). We therefore conclude that anatomical changes to the DS circuit arising during development are responsible for circuit dysfunction.
The fundamental change to DS circuit anatomy in Megf10 mutants is altered distribution of arbors and synapses, unlike other manipulations which simply serve to destroy SAC radial morphology or disrupt synaptic partnering among DS circuit cells (Sun et al., 2013; Duan et al., 2014; Kostadinov and Sanes, 2015; Peng et al., 2017). In Megf10 mutants, the combined effect of mosaic spacing defects and IPL laminar targeting errors is to disturb the regularity of SAC IPL innervation. As a result, some parts of the visual map become over-innervated (e.g. Figure 10C) while others are uninnervated (Figure 10D). In turn, ooDSGCs are recruited to the over-innervated regions and excluded from uninnervated gaps, likely causing sporadic local inhomogeneity in synapse density across visual space. According to some models of DS, which posit that the total amount of SAC inhibition is the key factor underlying DS responsiveness, these relatively small-scale changes would be considered unlikely to change circuit function (Taylor and Vaney, 2002; Demb, 2007). A more recent alternate view is that the fine spatial arrangement of glutamatergic inputs to SACs, and the synaptic balance of SAC and bipolar input onto ooDSGC dendrites, are both important for DS responses (Ding et al., 2016; Vlasits et al., 2016; Poleg-Polsky and Diamond, 2016; Sethuramanujam et al., 2016; Sethuramanujam et al., 2017). The finding that Megf10 mutants have DS tuning phenotypes suggests that local synaptic arrangements are indeed important for the DS computation. More broadly, this finding shows that the developmental mechanisms we describe here are important for enabling circuit function, raising the possibility that other circuits throughout the retina and CNS may use similar developmental mechanisms to establish their functional connectivity.
Materials and methods
Key resources table
| Reagent type | Designation | Source or reference | Identifier | Additional information |
|---|---|---|---|---|
| Antibody | Megf10: rabbit, 1:1000 | Kay et al. (2012) | ||
| Antibody | Sox2: rabbit, 1:500 | Abcam | ab97959 | |
| Antibody | Sox2: goat, 1:500 | Santa Cruz | sc-17320 | |
| Antibody | ChAT: goat, 1:400 | EMD Millipore | AB144P | |
| Antibody | Beta Galactosidase: rabbit, 1:5000 | other | Antibody was a gift of J.R. Sanes, Harvard | |
| Antibody | GFP: chicken, 1:1000 | Life Technologies | A10262 | |
| Antibody | GFP (Co-IP): rabbit, 1:1000 | Thermo Fisher Scientific | A-6455 | |
| Antibody | AP-2a: mouse, 1:200 | Developmental Studies Hybridoma Bank | 3B5 | |
| Antibody | RBPMS: guinea pig, 1:2000 | other | Antibody was a gift of N. Brecha, UCLA | |
| Antibody | Chx10: sheep, 1:300 | Exalpha | X1180P | |
| Antibody | Chx10: goat, 1:500 | Santa Cruz | sc-21690 | |
| Antibody | GAD65: rabbit, 1:1000 | Millipore | AB1511 | |
| Antibody | FLAG: mouse, 1:500 | Sigma Aldrich | F-1804 | |
| Antibody | VGLUT3 guinea pig | synaptic systems | 135 204 | |
| Antibody | Synaptotagmin-2 (Syt2), mouse, 1:250 | Zebrafish International Resource Center | ZDB-ATB-081002–25 | |
| Antibody | Isl1: mouse, 1:25 | Developmental Studies Hybridoma Bank | 39.4D5 | |
| Antibody | Internexin: rabbit, 1:1000 | EMD Millipore | AB5354 | |
| Antibody | Normal Rabbit IgG | Cell Signaling Technology | 2729S | |
| Antibody | Normal Mouse IgG | Cell Signaling Technology | 5415S | |
| Antibody | Alexa Fluor 488 AffiniPure Donkey Anti-Chicken: 1:1000 | Jackson ImmunoResearch | 703-545-155 | |
| Antibody | Alexa Fluor 488 AffiniPure Donkey Anti-rabbit: 1:1000 | Jackson ImmunoResearch | 711-545-152 | |
| Antibody | Alexa Fluor 488 AffiniPure Donkey Anti-goat: 1:1000 | Jackson ImmunoResearch | 705-545-147 | |
| Antibody | Alexa Fluor 488 AffiniPure Donkey Anti-mouse: 1:1000 | Jackson ImmunoResearch | 706-605-148 | |
| Antibody | Alexa Fluor 647 AffiniPure Donkey Anti-rabbit: 1:1000 | Jackson ImmunoResearch | 705-605-147 | |
| Antibody | Cy3-AffiniPure Donkey Anti-rabbit: 1:1000 | Jackson ImmunoResearch | 715-165-151 | |
| Antibody | Cy3-AffiniPure Donkey Anti-Guinea Pig: 1:1000 | Jackson ImmunoResearch | 706-165-148 | |
| Antibody | Cy3-AffiniPure Donkey Anti-Goat: 1:1000 | Jackson ImmunoResearch | 705-165-147 | |
| Antibody | IRDye 680RD Donkey anti-Mouse IgG (H + L): 1:1000 | Li-Cor Biosciences | 925–68072 | |
| Antibody | IRDye 800CW Donkey anti-Rabbit IgG (H + L): 1:1000 | Li-Cor Biosciences | 925–32213 | |
| Biological sample (AAV) | AAV9.hEF1a.lox.TagBFP.lox.eYFP.lox.WPRE.hGH-InvBYF(Harvard) | Penn Vector Core | AV-9-PV2453 | |
| Biological sample (AAV) | AAV9.hEF1a.lox.mCherry.lox.mTFP1.lox.WPRE.hGH-InvCheTF(Harvard) | Penn Vector Core | AV-9-PV2454 | |
| Chemical compound | Fetal Bovine Serum | Life Technologies | 16250–078 | |
| Chemical compound | Opti-MEM I Reduced Serum Medium | Thermo Fisher Scientific | 31985070 | |
| Chemical compound | Polyethylenimine (PEI), Linear (MW 25,000) | VWR/Polysiciences | 23966–2 | |
| Chemical compound | 16% Paraformaldehyde | Electron Microscopy Sciences | 15710 | |
| Chemical compound | Normal Donkey Serum | Jackson ImmunoResearch | 017-000-121 | |
| Other | Immun-Blot Low Fluorescence PVDF membrane | Bio-Rad | 1620264 | |
| Chemical compound | Fluoromount G | SouthernBiotech | 0100–01 | |
| Chemical compound | Hoechst 33258 | Invitrogen | H21491 | |
| Chemical compound | Isothesia: Isoflurane | Henry Schein | 11695–6776 | |
| Chemical compound | Tissue Freezing Medium | VWR | 15148–031 | |
| Chemical compound | Acrylamide/Bis solution | Bio-Rad | 161–0158 | |
| Chemical compound | 4x Laemmli Sample Buffer | Bio-Rad | 1610747 | |
| Chemical compound | Immun-Blot Low Fluorescence PVDF membrane | Bio-Rad | 1620264 | |
| Chemical compound | Odyssey Blocking Buffer | Li-Cor Biosciences | 927–40000 | |
| Chemical compound | Dynabeads Protein G for Immunoprecipitation | Thermo Fisher Scientific | 10003D | |
| Chemical compound | cOmplete, Mini, EDTA-free Protease Inhibitor Cocktail Tablets | Roche | 4693159001 | |
| commercial assay or kit | Bio-Rad DC Protein Assay Kit | Bio-Rad | 5000112 | |
| Strain (mus musculus) | Mouse: Megf10LacZ | Kay et al. (2012) | Megf10tm1b(KOMP)Jrs | |
| Strain (mus musculus) | Mouse: Megf10flox | this study | Megf10tm1c | see Methods, ‘Animals’ section |
| Strain (mus musculus) | Mouse: Ptf1a-cKO | Krah et al., 2015 | Ptf1atm3Cvw | |
| Strain (mus musculus) | Mouse: Isl1Cre | Jackson Labs 024242 | Isl1tm1(cre)Sev/J | |
| Strain (mus musculus) | Mouse: Hb9-GFP | Jackson Labs 005029 | B6.Cg-Tg(Hlxb9-GFP)1Tmj/J | |
| Strain (mus musculus) | Mouse: ChATCre | Jackson Labs 006410 | Chattm2(cre)Lowl | |
| Strain (mus musculus) | Mouse: Six3-Cre | Jackson Labs 019755 | Tg(Six3-cre)69Frty | |
| Strain (mus musculus) | Mouse: Kcng4Cre | Jackson Labs 029414 | Kcng4tm1.1(cre)Jrs | |
| Strain (mus musculus) | Mouse: Drd4-GFP | Huberman et al. (2009) | Tg(Drd4-EGFP)W18Gsat | |
| Strain (mus musculus) | Mouse: Gjd2-GFP | MMRRC | Tg(Gjd2-EGFP)JM16Gsat/Mmucd; RRID:MMRRC_030611-UCD | |
| Strain (mus musculus) | Mouse: Rosa26mTmG | Jackson Labs 007676 | Gt(ROSA)26Sortm4(ACTB-tdTomato,-EGFP)Luo | |
| Strain (mus musculus) | Mouse: Rosa26fGFP | Rawlins et al. (2009) | Gt(ROSA)26Sortm1(CAG-EGFP)Blh | |
| Strain (mus musculus) | Mouse: Rosa26Ai14 | Jackson Labs 007914 | B6.Cg-Gt(ROSA)26Sortm14(CAG-tdTomato)Hze/J | |
| Strain (mus musculus) | Mouse: ACTB:FLPe | Jackson Labs 003800 | B6;SJL-Tg(ACTFLPe)9205Dym/J | |
| Strain (mus musculus) | Mouse: Gad1-GFP | Jackson Labs 007677 | Tg(Gad1-EGFP)G42Zjh | |
| Strain (mus musculus) | Mouse: C57Bl6/J | Jackson Labs 000664 | C57BL/6J | |
| Cell line (human) | HEK293T | ATCC | 293T (ATCC CRL-3216) | Cell line was authenticated by ATCC at time of purchase |
| Recombinant DNA reagent | CMV-M10-FLAG | this paper | see Methods,‘Generation of MEGF10-ΔICD constructs’ section | |
| Recombinant DNA reagent | CMV-M10-GFP | Kay et al. (2012) | ||
| Recombinant DNA reagent | pCMV-MEGF10-ΔICD-GFP | Kay et al. (2012) | ||
| Recombinant DNA reagent | MEGF10-ΔICD-Flag | this paper | see Methods, ‘Generation of MEGF10-ΔICD Constructs’ section | |
| Recombinant DNA reagent | pAAV-EF1a-Brainbow-tagBFP-EYFP-WPRE | Addgene | 45185 | |
| Recombinant DNA reagent | pAAV-EF1a-Brainbow-mTFP1-Cherry-WPRE | Addgene | 45816 | |
| Software | Fiji/ImageJ | Schindelin et al. (2012) | ||
| Software | Prism | GraphPad | ||
| Software | NIS Elements | Nikon Instruments | ||
| Software | Custom JAVA scripts for spike sorting | Yu et al., 2017 | ||
| Software | Custom MATLAB scripts for data analysis | this paper | https://github.com/Field-Lab/megf10-dstuning | see Methods,‘Multielectrode Array Recordings’;‘Quantification and Statistical Analysis’ sections |
| Software | Matlab | Mathworks, Natick, MA | ||
| Software | Image StudioTM | LI-COR Biosciences | ||
| Software | Photoshop | Adobe | ||
| Sequence-based reagent | M10flagNotI_Rev | IDT | ATAGCGGCCGCttaCTTGTCGTCATCGTCTTTGTAGTCttcactgctgctgctgctgctg | |
| Sequence-based reagent | M10flag_Fwd | IDT | GGTACATGCCTGTGCGAAGCA | |
| Sequence-based reagent | Cyto9_flag_Rev1 | IDT | 5’ATAGCGGCCGCttaCTTGTCGTCATCGTCTTTGTAGTC TTCCTTCCTCTTCTGCTTGTGT |
Animals
All animal experiments were reviewed and approved by the Institutional Animal Care and Use Committee of Duke University. The animals were maintained under a 12 hr light-dark cycle with ad lib access to food and water. Retinas from adult (4–8 weeks old) Megf10–/– mutant mice and wild-type control mice with same genetic background were used for experiments performed on the multielectrode array (MEA). Animals were dark-adapted overnight prior to the experiment.
For this study, the following transgenic and mutant mouse lines were used: (1) Megf10tm1b(KOMP)Jrs (Kay et al., 2012), referred to as Megf10– or Megf10lacZ; (2) Ptf1atm3Cvw (Krah et al., 2015), referred to as Ptf1aflox or (when crossed to Cre mice) Ptf1a-cKO; (3) Isl1tm(cre)Sev (Yang et al., 2006), referred to as Isl1Cre; (4) Tg(Hlxb9-GFP)1Tmj/J (Trenholm et al., 2011), referred to as Hb9-GFP; (5) Chattm2(cre)Lowl (Rossi et al., 2011), referred to as ChatCre; (6) Tg(Six3-cre)69Frty (Furuta et al., 2000) referred to as Six3-Cre; (7) Kcng4tm1.1(cre)Jrs (Duan et al., 2014) referred to as Kcng4Cre; (8) Tg(Drd4-EGFP)W18Gsat (Huberman et al., 2009), referred to as Drd4-GFP; (9) Tg(Gjd2-EGFP)JM16Gsat, referred to as Gjd2-GFP; (10) Tg(Gad1-EGFP)G42Zjh, referred to as Gad1-GFP. Two Cre reporter strains were used that express membrane-targeted green fluorescent protein (mGFP) upon Cre recombination: (1) Gt(ROSA)26Sortm4(ACTB-tdTomato,-EGFP)Luo, also known as mT/mG (Muzumdar et al., 2007); (2) Rosa26fGFP (Rawlins et al., 2009). An additional Cre reporter strain was used that expresses tdTomato fluorescent protein upon Cre recombination: Gt(ROSA)26Sortm14(CAG-tdTomato)Hze (Madisen et al., 2010). See Key Resources table for repository stock numbers where applicable.
To produce Megf10flox mice, Megf10tm1a(KOMP)Jrs mice (Kay et al., 2012) were crossed to germline Cre strain B6;SJL-Tg(ACTFLPe)9205Dym/J, thereby generating a functional allele (also known as Megf10tm1c) in which exon four was flanked by loxP sites.
Cell culture
Request a detailed protocolHEK293T cells were obtained from, validated by, and mycoplasma tested by ATCC. The cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) with 10% bovine growth serum, 4.5 g/L D-glucose, 2.0 mM L-glutamine, 1% Penicillin/Streptomycin in 10 cm cell culture dishes. Cells were passaged every 2–3 days to reach confluence. Before splitting, culture media were removed and Dulbecco’s phosphate-buffered saline (D-PBS) was used to rinse cell layers as well as removing residual serum. Cells were detached from dish with 4 ml of 0.05% Trypsin and incubated at 37°C until cell layer is dispersed (about 5 min). Equal volume of complete culture media was added to the dish to inhibit protease activity. The suspension was centrifuged at 200 x g for 5 min. Supernatant was aspirated and the cells were suspended with appropriate amount of media and plated (1:4-1:8). Cells used for experiments were passaged no more than 10 times. Cell stocks were stored as 2 million cells per vial in complete culture media with 10% DMSO in liquid nitrogen.
Identification of DS circuit cell types using antibody and transgenic markers
SAC markers in mature retina
Request a detailed protocolAntibodies to choline acetyltransferase (ChAT) were used as a SAC marker in mice older than P5. This antibody stains SAC somata and their dendrites in the IPL (e.g. Figure 8A).
SAC markers in embryonic and neonatal retina
Request a detailed protocolAntibodies to ChAT and vesicular acetylcholine transporter, typically used as SAC markers in the mature retina, do not stain reliably in the embryonic and neonatal (P0-P3) mouse retina, precluding their use as markers during one of the key time periods of this study. We therefore characterized several other SAC markers that we found to be suitable for definitive SAC identification and their anatomical characterization in the E16-P3 period:
The Megf10lacZ allele (Kay et al., 2012) drives strong, selective β-galactosidase (βgal) expression in all SACs starting at embryonic day (E)17 (Figure 1B; Figure 1—figure supplement 1; data not shown). Horizontal cells are also labeled. Expression is strong enough to allow characterization of SAC dendrite anatomy at these early stages. Antibodies to Megf10 yield a similar staining pattern (Figure 1—figure supplement 1; Figure 4B,C), but staining of fine dendritic arbors was brighter with anti-βgal staining of Megf10lacZ mice, so this approach was used for most of our anatomical experiments analyzing the full SAC population at or before P3. In some such experiments, a Megf11lacZ allele (Kay et al., 2012) was also present; this allele drives βgal expression in essentially the same pattern as Megf10lacZ and therefore contributed to signal brightness. The presence of this allele had no apparent effect on SAC anatomy, in either wild-type or Megf10 mutant background.
Antibodies to Sox2 (Whitney et al., 2014) strongly label all SAC nuclei in the INL and GCL, starting at embryonic stages (Figure 1—figure supplement 1; Figure 2D,E,G). Progenitor cells in the ONBL are also labeled. This marker was typically used in conjunction with Megf10lacZ to provide definitive identification of SACs as βgal+Sox2+ cells.
Antibodies to internexin label SAC intermediate filaments, which localize in a polarized manner to the primary dendrite(s) and the side of the cell body from which they emerge (Figure 2—figure supplement 1). Primary dendrites were defined as any first-order dendrite branch, that is those arising directly from the cell body. Internexin is a selective marker of SAC in perinatal mouse retina, as previously shown in tree shrew (Knabe et al., 2007). RGC axons are also labeled (Figure 2—figure supplement 1).
Antibodies to Isl1 (Figure 2A) label all SAC nuclei, starting at cell cycle exit (Galli-Resta et al., 1997). A large subset of RGCs are also labeled. The Isl1Cre knock-in mouse (Yang et al., 2006) faithfully recapitulated this expression pattern (Figure 2A,B) and was used to study SAC anatomy at embryonic stages (see below for further details).
SAC single-cell labeling
Request a detailed protocolTo assess the single-cell morphology of individual SACs during early postnatal development, the ChatCre line was used. In contrast to mature retina (e.g. Figure 8B), in which all SACs were labeled, ChatCre expression was rare and sporadic in early postnatal retina (Figure 1C; Figure 1—figure supplement 1), as reported previously (Xu et al., 2016). Therefore, when crossed with Cre reporter mice to make ChatmG animals, the full anatomy of individual SACs was clearly delineated (e.g. Figure 2H–K). We did not typically observe Cre recombination in non-SAC cell types; nevertheless, we always co-stained with another SAC marker, either Sox2 or Megf10:βgal, to confirm the SAC identity of the cells that were analyzed.
ooDSGC markers
Request a detailed protocolTwo mouse lines were used, each of which labels distinct types of ooDSGCs. Hblx9-GFP (referred to as Hb9-GFP throughout the manuscript) labels the superior subtype of ooDSGC, while Drd4-GFP labels the posterior subtype of ooDSGC (Trenholm et al., 2011; Huberman et al., 2009).
DS-circuit bipolar cell markers
Request a detailed protocolFour types of bipolar cells have been shown to make monosynaptic connections with SACs and/or ooDSGCs: Types BC2, BC3a, BC5, and BC7 (Duan et al., 2014; Ding et al., 2016; Greene et al., 2016; Kim et al., 2014; Chen et al., 2014). OFF bipolar cells BC2 and BC3a were labeled, respectively, by antibodies to Syt2 and HCN4 (Wässle et al., 2009).
ON bipolar cells BC5 and BC7 were marked with Kcng4Cre (Duan et al., 2014) crossed to mGFP Cre reporter mice (denoted Kcng4mG). Labeling of BC7 was more prominent with the Rosa26 locus mGFP Cre reporter line that we used, compared to the cytosolic GFP reporter driven by Thy1 that was used by Duan et al. (2014).
Gjd2-GFP was also used to label BC5 bipolar cells (Figure 11—figure supplement 1). In adult retina, GFP was strongly expressed by a bipolar cell type that ramified in a laminar location typical of BC5 (Sidney Kuo, University of Washington, personal communication). We confirmed this expression pattern; weak expression in amacrine cells was also noted (Figure 11—figure supplement 1). At earlier developmental stages the amacrine cell staining was much stronger and filled many amacrine processes throughout the IPL, precluding use of this line for developmental studies of bipolar axons (M. Stogsdill and J.N.K, unpublished observations).
Immunohistochemistry
Retinal cross sections
Request a detailed protocolMice were anesthetized by isoflurane or cryoanesthesia (neonates only) followed by decapitation. Eyes were enucleated, washed in PBS, and fixed in PBS containing 4% formaldehyde (pH 7.5) for 1.5 hr at 4°C. After fixation, eyes were washed 3X with PBS and stored in PBS containing 0.02% sodium azide at 4°C until further processing. Retinas were dissected from the eyecup, cryoprotected by equilibration in PBS containing 30% sucrose, then embedded in Tissue Freezing Medium and frozen by submersion in 2-methylbutane chilled by dry ice. Tissue sections were cut on a cryostat to 20 µm and mounted on Superfrost Plus slides. Slides were dried on a slide warmer for 1 hr then stored at −80°C or used immediately.
For antibody labeling, slides were washed for 5 min with gentle agitation in PBS to remove embedding medium and blocked for 1 hr in PBS + 0.3% Triton X-100 (PBS-Tx) containing 3–5% normal donkey serum. Primary antibodies were diluted in blocking buffer, added to slides, then incubated overnight at 4° C. Slides were washed with PBS 3X for 10 min followed by incubation with secondary antibody diluted in PBS-Tx for 1–2 hr at RT. Slides were washed again with PBS 3X for 10 min then coverslipped using Fluoromount G.
Retinal whole-mounts
Request a detailed protocolTissue was processed as above up to the point of dissection from the eyecup. After dissection from eyecup, retinas were washed in PBS then blocked for 3 hr with agitation at 4° C in blocking buffer (constituted as described above). Primary antibodies were diluted in blocking buffer, added to retinas, and incubated for 5–7 days with gentle agitation at 4°C. Retinas were washed 3X with PBS over the course of 2 hr with gentle agitation. Secondary antibody was diluted in PBS containing 0.3% Triton X-100 and was added to retinas followed by incubation overnight at 4° C with gentle agitation. Retinas were washed again 3X in PBS over the course of 2 hr with gentle agitation. For mounting on slides, four radial incisions separated by 90° were made centripetally, approximately 1/3 the radius of the retina. Retinas were flattened on nitrocellulose paper photoreceptor side down and coverslipped with Fluoromount G.
Image acquisition and processing
Request a detailed protocolSections and whole-mounts were imaged on a Nikon A1 or an Olympus FV300 confocal microscope. Image Z-stacks (Z-resolution 0.4–0.5 µm for whole-mount images; 0.8–1.0 µm for cross-sections) were imported to Fiji (Schindelin et al., 2012), de-noised by median-filtering (0.5–2.0 pixel radius), and projected to a single plane. The portion of the stack selected for maximum-intensity projection was determined by the Z-volume of the structure to be depicted in the final image. Except where noted, data analysis and quantification was only performed using original stacks, not Z-projections. Color channels were assembled, and minor adjustments to brightness and contrast were made, in Adobe Photoshop. When images were to be compared, equivalent adjustments were performed on all images in the experiment. The width of the IPL is marked in many of the figures; this was determined by one of the following methods: (1) counterstaining with Hoechst to label all cell nuclei; (2) tdTomato fluorescence from unrecombined cells in mT/mG mice, which fills the IPL; (3) immunofluorescence against GAD65, which also fills the IPL; (4) autofluorescence signal intensity differences between soma layers and IPL.
Analysis of SAC anatomy in embryonic retina
Request a detailed protocolTo study SAC anatomy during embryonic stages, Isl1Cre was crossed to lox-stop-lox-mGFP Cre reporter mice (mT/mG or Rosa26GFPf; see Key Reagents) to generate Isl1mG animals. Timed-pregnant dams were sacrificed at E16 and eyes collected from embryos (n = 11 mice from three litters). Tissue was processed as described for postnatal eyes, except fixation time was 60 min. Cross-sections were stained with anti-GFP to reveal the morphology of Isl1mG-expressing neurons, as well as Sox2 to distinguish Isl1mG-positive SACs from RGCs. (All cells shown in Figure 2B–G were confirmed to be SACs by Sox2 co-labeling.) In combination with these markers, anti-internexin staining was used to assess orientation of primary dendrites. Location and/or presence of the IPL was determined using Hoechst nuclear staining, which revealed cell body-free neuropil regions, and/or by Isl1mG labeling of neuronal processes, which filled these neuropil regions (Figure 2—figure supplement 2). We assessed anatomy of mGFP+ migrating SACs in the ONBL, as well as SACs in the INBL that were concluding their migration. Morphology of ON SACs in the GCL could not be discerned due to Isl1 expression by RGCs (Figure 2A,B), but because displaced amacrine cells pause at the INL-IPL border before crossing to the GCL (Chow et al., 2015), the population of cells available to analyze might have included both ON and OFF SACs.
To measure the orientation of primary dendrites at E16 and P1, the angle ROI function in ImageJ was used. This function outputs an angle degree measurement (absolute value) between two line segments. The first line segment of the angle was drawn to follow the trajectory of the internexin+ primary dendrite; the endpoint was at the cell body. The second line segment of the angle was a plumb line to the IPL (i.e. it was drawn to intersect the IPL at ~90˚). As such, dendrites oriented exactly toward the IPL were assigned an angle of 0˚. At E16 the IPL was occasionally not present yet; in this case the second line segment was a plumb line to the inner limiting membrane. In cases where the internexin+ dendrite curved, we traced the initial trajectory of the dendrite as it emerged from the cell body. Dendrites were classified as projecting (1) towards the IPL; (2) toward the ONBL; or 3) tangentially, according to the angle scheme delineated in Figure 2—figure supplement 1E. Image stacks were randomly selected for analysis from a larger library of images; within each selected stack every SAC was traced. Isl1mG and Sox2 were used to confirm the SAC identity of each measured cell, as well as the trajectory of the internexin+ dendrite.
Characterization of SAC homotypic arbor network in soma layers
Request a detailed protocolThe homotypic nature of SAC soma-layer contacts was investigated by imaging single ChatmG-labeled OFF SACs in mice also carrying a single copy of the Megf10lacZ allele (Figure 2I,J). Anti-βgal staining was used to reveal the full SAC population, including arbors. En-face images were captured in Z-stacks spanning the INL and IPL; slices corresponding to each layer were separately Z-projected for display in Figure 2 and Figure 2—figure supplement 3. To quantify the frequency of SAC-SAC contacts, we used Z-stacks from P1 tissue to examine the trajectory and termination site of each dendritic tip in three dimensions. The fraction of ChatmG-labeled dendrites terminating on the βgal-positive soma or arbor of a neighboring SAC was quantified. To be counted, the putative contact needed to be confirmed in a single Z-stack slice; where necessary, 3D reconstructions and orthogonal views were used to confirm contact.
We also performed the same analysis on Z-stacks in which one channel had been flipped about the horizontal and vertical axes. This served as a negative control to measure the frequency with which GFP and βgal arbors interact by chance, given their density and geometry in the P1 retina. Sample sizes are given in main text and in Figure 2—figure supplement 3.
Generation and analysis of ‘solitary’ SACs
Reduction of SAC density using Ptf1aflox mice
Request a detailed protocolPtf1aflox mutant mice (Krah et al., 2015) were crossed into the Six3-Cre background to generate Ptf1a-cKO mice. Six3-Cre is expressed by retinal progenitors starting at E9.5 in a high-central-to-low-peripheral gradient (Furuta et al., 2000); Figure 3A). In central retina, where Cre is expressed in all progenitors, amacrine cells were completely absent but bipolar cells, RGCs, Müller glia, and photoreceptors remained (Figure 3B; Figure 3—figure supplement 1; data not shown). In peripheral retina, where Cre recombination was incomplete, amacrine cells derived only from Cre-negative progenitors (Figure 3C). Because the number of Cre-expressing progenitors in peripheral retina still vastly exceeded the number that escaped Cre, amacrine cell density in Ptf1a-cKO peripheral retina was markedly reduced compared to littermate controls (Figure 3A,B; Figure 3—figure supplement 1).
Quantification of dendrite phenotypes in solitary and touching SACs
Request a detailed protocolTo visualize SACs and quantify their arbor targeting frequencies in Ptf1a-cKO mice, we bred Megf10lacZ into the Ptf1aflox background. All Ptf1a-cKO and littermate control mice in these experiments carried one copy of the Megf10lacZ allele. SAC morphology was revealed with anti-βgal. Sox2 was used to confirm the SAC identity of all cells included in the experiment. SACs were scored as ‘solitary’ or ‘touching’ based on whether their dendrites contacted neighboring SACs in the same or adjacent sections. If this could not be determined (e.g. because the adjacent section was missing or damaged), the cell was excluded from further analysis. Because SACs were only present in Ptf1a-cKO peripheral retina, analysis of littermate control SACs was also limited to peripheral retina. In Ptf1a-cKO mice, SACs were more frequently found in the INL than the GCL and it is possible that the INL SACs were a mixed population of ONs and OFFs. Therefore, we did not distinguish between SAC subtypes for the analyses.
IPL projections of βgal-stained cells were examined, and cells were assigned to one of three categories: 1) no arbors projecting to the IPL; 2) Arbors enter the IPL but fail to stratify; 3) Arbors enter the IPL and ramify in a laminar pattern. Examples of the first category of solitary SACs are shown in Figure 3F, left, and Figure 3—figure supplement 1. Examples of the second category are shown in Figure 3F, right, and Figure 3—figure supplement 1. The third category is exemplified by all touching SACs shown (Figure 3E; Figure 3—figure supplement 1). Each cell in the dataset was also scored on an independent criterion: whether it projected to the soma layer (e.g. Figure 3D,F, white arrows).
For each animal in the experiment, the following was calculated and plotted in Figure 3G: (1) Percentage of SACs with projections to the soma layers; (2) percentage of SACs projecting to the IPL (i.e. the cells assigned to categories 2 and 3 above); (3) percentage of SACs with stratified IPL dendrites (i.e. the cells in category 3). Sample sizes: n = 3 wild-type littermates (28, 62, 32 cells analyzed in each animal); n = 4 Ptf1a-cKO animals (11, 35, 13, 12 solitary and 27, 44, 22, 23 touching SACs analyzed in each animal). Statistics: one-way ANOVA with Tukey’s post-hoc test.
Quantification of SAC projection phenotypes in ChatmG mice
Request a detailed protocolSingle SACs labeled in ChatmG and ChatmG;Megf10–/– mice were morphologically assessed in cross-sections. GFP signal was amplified with anti-GFP antibody staining. All GFP+ SACs on any given slide were imaged and analyzed, to avoid cell selection bias, with the exceptions of: 1) cells severed by the cryosectioning process; 2) cells with arbors that could not clearly be distinguished from those of their neighbors; 3) cells in the far retinal periphery, where sections were oblique to retinal layers, obscuring IPL strata. In experiments analyzing Megf10 mutants, littermates were always used as controls to avoid complications arising from the fact that the precise state of retinal development at the time of birth might vary from litter to litter.
A cell was scored as innervating the IPL if it ramified branched dendrites within the neuropil. Dendrites that entered the neuropil but did not branch or stratify (e.g. Figure 6D) were not sufficient. A cell was scored as projecting to the soma layer if arbors emanating from the cell soma or primary dendrite terminated or arborized in the INL (for OFF SACs) or GCL (for ON SACs). The arbor was required to be ~≥1 cell radius in length (i.e. small fine arbors were not counted). One other important exception that was not counted: We observed that many SACs at young ages had single unbranched arbors extending ~180˚ away from the IPL (e.g. Figure 2J,K – all four cells have such arbors, even the ones that do not project towards neighboring SAC somata). These processes were not counted for two reasons. First, their trajectory was such that they were unlikely to join the soma-layer dendrite network or contact neighboring somata. Second, these 180˚ arbors were sometimes still present in P5 SACs (Figure 2—figure supplement 2) and therefore they did not appear to be subject to the same developmental regulation as tangentially-directed arbors (Figure 2L). This observation suggests they are fundamentally different, and likely serve a different (as yet uncharacterized) purpose. No obvious difference in their frequency was observed between wild-type and Megf10 mutants.
To produce graphs in Figures 2L, 6E and 8F, the fraction of cells making ectopic projections – either to the soma layer or to inappropriate IPL sublayers – was calculated for each genotype and each time point. To determine whether a GFP+ IPL arbor was located in normal or abnormal IPL strata, Megf10:βgal was used as a counterstain. ChatCre was rarely expressed in OFF SACs at P0, making it difficult to obtain large sample sizes at this age. For this reason, and because soma-layer projection frequency did not appear to differ much between P0 and P1, the data from each time point was pooled for analysis of Megf10 litters.
Sample sizes for Figure 2L: P0, n = 25 OFF, 63 ON; P1, n = 51 OFF, 79 ON; P2, n = 46 OFF, 55 ON; P3, n = 33 OFF, 49 ON; P5, n = 15 OFF, 26 ON; P7, n = 23 OFF, 34 ON. Data were from four litters of mice, each of which was assessed at no less than two of these time points.
Sample sizes for Megf10; ChatmG experiments (Figure 6E; Figure 8F): Megf10 heterozygous littermate controls: P0/1, n = 11 OFF, 25 ON; P2, n = 25 OFF, 23 ON; P3, n = 17 OFF, 22, ON; P5, n = 16 OFF, 16 ON. Megf10 mutants: P0/1, n = 6 OFF, 25 ON; P2, n = 14 OFF, 20 ON; P3, n = 34 OFF, 41 ON; P5, n = 48 OFF, 54 ON. Data were from two litters of mice.
For the adult data reported in Figure 8F,a different procedure was used; see ‘Quantification of Mosaic Spacing Phenotypes’ section below.
Analysis of Chat-Megf10-cKO conditional mutants
Characterization of timing of MEGF10 deletion
Request a detailed protocolFor initial characterization of when MEGF10 protein is eliminated by the ChatCre driver line, the following experiment was performed: ChatCre; Megf10flox mice were intercrossed with ChatCre; Megf10lacZ carriers to generate ChatCre; Megf10flox/lacZ (Chat-Megf10-cKO) experimental animals and littermate controls (ChatCre; Megf10flox/+). These animals also carried a Rosa26 mGFP Cre reporter allele. Animals were sacrificed at P1, P3, and P5; retinas were cross-sectioned and immunostained for anti-MEGF10 (Figure 7—figure supplement 1). Comparisons were made across animals from the same litter to assess how MEGF10 immunoreactivity changed over time. Two litters were analyzed in this way, each yielding the same conclusion: MEGF10 immunoreactivity was largely eliminated by P5 in Chat-Megf10-cKO mice (Figure 7—figure supplement 1). At P3, overall MEGF10 levels were reduced, but most SACs still expressed detectable protein (Figure 7—figure supplement 1). The cells that lost MEGF10 immunoreactivity by P3 were not necessarily the same cells that recombined the mGFP reporter at the Rosa26 locus (Figure 7F,G). At P1, only a very small number of cells (<5 per retina) could be identified that lacked MEGF10 immunoreactivity; most of these were ON SACs although a few recombined OFF SACs were identified (Figure 7G). We conclude that a small fraction of SACs loses MEGF10 protein prior to P3, while the majority lose MEGF10 between P3 and P5. Further, ON SACs are somewhat more likely to lose MEGF10 before P3 than OFF SACs.
Assessment of morphological and IPL projection phenotypes
Request a detailed protocolTo ask if loss of MEGF10 prior to P3 affects dendritic targeting, ChatmG-labeled single SACs were identified in retinal cross-sections from Chat-Megf10-cKO and ChatCre; Megf10flox/+ control mice, as described above. Analysis was performed at P1 and P3; data in Figure 7H is from P3 only. All mGFP+ SACs were first scored as to whether they expressed MEGF10 protein (see Figure 7F,G). Subsequently, each cell was scored for soma-layer projection as described above for wild-type and Megf10–/– animals. This scoring was done blind to the cell’s MEGF10 expression status. The fraction of cells classified as either ‘soma-projecting’ or ‘IPL-only’ was calculated for MEGF10+ SACs, MEGF10– SACs, and littermate control SACs (Figure 7H). Sample sizes: n = 26 OFF, 18 ON cells from controls; 24 OFF, 19 ON MEGF10+ cells from Chat-Megf10-cKO; 9 OFF, 17 ON MEGF10– cells from Chat-Megf10-cKO.
To assess SAC stratification at maturity, cross-sections from P17 Chat-Megf10-cKO and littermate controls were stained for anti-ChAT. Four mutants and three littermate controls, from two litters, were examined.
Analysis of Six3-Megf10-cKO conditional mutants
Characterization of Cre recombination patterns
Request a detailed protocolBreeders carrying the relevant alleles were interbred to generate Six3-Cre; Megf10flox/lacZ (Six3 Megf10cKO) mice and littermate controls (Six3-Cre; Megf10+/lacZ or Cre– Megf10flox/lacZ). As noted above in Ptf1a section, Cre is expressed very early (~E9.5) in Six3-Cre retina, but expression is incomplete, with some parts of peripheral retina spared from Cre activity (Furuta et al., 2000). Therefore, all mice used for these experiments also carried the Rosa26GFPf Cre reporter, to reveal retinal regions that either lacked MEGF10 (GFP+ cells) or were spared from MEGF10 deletion (GFP– cells). Anti-MEGF10 staining confirmed that the GFP Cre reporter is a reliable marker of MEGF10 expression status (Figure 7—figure supplement 1).
Assessment of morphological phenotypes
Request a detailed protocolFor quantification of INL projection frequency at P2, Six3 Megf10cKO and littermate control whole-mount retinas were stained for βgal, Sox2, and anti-GFP. This staining marked SACs (Sox2 and βgal), revealed their dendritic morphology (βgal), and defined their MEGF10 expression status (GFP). Confocal stacks were acquired through the INL, extending to the IPL (which was clearly discernable due to dense βgal and GFP expression). The INL was defined as the region above this in the image stack, containing Sox2+ neurons. Cells that projected into the INL were clearly discernable due to their multipolar morphology with numerous dendritic protrusions (e.g. Figure 7B). Cells that did not project to the INL had a round morphology with only minor lateral branches less than one cell radius in length (Figure 7C). Each βgal-labeled SAC was scored as to whether it expressed GFP, and whether it projected lateral arbors into the INL. If the cell had only INL branches directed toward the IPL through the stack Z-plane, it was not counted as INL-projecting. Scoring was done in separate sessions so that the scorer was blind to GFP expression status when determining INL projections. Sample sizes: n = 117 SACs from two control mice; n = 302 GFP+ SACs and 149 GFP– SACs from 2 Six3-Megf10-cKO mice.
To assess SAC stratification in cross-sections, P2, P4, or P17 Six3-Megf10-cKO and littermate control retinas were sectioned and stained for anti-βgal (P2) or anti-ChAT (P17). The number of animals examined was: P2, four mutants, two controls; P4, two mutants, three controls; P17, two mutants, two controls.
Quantification of area covered by SAC dendritic arbors
Neonatal individual SAC arbor territory
Request a detailed protocolP0 ChatmG retinas were imaged in whole-mount preparations stained with anti-Sox2 and anti-GFP antibodies to identify single GFP+ SACs. To avoid cell selection biases, all labeled SACs with arbors that were clearly distinguishable from their neighbors were imaged and analyzed, except for far-peripheral cells that may have been damaged during mounting. At least four animals were imaged for each genotype. Z stacks were acquired through the GCL, IPL, and INL to encompass all arbors of a single cell. Images were imported into ImageJ, z-projected into a single plane, and polygons were drawn connecting the dendritic tips, nearest neighbor to nearest neighbor, until the dendritic field was captured. Area of this polygon was calculated using ImageJ. Sample sizes: OFF SACs, n = 16 wild-type and 16 Megf10–/–; ON SACs, n = 31 wild-type and 34 Megf10–/–. Statistics: two-tailed t-tests.
Adult individual SAC arbor territory
Request a detailed protocolIndividual SACs were labeled by injection of ChatCre mice with ‘Brainbow’ Adeno-associated virus (AAV) driving fluorophore expression in a Cre-dependent manner (Cai et al., 2013). The two Brainbow AAV9 viruses, encoding farnesylated fluorescent proteins that are targeted to the plasma membrane (University of Pennsylvania Vector Core), were mixed to 1.5 × 1012 genome copies per mL. Adult mice (P40-50) were anesthetized with ketamine-xylazine by intraperitoneal injection. Propraracaine hydrochloride (0.5%) ophthalmic solution (Akorn, Lake Forest, IL) was applied to the eye to provide local anesthesia. A 30 1/2G needle was used to make a small opening near the ora serrata, and 1 µl of virus was injected with a 33G blunt-ended Hamilton syringe intravitreally. Tissue was collected 3 weeks after the virus injection.
Retinas were stained in whole-mount with anti-GFP, anti-mCherry, and anti-mKate antibodies to reveal SACs. OFF SACs were not labeled in large numbers, so analysis was restricted to more abundantly labeled ON SACs. Imaging, image processing, and quantification were as for P0, except that only SACs in central and mid-peripheral retina were used to avoid confounding effects of eccentricity on arbor size. Sample sizes: n = 10 wild-type and 16 Megf10 mutant SACs.
ChAT arbor plexus retinal coverage
Request a detailed protocolEn-face images of adult (P46) ON or OFF SAC plexus were obtained from confocal Z-stacks (0.4 µm Z resolution) by performing maximum-intensity Z-projections of 2–4 optical slices encompassing the relevant layer. Using ImageJ, these images were then thresholded, converted to binary, and the percentage of the field of view covered by ChAT-positive arbors was calculated. All image stacks were obtained from central or mid-peripheral retina. Sample size: 9 fields of view from 2 Megf10 mutants and two heterozygous littermate controls were used to calculate average coverage for each genotype. Percent change is reported in the Results; total retinal coverage was as follows: Control ON, 65.9 ± 1.3%; control OFF, 70.6 ± 3.6%; mutant ON 56.9 ± 3.8%; mutant OFF, 60.0 ± 3.7% (mean ±S.D).
Hb9-GFP stratification
Request a detailed protocolP1-P2 retinas carrying Megf10lacZ and Hb9-GFP were co-stained for βgal and GFP. RGCs with dendrites that co-fasciculated with βgal-positive IPL strata were counted. Cells that projected to βgal-positive regions, but also filled non-SAC-projecting IPL regions, were not counted as co-fasciculated. To judge co-fasciculation, we used two criteria: 1) inspection of dendrite anatomy across the confocal stack; 2) fluorescence profiles of GFP and βgal channels across IPL (see next section below). Examples of cells falling into each category are provided in Figure 1 and Figure 1—figure supplement 2. See Results for sample sizes.
Quantitative assessment of IPL stratification level
Request a detailed protocolImages of retinal cross sections were processed in ImageJ. A vertical ROI (12.5 µm wide) was drawn to perpendicularly bisect the IPL strata, from the edge of the INL to the edge of the GCL. IPL stratification levels were reported as percentage of IPL width. Intensity was calculated for each pixel along the length of the ROI as an average across its width. Background (minimum pixel value) was subtracted; then, all pixel intensity values were normalized to the maximum value of that ROI. Location of fluorescent peaks was calculated as the pixel with maximum intensity; if multiple pixels had the same intensity the peak was defined as the center of the plateau. The procedure was typically performed on single confocal optical sections, but for some P1-2 cells, which have much smaller arbors, it was necessary to use a maximum-intensity projection of a small number of slices in order to fully capture dendrite morphology.
For BC5-BC7 arbor distance measurements (Figure 9F), distances as percentage of total IPL width were compared by one-way ANOVA/Tukey’s post-hoc test. n = 14 measurements from two control mice; n = 7 normal IPLs, 11 SAC clumps, 11 SAC gaps from 3 Megf10–/– mice.
Generation of Megf10-ΔICD constructs
Request a detailed protocolThe MEGF10-ΔICD-GFP construct was reported previously (Kay et al., 2012), which was originally made from pUbC-MEGF10-GFP (Addgene #40207). It encodes a version of MEGF10 in which the cytoplasmic domain is truncated after the 9th amino acid and replaced by GFP. Inclusion of those nine amino acids was necessary to achieve plasma membrane localization. For this study, it was subcloned into the pEGFPN3 plasmid, containing the CMV promoter, to make pCMV-MEGF10-ΔICD-GFP.
To make the MEGF10-ΔICD-Flag construct, Megf10 (truncated after the 9th intracellular domain amino acid as above) was PCR amplified from pUbC-MEGF10-GFP vector using M10flag_Fwd forward primer and Cyto9_flag_Rev1 reverse primer. Resulting PCR products were digested with NotI and AscI restriction enzymes and ligation cloned into pEGFPN3 vector linearized with corresponding restriction enzymes.
Assay for interaction of MEGF10-ΔICD constructs
Co-immunoprecipitation
Request a detailed protocolHEK293T cells were grown to 80% confluency. Cells were then transfected using a linear polyethylenimine (PEI) transfection reagent: DNA, PEI, and Opti-MEM were mixed in a 1:3:30 ratio and incubated for 10 min at room temperature then applied to confluent cells. Cells were harvested 48 hr post-transfection. Cells were lysed with NP-40 lysis buffer (1% NP-40, 150 mM NaCl, 50 mM Tris-Cl, and 1X proteinase inhibitor) by pipetting. Lysate was centrifuged at 14,000 x g at 4°C for 15 min. to remove insoluble material. The soluble protein fraction was quantified with Bio-Rad DC assay. For immunoprecipitation, 500 µl (1 µg/ µl) protein in NP-40 buffer lysis buffer was incubated overnight at 4°C with antibody (1 µl of chicken anti-GFP or 2 µl of mouse anti-Flag). Protein G Dynabeads (10 µl) were added to mixture for 1 hr at 4°C while rotating. Beads were sequestered by magnet and flow-through was removed. Beads were washed with 500 µl lysis buffer (3x) on ice then eluted with 30 µl 2X Laemmli containing 5% β-mercaptoethanol.
Western blot
Request a detailed protocolSamples were prepared in 2X Laemmli sample buffer, heated at 95°C for 10 min, and loaded onto SDS-acrylamide gel (running gel: 8% acrylamide/bis Tris-HCl with 0.1% SDS pH 8.8; stacking gel: 5% acrylamide pH 6.8; cross linked with TEMED and APS). Precision Plus Protein Dual Color Standards (BioRad) were used as a molecular weight marker. The gel was run on a BioRad mini gel running apparatus with SDS-PAGE running buffer (25 mM Tris, 192 mM glycine, 0.1% SDS). Electrophoresis was carried out at 50 V through the stacking gel then adjusted to 120 V until the dye front reached the lower end of the gel. BioRad Immobilon-FL PVDF membrane and Whatman filter paper were used with the BioRad mini cassette for transfer. Samples were transferred in 25 mM Tris, 192 mM glycine, 20% methanol at 100 V for 90 min. Membranes were blocked with PBS/Odyssey blocking buffer and stained with chicken anti-GFP 1:20000, mouse anti FLAG 1:20,000 overnight at 4°C with shaking. After washing with PBST for four times, membranes were stained with 1:20,000 secondary antibodies for 1 hr at room temperature. The membranes were washed with PBST four times and then rinsed with PBS and water. Finally, the membranes were imaged with LI-COR Odyssey using the Image Studio software.
Quantification of mosaic spacing phenotypes and their effects on SAC IPL projections
Regularity index
Request a detailed protocolRegularity of SAC cell body distribution in Six3-Megf10-cKO, Chat-Megf10-cKO, and littermate control mice was calculated as previously described (Kay et al., 2012). The Voronoi domain regularity index (VDRI) was used as a measure of regularity. It is calculated by first assigning a Voronoi domain to each cell in an array (Figure 9—figure supplement 1), and then calculating the mean and standard deviation of the domain areas. The VDRI is defined as the mean area divided by the standard deviation. Arrays that are less regularly distributed will have a lower VDRI because their domain sizes are more variable (and hence have a higher standard deviation).
P17 whole-mount retinas were stained with an antibody to ChAT and imaged en face. One eye was processed from each animal used in the experiment. For each eye, three confocal image stacks were obtained using a 20x objective (636.5 µm2 field of view). Images of INL SACs were analyzed using Fiji software. The location of each SAC in the field of view was marked; this information was used to count the number of SACs (Figure 9B) as well as define Voronoi domains belonging to each cell, using Fiji functions. The area of each Voronoi domain (excluding edges) was calculated in Fiji.
For statistical analysis of regularity effects across genotypes, we first calculated the per-animal average cell density and VDRI from the three acquired images. Differences between genotypes were then evaluated using one-way ANOVA and Fisher’s PLSD. Previously published Megf10 null and simulation data was also included for comparison (Kay et al., 2012). The simulations define the VDRI that would be expected for a randomly-arranged array of cells matched in size and density to real SACs. Data collection and analysis was virtually the same as in the previous study, allowing us to include these data in our statistical comparisons.
Effects of soma position upon IPL errors: single-cell analysis
Request a detailed protocolTo ask if soma position correlates with IPL errors, we first defined the ectopic projection status of each OFF SAC in a set of Z-stacks acquired from ChAT-immunostained retinal whole-mounts. Sample sizes: n = 515 cells from two control (Megf10+/–) mice; n = 584 cells from 2 Megf10 mutant mice. The Z-stacks encompassed, at different levels of the stack, SAC somata in the INL and their ramified arbors in the IPL. In Megf10 mutants, the OFF ectopic IPL arbor network and the typical OFF DS circuit sublayer were identified at different stack levels (Figure 8C,E). ChAT+ arbors arising from individual OFF SAC somata were traced through the stack to identify those that joined into the ectopic network. The fraction of SACs that did so was then calculated and plotted in Figure 8F. For Figure 8G, we further examined these stacks to look for SACs that made ectopic projections at the INL level.
Next, we defined the severity of mosaic spacing perturbations in the local neighborhood of each SAC. Because SAC position is random in Megf10 mutants, SACs might be more crowded or more isolated from their neighbors than in controls; or, by chance, they might be located at a fairly normal distance from their neighbors. The size of a cell’s Voronoi domain is influenced by the distance of all nearest neighbors (Figure 9—figure supplement 1), and therefore serves as a convenient measure of local cell density. For simplicity we refer to Voronoi domains as ‘territory size’ in Figure 9—figure supplement 1. The effect of local cell density upon IPL projection errors was determined by plotting the ectopic error rate for each 100 µm2 territory size bin (Figure 9—figure supplement 1). Sample size per bin, in order from smallest (<200 µm2) to largest (>1100 µm2): n = 32, 65, 89, 102, 91, 80, 30, 39, 24, 34.
Multielectrode array recordings
Isolation of retina, recording, and spike sorting
Request a detailed protocolTwo wild-type and two Megf10–/– animals were used for multielectrode array (MEA) recordings. Immediately following euthanasia, retinas were isolated under infrared (IR,>900 nm) illumination with the assistance of IR-to-visual converters. This preserved the photosensitivity of the retina during the dissection. Dissections were performed in sodium bicarbonate-buffered Ames’ solution (Sigma, St. Louis, MO) equilibrated with 5% CO2 +95% O2 to pH 7.4 and maintained at 32–34° C. Hemisection of the eye was performed along the ora serrata by first making a small incision, following which the vitreous was removed and the retina was isolated from the pigment epithelium and eye cup. A piece of dorsal retina (1–2 mm2) was dissected and placed RGC-side down on the planar MEA.
The MEA consisted of 519 electrodes with 30 μm inter-electrode spacing, covering a hexagonal region with 450 μm on a side (Field et al., 2010). The voltage on each electrode was digitized at 20 kHz and stored for post-hoc analysis. Details of recording methods and spike sorting have been described previously (Field et al., 2007). Spikes were identified using a threshold of four times the voltage standard deviation on each electrode. Principal component analysis applied to the ensemble of spike waveforms measured on each electrode provided a subspace for clustering spikes according to their shape. A Gaussian mixture model was used to cluster the spikes originating from individual RGCs. The clusters were manually inspected for each identified ooDSGC to ensure the spike waveforms were well isolated from other simultaneously recorded RGCs and all spikes were captured within each cluster. When a single cluster of spikes was captured by more than one Gaussian or when a single Gaussian included spikes from more than one cluster, the clustering was manually adjusted to generate a new set of initial conditions for re-fitting the mixture of Gaussians. Spike clusters with >10% estimated contamination based on refractory period violations, or spike rates < 1 Hz, were excluded from further analysis.
Visual stimulation and RGC responses
Request a detailed protocolVisual stimuli were focused on the photoreceptor outer segment, from an OLED display (Emagin, Inc.) with 60.35 Hz refresh rate. The mean intensity of the stimulus was 7000 photoisomerizations per rod per s, or 5000 photoisomerizations per cone per s for a cone containing all M-opsin. These estimates do not account for the effect of pigment self-screening. To measure the direction tuning of ooDSGCs as a function of contrast, a positive contrast bar (1200 µm wide) was presented on a gray background (Figure 12B). On each presentation, the bar moved in one of 12 equally spaced directions at 400 µm/s and was presented at one of the following (Weber) contrasts: 5%, 10%, 20%, 40%, 80%, 150% and 300%. Responses to a total of 8 trials were collected for every condition; stimulus conditions were presented pseudo randomly. Spike times were binned at 1 ms resolution for all subsequent analyses.
To distinguish DSGCs from other RGCs recorded on the MEA, square-wave drifting gratings were used. These gratings drifted in one of 12 different and equally spaced directions and at two different speeds (225 μm/s and 900 μm/s; spatial period 400 μm/cycle). DSGCs were identified based on their direction selectivity index (DSI) defined as:
calculated from responses to drifting gratings and moving bars. Here, is the number of spikes elicited to stimulus movement along the direction defined by the vector .
The distribution of DSIs across all recorded RGCs was bimodal, with DSGCs forming the high mode (Figure 12A). Based on these distributions, a DSI of 0.25 reliably identified DSGCs in wild-type and Megf10–/– retinas. ooDSGCs were isolated from ON DSGCs by their distinct ON and OFF responses to a bar entering and exiting the receptive field (Figure 12B). The total ooDSGC sample size obtained by this procedure was n = 80 from the two wild-type and n = 74 from the two Megf10–/– retinas. The paired Kolmogorov-Smirnov (KS) test was used to compare cumulative probability distributions from these two populations.
Analysis of ooDSGC response
Measurement of direction tuning width
Request a detailed protocolFirst, the direction tuning curve for each ooDSGC was obtained by calculating the number of spikes elicited across all trials for each direction of bar movement. Due to the circular nature of the data, the direction tuning curve was treated as circular normal distribution, also called von Mises distribution (Oesch et al., 2005), and the tuning width was measured as the circular standard deviation (, defined by
where is the second moment of the von Mises distribution:
This yielded a nonparametric estimate of the tuning curve width.
Measurement of direction tuning strength
Request a detailed protocolTo measure the strength of tuning, the difference between spike counts to motion in the preferred and null directions was normalized by the sum of these responses. The tuning curves were sampled at 30 degree intervals. To estimate the response in the preferred (null) direction, which could fall between sampled directions, a cosine-weighted average of the two strongest (weakest) responses was calculated. This yielded the following equation for measuring tuning strength:
where the summation is performed over the responses weighted by the cosine terms for the two nearest neighbor movement directions around the preferred direction and the null direction . This resulting index for tuning strength varied between zero and unity.
Measurement of direction tuning similarity between ON and OFF responses
Request a detailed protocolTo separately analyze the ON and OFF responses of ooDSGCs, we first defined temporal windows for each ooDSGC that distinctly separate the ON and OFF responses. This was done by passing high-contrast moving bars (150% and 300% contrast) through the receptive field. In the resulting spike rasters, ON and OFF response phases were clearly discernible (Figure 12B,F). The boundary for separating the ON and OFF responses was set halfway between the peak ON and OFF spike rate locations Figure 12F,G). Once the temporal boundary was defined, the preferred direction was calculated independently for the ON and OFF responses for each ooDSGC. The same ON-OFF temporal boundaries were used for all contrasts shown in Figure 12—figure supplement 1. The difference between the preferred directions, Δϕ, quantified the angular difference between the ON and OFF preferred directions (Figure 12F).
Analysis of ooDSGC subtypes
Request a detailed protocolooDSGC subtype classification was performed using the K-means clustering algorithm. This was done by first assigning a set of four initial seed values corresponding to the four cardinal directions of ooDSGCs (Oyster and Barlow, 1967). Next, the angular difference between the seed values (for first iteration) or the cluster means (for later iterations), and the preferred directions of each ooDSGCs was calculated. The cluster for which the angular difference was minimum was the cluster to which an ooDSGC was assigned. This yielded the four subpopulations of ooDSGCs described in Figure 12—figure supplement 1.
Quantification and statistical analysis
Request a detailed protocolStatistical analysis was performed using GraphPad Prism software (anatomy/development studies) or using custom JAVA-based software and MATLAB software (physiology studies). This software is available, together with the primary data it was written to analyze, at a public repository (Roy and Field, 2017; https://github.com/Field-Lab/megf10-dstuning; copy archived at https://github.com/elifesciences-publications/megf10-dstuning). Statistical tests used for each experiment are given in the appropriate Materials and methods section above, and/or in the figure legends. Sample sizes for each experiment are given in the appropriate Methods section above or else in the Results. p-Values (α = 0.05) are given in figure legends, or in the Results if no figure is shown. Error bars are defined in figure legends. Exact p-values are reported unless the value was less than 1.0 × 10−7.
References
-
Inhibition in ON-OFF directionally selective ganglion cells of the rabbit retinaJournal of Neurophysiology 69:2174–2187.https://doi.org/10.1152/jn.1993.69.6.2174
-
Synaptic laminae in the visual system: molecular mechanisms forming layers of perceptionAnnual Review of Cell and Developmental Biology 29:385–416.https://doi.org/10.1146/annurev-cellbio-101011-155748
-
Improved tools for the Brainbow toolboxNature Methods 10:540–547.https://doi.org/10.1038/nmeth.2450
-
Receptive field properties of bipolar cell axon terminals in direction-selective sublaminas of the mouse retinaJournal of Neurophysiology 112:1950–1962.https://doi.org/10.1152/jn.00283.2014
-
Molecules and mechanisms that regulate multipolar migration in the intermediate zoneFrontiers in Cellular Neuroscience 8:386.https://doi.org/10.3389/fncel.2014.00386
-
Spatial properties and functional organization of small bistratified ganglion cells in primate retinaJournal of Neuroscience 27:13261–13272.https://doi.org/10.1523/JNEUROSCI.3437-07.2007
-
Visual coding with a population of direction-selective neuronsJournal of Neurophysiology 114:2485–2499.https://doi.org/10.1152/jn.00919.2014
-
Mosaics of islet-1-expressing amacrine cells assembled by short-range cellular interactionsJournal of Neuroscience 17:7831–7838.https://doi.org/10.1523/JNEUROSCI.17-20-07831.1997
-
Early development of amacrine cells in the mouse retina: an electron microscopic, serial section analysisThe Journal of Comparative Neurology 179:277–300.https://doi.org/10.1002/cne.901790204
-
Sox11 balances dendritic morphogenesis with neuronal migration in the developing cerebral cortexJournal of Neuroscience 36:5775–5784.https://doi.org/10.1523/JNEUROSCI.3250-15.2016
-
The New Visual NeurosciencesDevelopment of retinal arbors and synapses, The New Visual Neurosciences, Cambridge, MIT Press.
-
Development of starburst cholinergic amacrine cells in the retina of Tupaia belangeriThe Journal of Comparative Neurology 502:584–597.https://doi.org/10.1002/cne.21324
-
Development of dendritic form and functionAnnual Review of Cell and Developmental Biology 31:741–777.https://doi.org/10.1146/annurev-cellbio-100913-013020
-
Two modes of radial migration in early development of the cerebral cortexNature Neuroscience 4:143–150.https://doi.org/10.1038/83967
-
Many paths to synaptic specificityAnnual Review of Cell and Developmental Biology 25:161–195.https://doi.org/10.1146/annurev.cellbio.24.110707.175402
-
Fiji: an open-source platform for biological-image analysisNature Methods 9:676–682.https://doi.org/10.1038/nmeth.2019
-
Developmental relationship between cholinergic amacrine cell processes and ganglion cell dendrites of the mouse retinaThe Journal of Comparative Neurology 456:154–166.https://doi.org/10.1002/cne.10509
-
Diverse synaptic mechanisms generate direction selectivity in the rabbit retinaJournal of Neuroscience 22:7712–7720.https://doi.org/10.1523/JNEUROSCI.22-17-07712.2002
-
Direction selectivity in the retina: symmetry and asymmetry in structure and functionNature Reviews Neuroscience 13:194–208.https://doi.org/10.1038/nrn3165
-
Anatomy and spatial organization of Müller glia in mouse retinaJournal of Comparative Neurology 525:1759–1777.https://doi.org/10.1002/cne.24153
-
Sox2 regulates cholinergic amacrine cell positioning and dendritic stratification in the retinaJournal of Neuroscience 34:10109–10121.https://doi.org/10.1523/JNEUROSCI.0415-14.2014
-
Cone contacts, mosaics, and territories of bipolar cells in the mouse retinaJournal of Neuroscience 29:106–117.https://doi.org/10.1523/JNEUROSCI.4442-08.2009
-
Isl1Cre reveals a common Bmp pathway in heart and limb developmentDevelopment 133:1575–1585.https://doi.org/10.1242/dev.02322
-
Cell type-specific changes in retinal ganglion cell function induced by rod death and cone reorganization in ratsJournal of Neurophysiology 118:434–454.https://doi.org/10.1152/jn.00826.2016
Decision letter
-
Jeremy NathansReviewing Editor; Johns Hopkins University School of Medicine, United States
In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.
Thank you for submitting your manuscript "Formation of retinal direction-selective circuitry initiated by starburst amacrine cell homotypic contact” to eLife. Three experts reviewed your manuscript, and their assessments, together with my own, form the basis of this letter. As you will see, all of the reviewers were impressed with the importance and novelty of your work.
I am including the three reviews at the end of this letter, as there are a variety of specific and useful suggestions in them. In the written reviews and in the subsequent discussion among the reviewers, the general points that emerged are that (1) additional data and/or data quantification would improve the manuscript at various points, (2) some of the interpretations seem to be too strong and should be more nuanced, (3) there is a massive amount of data and analysis for the reader to digest and anything that you can do to help the reader in this regard would be welcome and will increase the impact of the study, and (4) we feel that the manuscript would benefit from a clear focus on the importance of homotypic neurite interactions mediated by megf10 in (the rate of) sublayer formation, rather than stretching to highlight more speculative concepts.
Reviewer #1:
The authors use the development of starburst amacrine cells in the mouse retina as a model for understanding how the laminar organization of the retina arises during development. This is an excellent model system since this cell type is thought to be one of the earliest organizers of the inner-plexiform layer (IPL), a structure within the retina in which specific cell types arborize to form synaptic connections with appropriate partners.
The approaches and analyses used in this paper are quite powerful. The authors use a wide array of transgenic mice to establish several novel findings regarding the development of SACs. For example, they take advantage of a ChAT-Cre mouse, which has weak expression early in development, or Islet1-Cre mice which label SACs, crossed with a membrane-GFP reporter to obtain images of individual SACs during many stages of development, including migration, during which they have both trialing edge and leading edge processes (around E16), prior to innervating the IPL when they make transient contacts with nearby SACs (around P0), and finally then innervation (P2). Not only are these images stunning, but also, they are quite informative, leading the authors to hypothesize that those transient contacts with neighboring SACs are critical for normal lamination. To further motivate this hypothesis, the authors show that the protein internexin – which establishes polarity -- is oriented toward other SACs while migrating and then turn toward the IPL when they innervate the IPL.
To test this hypothesis the authors again use several approaches. First, they prevent expression of a transcription factor involved of differentiation of amacrine cells in central retina and sparsely in peripheral retina to establish areas of retina where some SACs are isolated, meaning they are so sparse that they cannot make homotypic contacts with other SACs. These SACs maintain an immature phenotype and do not innervate the IPL while maintaining processes in the cellular layers, something never seen in normal mice. They then do a variety of manipulations of the membrane protein Megf10-in some cases you have a SAC with normal levels of Megf10 surrounded by mutant SACs lacking Mef10 or vice versa – a mutant SAC surrounded by normal SACs. In both cases, if a SAC is not making signaling with its neighbors via Megf10, then it has the same phenotype as a "isolated SAC".
(Note authors also do nice control experiments to show that the lamination phenotypes cannot be explained by the disruption of SAC mosaic that Kay previously described in Megf10 mutant mice).
The authors then go on to determine the impact of this early disruption of the lamination on the adult. Surprisingly, there appears to be some recovery, in that many of the mutant or isolated SACs do form lamination in the correct place but in addition, they have ectopic branches sometimes that venture into the INL. In one of the most amazing aspects of this study, the authors show that the pre- and postsynaptic targets of the SACs actually change their lamination – essentially keeping their relative associations with the SAC processes, even when those processes are in the wrong location.
Finally, the authors show the impact of this rearrangements on function and find that there is a broadening of direction selective tuning. Interestingly, they find that one of the phenotypes is that ON vs. OFF direction preference are more weakly aligned in Megf10 mutant mice than in WT.
This study is extremely well done – as noted already multiple approaches are used for all experiments. The authors do an excellent job quantifying their results and they are quite conservative in their interpretations. Every "i" is dotted and every "t" is crossed. This is surely going to be a definitive study not only on starburst amacrine cell development but also the role that this cell plays in organizing the entire direction selective circuit. The authors are to be commended.
1) For all figures based on bar plots and SEM error bars (example, Figure 8B), the authors need to present the variance some way. Either the authors need to show individual data points, or they need to use standard deviation for error bars. Example Figure8B
2) Subsection “Timing of DS circuit IPL sublayer formation”: Authors need to clarify how they determined "percent stratified". For example, they state only "30% of ooDSGCs co-fasciculated with SAC arbors". Is this percent of dendrites or percent of cells? What qualified as "co-fasciculated"?
3) Figure 2J-L – purple color should labelled "β-gal" rather than "SAC" for consistency with other figures (e.g. Figure 5).
Reviewer #2
Circuits in the retina are organized such that pre and postsynaptic cells elaborate and confine their processes and synaptic connections within specific sublayers of the neuropil. The sequence of early developmental steps giving rise to these lamination patterns and the molecular cues that are involved are not well understood. Ray et al., carried out an extensive set of experiments capitalizing on genetic tools, imaging and electrophysiology to fill this knowledge gap. They focused on a well characterized retinal circuit, involving inhibitory starburst (SAC) amacrine cells and their postsynaptic partners, the direction-selective (DS) ganglion cells. The authors propose that homotypic interactions direct the processes of SACs at the end stages of their migration to ramify in the IPL. They show convincingly that Megf10-mediated interactions are needed for timely SAC lamination, and if disrupted, leads to broader tuning of direction-selective responses in the DS ganglion cells. Overall, the data is of high quality, quantification and statistical analyses are provided and many findings are supported by the data presented. However, in some places, the conclusions are over-stated; such claims need to be substantiated by further analysis.
1) A major conclusion is that SAC processes make homotypic contact with neighboring SAC cell bodies – this contact is what drives elaboration of SAC processes in the IPL.
The evidence for homotypic contact is not strong. It is difficult to gain a good sense of what is happening based on the images, especially from the frozen sections. Figure 1J (an en face view) provides a bit more support but further analysis is necessary. At the minimum, using en face image stacks of the labeled SACs, the authors should perform a control by flipping one confocal channel and then re-measuring overlap between the channels, and demonstrate that the overlap is above chance. To confirm that there is contact, electron microscopy (EM) is necessary. This is because the diffraction limit of light does not enable one to conclude from two labeled structures alone that there is 'contact'. If the authors perform the overlap measurements with the control, then I think it is fine to conclude that there is an association that is not random, and these are presumed to be contacts.
2) '[…]deletion of Megf10 causes a profound impairment of IPL-directed SAC dendrite growth, preventing timely sublayer formation'.
It is clear that the processes of SACs, especially the OFF SCAs, in the Megf10 KO have not reached or contribute to a singular plexus in the IPL by P3 (although they do). But at P3 in the mutant, the majority of the processes appear directed towards the IPL (Figure 5B). It would be good to clarify whether the authors mean that Megf10 is important for the rate at which the SAC process lamination occurs, or they also mean, the direction of process outgrowth.
3) SAC-BC repulsion regulates bipolar cell axon lamination.
This isn't so clear to me because it is difficult to come to this conclusion based on the images. The small displacements in BC5+7 may simply be because where SACs occupy space, BC5+7 cannot elaborate their axons because the SAC process network is very dense. Where there is a gap in the SAC plexus, these BC axon terminals do come closer together, and this could be because space is now available rather than repulsion is absent.
Results section:
1) Throughout the text, the authors make statements about the behavior of the SACs, which are not strongly supported by the images. Here are some examples:
i) Subsection “Homotypic contact is required for SAC IPL innervation and dendrite lamination”: "Misprojecting DACs are still in contact with numerous other amacrine[…]" and legend for Figure 3—figure supplement 1G: "SACs that made errors had extensive interactions with GAD65+ amacrine cells". It not possible to infer 'contact' or 'interactions' from these data. Please explain.
ii)Subsection “SAC IPL errors induce laminar targeting errors by their DS circuit partners”: 'No changes were seen in Syt2-labeled BC6, suggesting a specific effect on the bipolar cells that make extensive contact with SACs.' The BC6 axons are much further away from the SAC ON plexus compared to the BC7 axon terminals. Thus, I am not sure one can readily conclude that 'DS-circuit bipolar cells respond to SAC attractive cues'.
2) The authors have really put in an enormous effort to obtain a comprehensive study. There is a lot of data and the supplements that complement the Figures are important. But I wonder if the authors would consider streamlining some parts because it is really a bit overwhelming for the reader to try and assimilate all the information relevant to each major conclusion by going back and forth from the main figures to the supplements. For example, the various effects on bipolar cells in the Megf10-/- retina; it would be helpful to at least provide the summary in a single figure for the bipolar cell types analyzed (BC2,3a,5,6,7).
3) The mistargeting of BC2 in Megf10-/- to where there are ectopically located SAC processes is intriguing and clear that there is a bipolar axon terminal lamination defect associated with abnormally placed SAC processes (but for a BC type that is not part of the SAC-DS circuit?). However, to establish a causal link to SACs, it would help to show that lamination of other retinal neurons that are associated with BC2 bipolar cells (if that is known) are unperturbed in their lamination (i.e. are the BC2's following their targets?).
4) Figure 2—figure supplement 2: There should be quantitation for the E16 observations to substantiate SACs projecting to each other. The neuroblastic layer is dense with nuclei – where the processes appear to 'project', they seem to be just coursing around the nuclei.
5) Are SACs not expected to be labeled in the Gad1-GFP line?
6) It would be helpful to indicate in all the figures whether we are looking at a single image plane or a z-projection of several optical planes. This will help the reader gain a sense of just how 'close' the labeled structures are. What was the z-dimension step size for the confocal images on the wholemounts (not sure I found this in the Materials and methods section).
7) Where possible, labeling the borders of the IPL would be helpful- presumably these are defined always by labeling of cell nuclei?
8) OFF SACs show more dramatic ectopic (somatic) processes than ON SCs in the Megf10 mutant. But this defect does not seem to have a physiological correlate – or does it?
9) Gad1-GFP line should be listed in the transgenics used.
10) Has the expression of cre in the Chat-cre line been documented somewhere? This is important with respect to separating the early and late deletion of Megf10.
11) Figure 8—figure supplement 1: Please provide a more detailed explanation of how the 2D cross-correlation analysis was performed.
12) Although it does appear that Meg10 plays two roles- one in somal mosaic positioning and the other in lamination of the processes of SACs, it isn't clear that timing alone separates these roles. If mosaic arrangements are due to repulsive SAC-SAC interactions, then fasciculation of the SAC processes must be an adhesion-based, perhaps even 'attractive' mechanism. Is it known if Megfs have other downstream pathways that are not 'repulsive'?
13) Could the authors elaborate a bit more in the Discussion section on how the lamination of the OFF and ON SAC plexuses become better confined to a single lamina (with exceptions) by adulthood?
Reviewer #3
This study by the Kay lab investigates the role of Megf10 in the development of starburst amacrine cells (SACs). In a previous study by Kay and colleagues (Kay et al., 2012), the mosaic spacing of SACs is impaired in Megf10 knockout mice. In this manuscript, Ray and colleagues examine the early developmental stages of SACs starting at E16. They found that Megf10 is required for the timely stratification of SAC neurites at the inner plexiform layer (IPL). In the absence of Megf10, SACs fail to eliminate ectopic neurites and unable to stratify at the correct IPL depth in the first few postnatal days (P0-P2). This developmental deficit leads to mistargeting and gaps of SAC neurites in the IPL in the mature retina. Furthermore, using elegant genetic tools, the authors demonstrate that Megf10 function requires homotypic interaction of megf10 expressed in neighboring SACs. Overall, this study advances the current understanding of sublayer formation in the retina. It also provides a missing link between the altered developmental process of SACs and the deficits in the mature retina of Megf10 mutants. My major concerns are about some of the interpretations of the data and analysis.
1) One developmental feature that is prominently mentioned in this study is a transient homotypic network of SAC dendrites in the inner nuclear layer (INL) between E16 and P1. The authors describe the neurites outside the IPL to be selectively targeted to the neighboring SAC somas and neurites in the INL. However, an alternative scenario is that SACs during this period have random growth of neurites to both INL and IPL. After P1, the INL neurites are selectively pruned, and the IPL neurites further grow and stratify in the correct IPL sublayer. I did not find sufficient evidence supporting the presence and functional significance of a homotypic network of SAC arbors at the INL. In particular,
- Subsection “Early SAC projections target neighboring SAC somata”: The authors describe SAC neurites in the INL as "soma-directed" neurites. It is unclear to me whether these neurites are "directed" at somas, or they are just randomly distributed in 3D. There is no quantification that shows "selective" somatic targeting over random growth. In fact, in both Figure 2E (E16) and Figure 2H (P0), SACs send more neurites to IPL than to INL.
- Subsection “Early SAC projections target neighboring SAC somata”: The authors claim that "P1 INL arbors selectively contacted somata or arbors of SAC neighbors". I have difficulty recognizing this pattern as "selective" since the INL at this stage is densely covered by SAC somas and ectopic neurites. The INL neurites may always end next to neighboring somas or neurites just because of the high density.
- Subsection “Early SAC projections target neighboring SAC somata”: The authors describe that On SACs send fine soma-directed branches from their IPL arbors (Figure 2—Figure supplement 2). Again, how do the authors distinguish "soma-directed" projection from random contacts due to imperfect stratification?
2) Subsection “Timing of DS circuit IPL sublayer formation”: The authors mention that SACs are stratified within the expected IPL sublayer, but ooDSGC dendrites are unstratified. This needs to be supported by quantifications. In Figure 1—figure supplement 1, panel C, both SACs and the ooDSGC appear to be unstratified in a similar way.
3) Subsection “Homotypic contact is required for SAC IPL innervation and dendrite lamination”: "the projections were typically more elaborate than those observed in wild-type retina". Please provide a quantitative analysis to support this claim.
4) Subsection “Requirement for MEGF10 in SAC IPL innervation and sublayer formation”: "this phenotype is not due to aberrant SAC migration.…" However, from Figure 5A and Figure 5B, Off SAC somas in MEGF10 mutants are farther away from IPL than those in the wild type. And there are significantly less On SACs in the mutant. This is consistent with delayed migration of SACs, which might also account for the delayed stratification phenotype in the mutants.
5) Subsection “Requirement for MEGF10 in SAC IPL innervation and sublayer formation”: "Rather, the absence of SAC sublayers was due to innervation of the soma layers instead of the IPL". I do not see a clear causal relationship here.
6) Related to point 1 above, subsection “Requirement for MEGF10 in SAC IPL innervation and sublayer formation”: To support the statement that P1 SACs in Megf10 mutants "preferentially" project neurites to the soma layer, the authors need to show that each SAC in Megf10 mutants sends more neurites in the INL than IPL.
7) Subsection” SAC errors persist to adulthood in Megf10 mutants”: The authors describe the laminar disruptions in mature retinas of Megf10 mutants occur "sporadically and at apparently arbitrary retinal locations." It is important to determine the percentage of IPL that is disrupted or shows ectopic stratification in Megf10 mutants. It is surprising that close to 100% SACs in Megf10 mutant mice show ectopic IPL projection in the adult (Figure 8F), but the laminar disruption shown in this manuscript (e.g. Figure 7A,B, Figure 8A and Figure9D) and the previous paper by Kay et al., 2012 seems to occur for a small fraction of the IPL length.
8) The authors use Six3-Megf10 cko and Chat-Megf10 cko to argue for separate time windows for laminar formation and somatic spacing. However, the data indicate that these two processes occur in overlapping temporal windows. First, irregular soma spacing with gaps in MEGF10 null retinas is very obvious in Figure 5A, indicating that mosaic spacing is already affected at P0. Second, earlier deletion of megf10 in Six3-cko mice causes more sever deficit in somatic spacing than in Chat-megf10 cko mice, indicating that somatic mosaics formation requires early megf10 signaling.
9) Please explain the rationale of the cross-correlation analysis in Figure 8—figure supplement 1. Why is there a strong correlation between the locations of SAC somas and the corresponding locations in the IPL layer? Given the morphology of SACs, it is unexpected to see such a correlation. In addition, statistical test needs to be performed for Figure 8—figure supplement 1C to claim a difference between mutant and control groups.
- Subsection “Early SAC projections target neighboring SAC somata”: Please clarify the statement "At P0-1, other IPL layers do not yet exist.…" Do the authors mean that there is no neurites in the IPL other than those of SACs?
What are the dashed lines of different depths in Figure 5E? How do the authors determine the border of IPL and INL given that SACs are the only amacrine cell type in early development? Delayed migration towards IPL and/or aberrant neurite growth can obscure the classification of IPL borders.
https://doi.org/10.7554/eLife.34241.033Author response
Reviewer #1:
The authors use the development of starburst amacrine cells in the mouse retina as a model for understanding how the laminar organization of the retina arises during development. This is an excellent model system since this cell type is thought to be one of the earliest organizers of the inner-plexiform layer (IPL), a structure within the retina in which specific cell types arborize to form synaptic connections with appropriate partners.
[…]
This study is extremely well done – as noted already multiple approaches are used for all experiments. The authors do an excellent job quantifying their results and they are quite conservative in their interpretations. Every "i" is dotted and every "t" is crossed. This is surely going to be a definitive study not only on starburst amacrine cell development but also the role that this cell plays in organizing the entire direction selective circuit. The authors are to be commended.
We thank the reviewer for these kind words!
1) For all figures based on bar plots and SEM error bars (example, Figure 8B), the authors need to present the variance some way. Either the authors need to show individual data points, or they need to use standard deviation for error bars.
We have added individual data points to these graphs as requested. See Figures 5C, Figure 8B, Figure 9B and Figure 11E.
2) Subsection “Timing of DS circuit IPL sublayer formation”: Authors need to clarify how they determined "percent stratified". For example, they state only "30% of ooDSGCs co-fasciculated with SAC arbors". Is this percent of dendrites or percent of cells? What qualified as "co-fasciculated"?
We scored the fraction of ooDSGC cells, not dendrites, that were co-fasciculated; this has been clarified in subsection “Timing of DS circuit IPL sublayer formation” and subsection “Hb9-GFP stratification”. We have also added further details on how ooDSGCs were scored as co-fasciculated (same subsection “Hb9-GFP stratification”). In response to reviewer 3, we added an additional step to the scoring procedure: We now have examined fluorescence profile plots across the IPL for the SAC and ooDSGC channels, in addition to examining anatomy in Z-stacks. These are shown in Figure 1E.
The paragraph in the Materials and methods section now reads:
“P1-P2 retinas carrying Megf10lacZ and Hb9-GFP were co-stained for βgal and GFP. RGCs with dendrites that co-fasciculated with βgal-positive IPL strata were counted. […] Examples of cells falling into each category are provided in Figure 1 and Figure 1—figure supplement 2.” (subsection “Hb9-GFP stratification”).
The fluorescence profile plots were not always dispositive for scoring stratification, because at P1-2 a profile line could not always be drawn to avoid SAC cell bodies within or abutting the IPL (which would give a false “SAC band”). For this reason, we relied on expert judgment combining both visual inspection and profile plot data. For transparency about our scoring system we provide examples of “stratified” and “diffuse” ooDSGCs in Figure 1—figure supplement 2.
3) Figure 2J-L – purple color should labelled "β-gal" rather than "SAC" for consistency with other figures (e.g. Figure 5).
This has been fixed.
Reviewer #2
Circuits in the retina are organized such that pre and postsynaptic cells elaborate and confine their processes and synaptic connections within specific sublayers of the neuropil. The sequence of early developmental steps giving rise to these lamination patterns and the molecular cues that are involved are not well understood. Ray et al., carried out an extensive set of experiments capitalizing on genetic tools, imaging and electrophysiology to fill this knowledge gap. They focused on a well characterized retinal circuit, involving inhibitory starburst (SAC) amacrine cells and their postsynaptic partners, the direction-selective (DS) ganglion cells. The authors propose that homotypic interactions direct the processes of SACs at the end stages of their migration to ramify in the IPL. They show convincingly that Megf10-mediated interactions are needed for timely SAC lamination, and if disrupted, leads to broader tuning of direction-selective responses in the DS ganglion cells. Overall, the data is of high quality, quantification and statistical analyses are provided and many findings are supported by the data presented. However, in some places, the conclusions are over-stated; such claims need to be substantiated by further analysis.
1) A major conclusion is that SAC processes make homotypic contact with neighboring SAC cell bodies – this contact is what drives elaboration of SAC processes in the IPL.
The evidence for homotypic contact is not strong. It is difficult to gain a good sense of what is happening based on the images, especially from the frozen sections. Figure 1J (an en face view) provides a bit more support but further analysis is necessary. At the minimum, using en face image stacks of the labeled SACs, the authors should perform a control by flipping one confocal channel and then re-measuring overlap between the channels, and demonstrate that the overlap is above chance.
We agree with the reviewer that claims concerning the homotypic nature of the SAC INL network required further evidence. We have added a new figure (Figure 2—figure supplement 3) to address this concern. This figure includes the channel-flipping analysis – we thank the reviewer for this suggestion. We found that the “random” rate of contact is about 15% (Figure 2—figure supplement 3). This is significantly lower than the 88.8% contact rate we observed in the real data; therefore, we conclude that the interaction between GFP+ individual SACs and the broader βGal+ SAC arbor network is likely to be selective and specific. Relevant Materials and methods section updates have been added to subsection “Characterization of SAC homotypic arbor network in soma layers”.
Also included in this new figure (Figure 2—figure supplement 3) is a more detailed documentation of the contacts sites. We provide an example of a putative contact between an individual GFP+ SAC arbor and an adjacent GFP-negative βGal+ arbor, showing in single optical planes, and in orthogonal views of 3D stacks, that the putative contact is apparent from all three perspectives. We now make clear in the Methods (subsection “Characterization of SAC homotypic arbor network in soma layers”) that this kind of analysis was performed during our quantification to confirm any contacts that were uncertain based on inspection of the individual Z slices.
For the main figure (Figure 2H,I), we chose to show a maximum-intensity projection of a larger Z volume encompassing the entire INL arbor. This is because the INL arbors often traversed many Z planes; therefore, only a maximum intensity projection gives the reader a real sense of the overall arbor morphology. However, as we now make clear in subsection “Characterization of SAC homotypic arbor network in soma layers” we did not use these Z-projections for quantification. Instead, we only counted putative contacts when they were confirmed in a single Z-plane, and/or with orthogonal views of the stack as illustrated in Figure 2—figure supplement 3.
To confirm that there is contact, electron microscopy (EM) is necessary. This is because the diffraction limit of light does not enable one to conclude from two labeled structures alone that there is 'contact'. If the authors perform the overlap measurements with the control, then I think it is fine to conclude that there is an association that is not random, and these are presumed to be contacts.
We agree that EM would be required to definitively demonstrate contacts, but we also agree with the reviewer’s comment that contacts are often inferred from light microscopy as long as appropriate anatomical analysis/controls are performed. We would argue that our revised manuscript, with 3D rotations and the cell flipping controls, now allows us to claim that homotypic contacts exist, at least to the resolution provided by light microscopy.
2) '[…]deletion of Megf10 causes a profound impairment of IPL-directed SAC dendrite growth, preventing timely sublayer formation'.
It is clear that the processes of SACs, especially the OFF SCAs, in the Megf10 KO have not reached or contribute to a singular plexus in the IPL by P3 (although they do). But at P3 in the mutant, the majority of the processes appear directed towards the IPL (Figure 5B). It would be good to clarify whether the authors mean that Megf10 is important for the rate at which the SAC process lamination occurs, or they also mean, the direction of process outgrowth.
The reviewer is right that we used sloppy and confusing wording here in our original submission. We have removed all references to “soma-directed” and “IPL-directed” arbors and replaced them with descriptions of where SACs are ramifying/elaborating their arbors (i.e. within the soma layer or within the IPL). For example, in place of the sentence quoted by the reviewer, we now write, “These studies (i.e. deletion of Megf10) revealed a severe deficit in IPL dendrite arborization” (subsection “Requirement for MEGF10 in SAC IPL innervation and sublayer formation”).
3) SAC-BC repulsion regulates bipolar cell axon lamination.
This isn't so clear to me because it is difficult to come to this conclusion based on the images. The small displacements in BC5+7 may simply be because where SACs occupy space, BC5+7 cannot elaborate their axons because the SAC process network is very dense. Where there is a gap in the SAC plexus, these BC axon terminals do come closer together, and this could be because space is now available rather than repulsion is absent.
The reviewer’s model is entirely plausible. We did not mean to argue against this possibility; it is one of many ways in which repulsion can be implemented. The point we sought to make in describing our observations is this: If SAC arbors occupy an IPL location, BC3a/BC5/BC7 bipolar cell terminals are excluded from that location. This stands in stark contrast to the behavior of ooDSGC dendrites: If SACs occupy an IPL location, ooDSGCs will occupy it too. Because some cells can enter while others cannot, we feel that describing this as “repulsion” is reasonable; the SAC barrier is not so impenetrable as to exclude every arbor.
Nevertheless, we take the reviewer’s point that caution is needed in how we describe this finding. In the Results section and the new model figure (Figure 11F), we do not use the word “repulsion” but instead simply describe where the bipolar cell axons go when confronted with a SAC ectopia or gap. We have revised the relevant portions of the Discussion section starting at subsection “SACs as a scaffold for DS circuit assembly”) to soften our claims, referring to this as “possible repulsion” and “arbor exclusion.” We also draw a contrast with ooDSGCs to make the point that there is some selectivity to the putative repulsion (Results section).
1) Throughout the text, the authors make statements about the behavior of the SACs, which are not strongly supported by the images. Here are some examples:
i) Subsection “Homotypic contact is required for SAC IPL innervation and dendrite lamination”: "Misprojecting DACs are still in contact with numerous other amacrine[…]" and legend for Figure 3—figure supplement 1G: "SACs that made errors had extensive interactions with GAD65+ amacrine cells". It not possible to infer 'contact' or 'interactions' from these data. Please explain.
We apologize for imprecise wording here. What we should have said was: There is a strong likelihood of contact/interactions between SAC arbors and GAD65+ amacrine arbors in the IPL. To support this claim, we have gone back to examine the original confocal Z-stacks from the P15 Ptf1acKO experiment. We find that the images in the figure cited by the reviewer (now Figure 3—figure supplement 2F) are entirely representative of the vast majority of our images from this experiment: GAD65+ arbors completely filled the IPL in all Z planes examined. Thus, when SACs projected into the IPL, they were highly likely to interact with if not contact GAD65+ arbors, as illustrated in Figure 3—figure supplement 2F.
Based on this analysis, and mindful of the reviewer’s wise suggestion to be careful in how these images are described, we now write in the Results section: “Misprojecting SACs were still closely opposed to numerous other amacrine cells, and their arbors were intermingled in the IPL, strongly suggesting that generic amacrine interactions are not sufficient to ensure normal dendrite targeting.”
The legend to Figure 3—figure supplement 2F has also been updated to read: “SACs that made errors likely interacted in the IPL with GAD65+ amacrine cell arbors (F, arrows) because these arbors completely filled the IPL in the region innervated by the SAC dendrites.”
ii)Subsection “SAC IPL errors induce laminar targeting errors by their DS circuit partners”: 'No changes were seen in Syt2-labeled BC6, suggesting a specific effect on the bipolar cells that make extensive contact with SACs.' The BC6 axons are much further away from the SAC ON plexus compared to the BC7 axon terminals. Thus, I am not sure one can readily conclude that 'DS-circuit bipolar cells respond to SAC attractive cues'.
Because the analysis of BC6 was “data not shown,” we have chosen to remove the sentence rather than perform a more comprehensive analysis. We now make claims only about the behavior of BC2, BC3a, BC5, and BC7.It is important to note that BC6 (unlike the four bipolar types we did analyze) is not currently thought to be a monosynaptic partner of SACs or ooDSGCs, so removing this data does not alter our conclusion that DS circuit bipolar cells change their IPL lamination to follow SAC errors.
2) The authors have really put in an enormous effort to obtain a comprehensive study. There is a lot of data and the supplements that complement the Figures are important. But I wonder if the authors would consider streamlining some parts because it is really a bit overwhelming for the reader to try and assimilate all the information relevant to each major conclusion by going back and forth from the main figures to the supplements. For example, the various effects on bipolar cells in the Megf10-/- retina; it would be helpful to at least provide the summary in a single figure for the bipolar cell types analyzed (BC2,3a,5,6,7).
We thank the reviewer for this helpful suggestion. We have included bipolar cell model figures as the reviewer requested (for clarity, we made separate model figures to put next to the primary data in the main and supplementary figures). To the larger point about streamlining: As we noted in the beginning of this document, we have added many new model figures, broken up large figures, and shortened the Results section to make the story easier to follow. One beneficial effect of our figure reorganization is that we made room to move certain key panels from the supplementary to the main figures. This should aid the reader in finding the data they are looking for.
3) The mistargeting of BC2 in Megf10-/- to where there are ectopically located SAC processes is intriguing and clear that there is a bipolar axon terminal lamination defect associated with abnormally placed SAC processes (but for a BC type that is not part of the SAC-DS circuit?).
BC2 is part of the SAC-DS circuit according to Duan et al., (2014) and Kim et al., (2014). Serial block-face EM reconstructions in Ding et al., (2016) show that, among the bipolar cell types, BC2 and BC3a make the most synapses onto OFF SACs, along with BC1.
However, to establish a causal link to SACs, it would help to show that lamination of other retinal neurons that are associated with BC2 bipolar cells (if that is known) are unperturbed in their lamination (i.e. are the BC2's following their targets?).
To our knowledge specific non-DS circuit targets of BC2 (and BC3a) have not been definitively identified. So additional experiments here are not possible. However, the reviewer’s point is well taken: It is possible that starbursts affect BC2 (and the other BC cell types) indirectly, by moving an intermediary cell type to a new laminar location. Such a mechanism would be entirely consistent with the “SAC as early scaffold” model, but it would imply that the direct bipolar cell interaction happens with the intermediary cell type instead of with SACs themselves. We now acknowledge this possibility in the Discussion section: “SACs might achieve their scaffolding functions directly, by providing guidance cues to their partners; or they may do so indirectly, by patterning the IPL projections of an intermediary cell type that in turn guides later-arriving projections.”
We also softened our claim concerning SAC-BC interactions in the Results section, referring now to SAC “guidance strategies” rather than “guidance cues”.
4) Figure 2—figure supplement 2: There should be quantitation for the E16 observations to substantiate SACs projecting to each other. The neuroblastic layer is dense with nuclei – where the processes appear to 'project', they seem to be just coursing around the nuclei.
As requested we have performed quantitative analysis of the internexin+ primary dendrites; more on this below. But first we should address a confusion that may have arisen for the reviewer due to imprecise language in our previous submission. We did not mean to claim that all SACs project directly towards each other; indeed, it would be impossible to conclude this from the internexin staining pattern. Rather we are making a point about the angle of primary dendrite orientation: If a SAC projects in the tangential plane, within the INBL (or INL/GCL at later stages), it is making a projection in the direction of neighboring somata rather than towards the IPL. By projecting in the tangential plane, E16 SACs make it possible to contact each other’s somata or soma-layer arbors. By contrast, P1-2 primary dendrites are oriented towards IPL and therefore quite unlikely to establish such contacts (although previously established contacts do persist through this period).
We have revised the Results to make clear that the internexin analysis is aimed at determining whether SACs orient their primary dendrites towards the IPL, or in the tangential plane where they might contact neighboring somata. To make this determination, we measured the angle of internexin+ primary dendrites at E16, and at P1 for comparison. This analysis, now presented in Figure 2F, shows that E16 SACs have a broad range of primary dendrite projection angles, including many that project tangentially within the INBL. By contrast, P1 SACs project almost exclusively towards the IPL. A schematic illustrating how the analysis was done is provided in Figure 2—figure supplement 1E. The Materials and methods section have also been updated to describe this analysis.
5) Are SACs not expected to be labeled in the Gad1-GFP line?
This is an important question; we thank the reviewer for spurring us to address it. Empirically, we found that SACs are labeled only rarely in the Gad1-GFP line. A picture is now provided in Figure 5—figure supplement 1 showing lack of overlap between GFP and the SAC marker Sox2.
6) It would be helpful to indicate in all the figures whether we are looking at a single image plane or a z-projection of several optical planes. This will help the reader gain a sense of just how 'close' the labeled structures are. What was the z-dimension step size for the confocal images on the wholemounts (not sure I found this in the Materials and methods section).
This is an important question; we thank the reviewer for spurring us to address it. Empirically, we found that SACs are labeled only rarely in the Gad1-GFP line. A picture is now provided in Figure 5—figure supplement 1 showing lack of overlap between GFP and the SAC marker Sox2.
7) Where possible, labeling the borders of the IPL would be helpful- presumably these are defined always by labeling of cell nuclei?
We have added vertical white lines or dashed horizontal lines to mark the IPL width/borders in many of the figures (e.g. Figure 1, Figure 4, Figure 5, Figure 7, Figure 8 and Figure 9). IPL borders were defined in various ways depending on the experiment: (1) labeling of cell nuclei; (2) tdTomato fluorescence from unrecombined cells in mTmG mice, which fills the IPL; (3) immunofluorescence against GAD65, which also fills the IPL; (4) autofluorescence signal intensity differences between soma layers and IPL. These details have been added to subsection “Image acquisition and processing”.
8) OFF SACs show more dramatic ectopic (somatic) processes than ON SCs in the Megf10 mutant. But this defect does not seem to have a physiological correlate – or does it?
This is a very interesting question. To ask if the effects on DS tuning width and strength were similar for the OFF and ON DS responses, we have now performed additional data analysis. We found that the magnitude of the mutant effect size was the same for both ON and OFF.
However, we are choosing not to include the analysis in the paper, for two reasons. First, it turns out that the necessary analysis was fairly complicated and would require its own multipanel figure to explain in a clear and satisfying way. We are reluctant to burden the reader with an entire new multi-panel figure that is mostly peripheral to the paper’s main point. Second, our analysis uncovered some interesting differences between the ON and OFF tuning curves for wild-type ooDSGCs. These are interesting enough that they merit a more complete unpacking in the context of a separate study; they really should not be buried in the supplementary figures of this paper.
The fact that the reviewer raised this point suggests to us that it is worth addressing in the paper. Given the above considerations we elected to make this “data not shown” and mention it in the legend to Figure 12. We write: “Mutant ooDSGC population is tuned more broadly (D, right shift of red curve) and more weakly (E, left shift of red curve) than WT. Similar results were obtained when ON and OFF responses were considered separately (not shown).”
We are extremely confident in the outcome of our analysis, so we think this is a reasonable way to balance general readability concerns with the specific interests of a subset of our readers.
9) Gad1-GFP line should be listed in the transgenics used.
We regret this oversight; it has been fixed (Key Resources Table; subsection “Animals”.)
10) Has the expression of cre in the Chat-cre line been documented somewhere? This is important with respect to separating the early and late deletion of Megf10.
We regret this oversight; it has been fixed (Key Resources Table; subsection “Animals”.)
11) Figure 8—figure supplement 1: Please provide a more detailed explanation of how the 2D cross-correlation analysis was performed.
For the sake of clarity and streamlining the manuscript, we have decided to remove this analysis. It is only useful for making a very abstract and theoretical point, which is somewhat peripheral to the manuscript’s main point. Moreover, it is very complicated to explain. For further details please see response to reviewer #3 (comment 3) below.
Given that this has been removed, we have now revised our conclusions about the interactions between mosaic phenotypes and IPL errors. We now write: “While mosaic spacing errors do not account for the Megf10 mutant ectopic IPL phenotype, we cannot exclude the possibility that the placement of IPL arbor gaps might be at least partly explained by soma position.” (subsection “Formation of SAC IPL sublayers”).
12) Although it does appear that Meg10 plays two roles- one in somal mosaic positioning and the other in lamination of the processes of SACs, it isn't clear that timing alone separates these roles.
We did not mean to imply that timing is the only thing that separates the two MEGF10 roles. No doubt the mosaic positioning effect is mediated by its own cellular mechanisms; these would constitute additional factors, aside from timing, that separate the two MEGF10 functions. We apologize for being unclear on this point. Instead we meant to say that, since timing can separate the two roles, this demonstrates that they are indeed separate. During the earliest stages of SAC-SAC interactions (i.e. prior to P3), it is clear that MEGF10 influences both mosaic spacing and laminar targeting. If this were all we knew about MEGF10, it would be entirely possible that the laminar targeting and mosaic spacing phenotypes are really one and the same, entirely inseparable. The late deletion experiment indicates that is not the case – one phenotype can be induced without the other. This is the key point we sought to make. We have added a new model figure summarizing the results of the conditional mutant experiments, which should help clarify this point (Figure 9C).
If mosaic arrangements are due to repulsive SAC-SAC interactions, then fasciculation of the SAC processes must be an adhesion-based, perhaps even 'attractive' mechanism. Is it known if Megfs have other downstream pathways that are not 'repulsive'?
We agree with the reviewer that fasciculation – a key step in sublayer formation – must be adhesion/attraction based. In fact, we make this point in subsection “Formation of SAC IPL sublayers”. As we now try to suggest in the revised version of this paragraph, our data indicates that MEGF10 is not acting in a repulsive fashion, at least when it comes to IPL innervation. However, it is not the attractive/adhesion cue either, as shown by the fact that layers ultimately form in its absence. Instead, it seems to be a signal for a change in growth mode. Whether MEGF10 acts in this way during mosaic formation remains to be determined in future work. The revised text is quoted below in answer to the next query.
13) Could the authors elaborate a bit more in the Discussion section on how the lamination of the OFF and ON SAC plexuses become better confined to a single lamina (with exceptions) by adulthood?
The honest answer here is that we do not really know how the sublayers ultimately form in Megf10 mutants. In subsection “SAC errors persist to adulthood in Megf10 mutants” we point out that there are compensatory mechanisms that ultimately drive SACs to innervate the IPL and fasciculate there into sublayers (albeit imperfectly).
This comment from reviewer 2, together with the previous one, spurred us to elaborate in the Discussion section on one possible compensatory mechanism. We point out that there likely exist attractive/adhesive cues that promote homotypic SAC fasciculation – a key step in layer formation (subsection “Formation of SAC IPL sublayers”). This system remains intact in mutants, and therefore could be one of the compensatory mechanisms. The revised paragraph reads:
“It is generally assumed that sublayer formation has two basic molecular requirements: (1) Attractive/adhesive molecules that mediate co-fasciculation of stratified arbors; and (2) repulsive cues that prevent straying of arbors into other sublayers (Lefebvre et al., 2015; Sanes and Yamagata, 2009). Our MEGF10 studies suggest an additional, earlier requirement for cell-cell interactions that occur prior to neuropil innervation. The purpose of this surprisingly early SACSAC interaction, we propose, is to ensure that SACs grow dendrites at the right time and place to co-fasciculate with their SAC neighbors. The molecular basis of this homotypic co-fasciculation – clearly another essential player in sublayer formation – remains to be determined. MEGF10 is probably not involved; the cofasciculation system appears intact in Megf10 mutants given that sublayers do eventually form. Perhaps this system is part of the mechanism that compensates for loss of MEGF10 to ultimately generate the sublayers.”.
Reviewer #3
This study by the Kay lab investigates the role of Megf10 in the development of starburst amacrine cells (SACs). In a previous study by Kay and colleagues (Kay et al., 2012), the mosaic spacing of SACs is impaired in Megf10 knockout mice. In this manuscript, Ray and colleagues examine the early developmental stages of SACs starting at E16. They found that Megf10 is required for the timely stratification of SAC neurites at the inner plexiform layer (IPL). In the absence of Megf10, SACs fail to eliminate ectopic neurites and unable to stratify at the correct IPL depth in the first few postnatal days (P0-P2). This developmental deficit leads to mistargeting and gaps of SAC neurites in the IPL in the mature retina. Furthermore, using elegant genetic tools, the authors demonstrate that Megf10 function requires homotypic interaction of megf10 expressed in neighboring SACs. Overall, this study advances the current understanding of sublayer formation in the retina. It also provides a missing link between the altered developmental process of SACs and the deficits in the mature retina of Megf10 mutants. My major concerns are about some of the interpretations of the data and analysis.
We thank the reviewer for this comment; we are glad they agree that our work advances the field on several fronts.
1) One developmental feature that is prominently mentioned in this study is a transient homotypic network of SAC dendrites in the inner nuclear layer (INL) between E16 and P1. The authors describe the neurites outside the IPL to be selectively targeted to the neighboring SAC somas and neurites in the INL. However, an alternative scenario is that SACs during this period have random growth of neurites to both INL and IPL. After P1, the INL neurites are selectively pruned, and the IPL neurites further grow and stratify in the correct IPL sublayer. I did not find sufficient evidence supporting the presence and functional significance of a homotypic network of SAC arbors at the INL.
In addressing the question of whether a network of SAC INL arbors is present, we would first point out that our data show definitively that SACs make extensive transient projections to the soma layers (particularly the INL). The extent of these projections is quite surprising, since before now it had generally assumed that retinal neurons project stable arbors only within the plexiform layers, even during early development. In the initial submission we showed that most (~88%) of these INL SAC arbors contact a neighboring SAC soma or arbor. This finding led us to claim that the INL network is homotypic. The reviewer raises the quite reasonable question of whether this rate of apparent contact might arise even if “SACs during this period have random growth of neurites.” We have now performed a flipped-channel analysis to investigate this question (Figure 2—figure supplement 3). We found that the “random” rate of contact (in the flipped condition) is about 15% (Figure 2—figure supplement 3). This is significantly lower than the 88% contact rate we observed in the real data; therefore, we conclude that the interaction between GFP+ individual SACs and the broader βGal+ SAC arbor network is likely to be selective and specific.
We believe this finding should go a long way towards alleviating the reviewer’s concern, raised in the “alternative scenario,” that INL projections are random. The remaining aspects of the alternative scenario are not necessarily in conflict with what we propose is happening in embryonic retina. The reviewer’s alternative scenario mainly concerns how the soma-layer network arises. For instance, the alternative scenario posits that E16 SACs project in random directions. It is entirely possible, given our new data (Figure 2F), that this is in fact the case, but those dendrites within the INBL that contact neighboring SACs become stabilized, leading to homotypic network formation. We have not investigated in detail how the network arises, beyond showing that E16 SACs project into the soma layer such that they could plausibly contact their neighbors. Nevertheless, showing that this homotypic soma-layer network exists is a novel anatomical discovery; it motivates the rest of the study and, we hope, will change how the field thinks about early dendrite differentiation in retinal interneurons. How exactly the network arises is certainly an important question, but we feel it is best left for future studies.
In particular,
- Subsection “Early SAC projections target neighboring SAC somata”: The authors describe SAC neurites in the INL as "soma-directed" neurites. It is unclear to me whether these neurites are "directed" at somas, or they are just randomly distributed in 3D. There is no quantification that shows "selective" somatic targeting over random growth. In fact, in both Figure 2E (E16) and 2H (P0), SACs send more neurites to IPL than to INL.
As noted in our response to reviewer 2, the choice of “somadirected” and “IPL-directed” as descriptors of SAC arbors was a poor one, and confusing. While a tangential projection within the soma layer might be considered “soma-directed,” the reviewer is right that we cannot establish the precise guidance parameters for any given dendrite; nor is this necessary to make the point that we wish to make. What we intended to communicate in the initial submission, and what we believe we now communicate in this revision, was the following: E16 SACs direct primary dendrites tangentially, within the soma layer, while postnatal SACs do not. By projecting tangentially within the INBL, embryonic SACs adopt a trajectory where they might contact their neighbors, thereby giving rise to the soma-layer contacts that we observed at E16 and that persist until P2-3 (Figure 2D,G).
To clarify this point we have replaced the term “soma-directed” with “soma-layer projections,” and have made other edits to subsection “Early SAC projections target neighboring SAC somata” that are consistent with the model described above. Further, we now add quantitative data supporting the existence of tangential soma-layer projections at E16 but not P1 (Figure 2F), and we show that the result of the E16 events is establishment of a homotypic INL network, not a random one (Figure 2—figure supplement 3). While there may be a developmental stage where arbors are distributed randomly in 3D, as the reviewer suggests, the ultimate outcome is a homotypic network, perhaps through selective stabilization of certain arbors. It will be interesting to investigate this possibility in future studies.
The reviewer is correct that many E16 primary dendrites are also directed towards the IPL. This is not in conflict with our model, as long as arbors are also projecting tangentially such that they could establish a soma-layer network. Our newly added data shows that this is the case – tangential and IPL-directed projections account for a similar fraction of total primary dendrites analyzed, with tangential being slightly more frequent (Figure 2F).
The key revised portion of the Results section reads:
“The majority of INBL SACs engaged in these soma-layer contacts, such that a GFP+ arbor network connected them (Figure 2G). Analysis of primary dendrite orientation indicated that soma-layer contacts likely arose due to projections targeted within this layer: Unlike mature SACs, which exclusively project their primary dendrites towards the IPL, many E16 SACs projected tangentially through the INBL – i.e., towards neighboring somata (Figure 2E,F).”.
- Subsection “Early SAC projections target neighboring SAC somata”: the authors claim that "P1 INL arbors selectively contacted somata or arbors of SAC neighbors". I have difficulty recognizing this pattern as "selective" since the INL at this stage is densely covered by SAC somas and ectopic neurites. The INL neurites may always end next to neighboring somas or neurites just because of the high density.
As noted above, we have addressed this concern by flipping one channel of the confocal stacks to obtain an expected rate of interaction if arbors were randomly distributed (Figure 2—figure supplement 3). The “random” rate of contact (in the flipped condition) is about 15%, significantly lower than the ~88% contact rate we observed in the real data.
- Subsection “Early SAC projections target neighboring SAC somata”: The authors describe that On SACs send fine soma-directed branches from their IPL arbors (Figure 2—figure supplement 2). Again, how do the authors distinguish "soma-directed" projection from random contacts due to imperfect stratification?
In the figure panels cited by the reviewer (now Figure 2—figure supplement 3E,F) we provide specific examples of cases where ON SACs branched to contact SAC somata located either within the IPL, or at the IPL/GCL border. We have not performed the same quantitative specificity analysis as we did for OFF SACs. But we argue that the anatomy shown in Figure 2—figure supplement 3E,F is consistent with ON SACs behaving similarly to OFF SACs. The figure shows: (1) the existence of branching morphology that brings dendrites into the vicinity of neighboring homotypic somata; and (2) the existence of actual dendro-somatic SAC-SAC contact. Thus, ON SACs, like OFF SACs, create the possibility for soma contact by projecting into the soma layer, and this possibility is realized, at least in some cases. We propose that the most likely interpretation of these ON-OFF similarities is that both SAC subtypes are engaging in homotypic interactions.
However, because we have not performed the specificity analysis, we softened the language used to describe the ON SAC anatomy. We have removed the word “selective” when describing ON SAC interactions, and we now write, “ON SACs also made soma layer projections between P0-P3 that contacted neighboring SAC somata” (subsection “Early SAC projections target neighboring SAC somata”).
2) Subsection “Timing of DS circuit IPL sublayer formation”: The authors mention that SACs are stratified within the expected IPL sublayer, but ooDSGC dendrites are unstratified. This needs to be supported by quantifications. In Figure 1—figure supplement 1, panel C, both SACs and the ooDSGC appear to be unstratified in a similar way.
We thank the reviewer for a suggestion that undoubtedly strengthens the paper. We have now performed fluorescence profile plots across the IPL for P1 and P2 dual-channel SAC & ooDSGC images, similar to what we had previously included for mature retina (old Figure 9; new Figure 10 and Figure 11). Representative examples are now shown in Figure 1E. These plots show clear sublaminar peaks for SAC dendrites at both P1 and P2, while ooDSGC dendrites are distributed more broadly across the IPL (Figure 1E). Importantly, the representative profile chosen for P1 is from the cell that the reviewer questions in the comment above; the profile clearly shows stratification of SACs but not ooDSGCs. While seeking to improve the clarity of this figure and its accompanying supplementary figure (Figure 1—figure supplement 2), we also chose a better representative image to include in panel Figure 1D. The previous panel 1D cell is now in the supplement.
3) Subsection “Homotypic contact is required for SAC IPL innervation and dendrite lamination”: "the projections were typically more elaborate than those observed in wild-type retina". Please provide a quantitative analysis to support this claim.
This is a good suggestion, but unfortunately, we have not been able to devise a satisfactory way to do it. We attempted to measure arbor coverage area and other dendrite anatomical parameters in whole-mount, en face views of solitary and touching SACs, but the βGal labeling was too dim to permit such an analysis. Measuring these parameters in crosssection is difficult, because arbors may be severed at the sectioning border. We felt that any quantitative measure we might obtain from these samples would therefore show only a relative difference between the two groups, rather than providing an absolute measure (e.g. arbor area, total arbor length, or Sholl branch complexity) that could be compared productively across experiments and across labs. Because this relative difference in arbor complexity is already quite readily apparent from a qualitative inspection of the micrographs (Figure 3D, arrow vs. Figure 3F; Figure 3—figure supplement 2), we argue that quantification would add little value. Indeed, as shown in these figures, many solitary SACs have a dendrite morphology completely outside the range of anything observed for controls. Quantifying dendritic material in the INL would fail to accurately describe this qualitative morphological change.
Furthermore, we already showed quantifications of the number of solitary SACs projecting to the soma layer – this was increased relative to both controls (Figure 3G). So, when we assert that solitary SACs over-innervate the soma layer, we have both qualitative and quantitative data to back up this claim.
To aid the reader in assessing the validity of our qualitative claim, we created a supplementary figure (Figure 3—figure supplement 2A-D) devoted to providing multiple examples of solitary SAC dendrite morphology. The text has also been revised (legend to Figure 3—figure supplement 2) to clarify that the example images are indeed representative.
4) Subsection “Requirement for MEGF10 in SAC IPL innervation and sublayer formation”: "this phenotype is not due to aberrant SAC migration.……" However, from Figure 5A and Figure 5B, Off SAC somas in MEGF10 mutants are farther away from IPL than those in the wild type.
We agree with the reviewer that this is too strong a statement; we have clarified by specifying that we are referring here specifically to “radial migration” and that it occurred “similar to littermate controls.” However, based on the evidence provided, we feel comfortable making our revised assertion that “Loss of IPL innervation was not due to aberrant SAC radial migration.” (subsection “Requirement for MEGF10 in SAC IPL innervation and sublayer formation”). While it is true, as the reviewer points out, that P1 OFF SACs are clearly farther from the IPL in mutants than in controls, we find no evidence for a major difference at P0. This can be observed in Figure 5A, where INL SACs are positioned adjacent to the IPL in both mutants and littermate controls. It is also clear from our counts of INL SACs at P0 (reported in the text in subsection “Requirement for MEGF10 in SAC IPL innervation and sublayer formation”), which showed no difference in the number of SACs arriving in the INL by this age. Finally, one can appreciate in cross-sections (Figure 5A; data not shown) that mutant SACs do not arrest their migration in the ONBL; they successfully reach inner retina by P0, in normal numbers. Together these data are consistent with a model whereby SACs complete their radial migration more or less normally, arriving at the IPL, but then moving away from it (possibly as a result of failing to innervate it properly) between P0 and P1.
We have not measured the soma-IPL distance at P0, so it is possible that there may be some subtle effects of Megf10 on final position after migration. For this reason, we agree with the reviewer that we should not say radial migration was “normal.” But the issue here is whether the migration defect (if any) is sufficient to explain the lack of IPL innervation. If the SACs are farther from the IPL at P0, it is not by much; indeed, as we now point out in the text (subsection “Requirement for MEGF10 in SAC IPL innervation and sublayer formation”), most SACs are able to send at least some dendrites into the IPL, suggesting that they migrated to a location from which IPL innervation was feasible. However, once they enter the IPL, these dendrites do not ramify and stratify as control dendrites do. These observations strongly suggest that any migration/radial positioning defects are unlikely to be the primary cause of IPL innervation defects in Megf10 mutants. The relevant section of the manuscript now reads:
“Loss of IPL innervation was not due to aberrant SAC radial migration, because, at P0, mutant SACs had reached the inner retina in normal numbers (wild-type, 2600 ± 287 SACs/mm2; mutant, 3153 ± 145 SACs/mm2; p = 0.144, 2-tailed t-test; n = 3 each group), and were positioned adjacent to the IPL, similar to littermate controls (Figure 5A). Furthermore, most mutant SACs sent at least some arbors into the IPL at P0-1 (Figure 5A,C; Figure 6C), suggesting that they migrated to a location from which IPL innervation was feasible (Figure 5A-C; Figure 6C). However, the mutant SAC arbors that reached the IPL appeared undifferentiated, with a lack of space-filling branches (Figure 5C,E).” (subsection “Requirement for MEGF10 in SAC IPL innervation and sublayer formation”).
In addition to this change, we also point out in the legend to Figure 5 that “mutant somata abut the IPL at P0, suggesting their radial migration was similar to controls. By P1, OFF somata have moved apically.”
And there are significantly less On SACs in the mutant. This is consistent with delayed migration of SACs, which might also account for the delayed stratification phenotype in the mutants.
The reviewer makes an interesting observation concerning the ratio of OFF to ON SACs. Chow et al. (2015) showed that displaced amacrine cells cross the IPL using a migratory strategy that is distinct from radial migration: At the completion of their radial migration to the IPL, cells flatten in the tangential plane and extend bipolar arbors within the IPL. We observed SACs with this morphology (Figure 2—figure supplement 3) suggesting that ON SACs likely cross the IPL in this way.
Inspired by the reviewer’s suggestion, we did a pilot study on a small number of sections from a small number of P0 animals (5 fields of view from 2 Megf10 mutants and 2 littermate controls). We found that, while the total number of SACs in the INL and GCL was unchanged between mutant and control, the INL/GCL ratio was indeed slightly higher in the mutants, as the reviewer observed (67% INL in mutants, 54% INL in controls). The fact that total SAC number was similar suggests that radial migration is completed largely normally, consistent with data reported in the manuscript (for further details see our response above to comment 3H). However, the altered INL/GCL ratio suggests there may be a requirement for MEGF10 in the IPL crossing behavior discovered by Chow et al.
Proving that this is in fact the case would require substantial further experiments, which we feel are outside the scope of this study. Even if the phenotype were confirmed, it would not change our conclusion that migratory phenotypes cannot explain the loss of IPL innervation (subsection “Requirement for MEGF10 in SAC IPL innervation and sublayer formation”). This is because IPL crossing happens after radial migration is complete; moreover, it happens concurrently with initial IPL innervation rather than prior to it (Chow et al., 2015). If something has gone wrong with crossing, the problem likely occurred during or after initial interaction with the IPL and would not have prevented such interactions.
The reviewer suggests that delayed crossing might be responsible for delayed stratification in Megf10 mutants. This is a reasonable hypothesis. However, given our finding that IPL innervation is delayed in mutants, it is equally reasonable to propose the opposite hypothesis: If mutant SACs are unable to properly send arbors into the IPL, this might delay the onset of crossing behavior. Since both possibilities are reasonable, additional work will be required to distinguish between them (if it turns out that MEGF10 mutants do in fact have an IPL crossing phenotype). It will be quite interesting to carry out future studies to resolve this issue. However, we feel that such work is clearly separate from this study, and not necessary to support our conclusions.
To take into account the possibility that IPL crossing might not be normal in Megf10 mutants, we have revised our migration claim to specifically address radial migration, rather than migration generally (subsection “Requirement for MEGF10 in SAC IPL innervation and sublayer formation” and the legend to Figure 5).
5) Subsection “Requirement for MEGF10 in SAC IPL innervation and sublayer formation”: "Rather, the absence of SAC sublayers was due to innervation of the soma layers instead of the IPL". I do not see a clear causal relationship here.
Thank you for pointing out our imprecise wording. We did not mean to imply that the soma layer overinnervation causes absence of sublayers; this sentence has been removed. The absence of sublayers is attributable to the IPL innervation deficits, as we now make clear in the portion of the Results section that replaces this sentence (subsection “Requirement for MEGF10 in SAC IPL innervation and sublayer formation”).
6) Related to point 1 above, Subsection “Requirement for MEGF10 in SAC IPL innervation and sublayer formation”: To support the statement that P1 SACs in Megf10 mutants "preferentially" project neurites to the soma layer, the authors need to show that each SAC in Megf10 mutants sends more neurites in the INL than IPL.
Perhaps “preferentially” was too strong of a word choice; we have removed it. The section describing Megf10–/– phenotypes (subsection “Requirement for MEGF10 in SAC IPL innervation and sublayer formation”) has been extensively re-written to avoid over-interpretation. We now focus on pointing out that mutant SACs have two phenotypes: underinnervation of the IPL (Figure 5) and the overinnervation of the soma layers (Figure 6).
To address the reviewer’s concern, we have also added new figure panels and analysis to support the assertion that P1 SACs overinnervate the INL:
1) We now show in the main figures, using Megf10:βGal, that the mutant P1 INL arbor network is more extensive than the control network. This was previously in the supplementary figures, but we have moved it to a much more prominent main figure location (Figure 6C) because of the importance of this data.
2) We highlight (subsection “SAC dendrite targeting requires transcellular MEGF10 signaling”) an important finding that was present in the earlier submission but did not receive sufficient emphasis: At P1, the number of SACs projecting to the INL is the same in mutants and controls (Figure 6E). If the same number of SACs participate in the INL network, but the network is denser in mutants (Figure 6C), it stands to reason that the typical mutant SAC must have a larger/more elaborate INL arbor than the typical control SAC.
3) In support of this inference, we now provide high-magnification en-face images of representative control and mutant SACs, documenting more elaborate INL arbors in mutants (Figure 6B).
These new data bolster the lower-magnification views included in our original submission (Figure 5A,B; Figure 6A), which clearly show at P1, and especially at P3, that the amount of dendritic material in the mutant INL is dramatically higher than in controls. Indeed, between P0 and P3, the amount of dendritic material in the mutant INL is stable or even increases, while it decreases in controls (Figure 5A-B). While we cannot claim (nor did we intend to claim in the original submission) that each individual SAC sends more neurites to the INL than the IPL, we feel our data strongly supports the following claim: At the population level, mutant SACs overinnervate the INL during the period when their IPL arbors are underdeveloped. The relevant portion of the Results section has been extensively revised to reflect this point.
7) Subsection” SAC errors persist to adulthood in Megf10 mutants”: The authors describe the laminar disruptions in mature retinas of Megf10 mutants occur "sporadically and at apparently arbitrary retinal locations." It is important to determine the percentage of IPL that is disrupted or shows ectopic stratification in Megf10 mutants.
We agree with the reviewer about the importance of documenting the extent of Megf10 IPL phenotypes. We showed in the initial submission that ~80% of OFF SACs participate in the ectopic network (Figure 8F), highlighting the prevalence of these errors. At the reviewer’s request we now extend our quantitative analysis to the IPL gap phenotype: We measured the area covered by the SAC dendritic plexus in ON and OFF sublayers and found that Megf10 mutant SACs cover ~15% less retinal area than control SACs (15.0 ± 0.9% decrease for OFF, 13.7 ± 4.0% decrease for ON, mean ± SD). These data are now included in subsection “SAC errors persist to adulthood in Megf10 mutants”, and the methodology is described in subsection “Characterization of SAC homotypic arbor network in soma layers”.
Retinal coverage measurements were obtained by segmenting en-face images of the ChAT+ SAC plexus and calculating the field-of-view fraction occupied by SAC arbors. We attempted to use this approach to quantify en-face images of the ectopic OFF SAC network, but we could not obtain accurate measurements due to the presence of cell bodies (either out of focus or partially in focus) within the images at this IPL level. Even though we could not perform the exact analysis suggested by the reviewer, we would argue that our measurements of the fraction of SACs that make errors (Figure 8F; Figure 9—figure supplement 1D) provides a reasonable quantitative description of error prevalence.
It is surprising that close to 100% SACs in Megf10 mutant mice show ectopic IPL projection in the adult (Figure 8F), but the laminar disruption shown in this manuscript (e.g. Figure 7A,B, Figure 8A and Figure9D) and the previous paper by Kay et al., 2012 seems to occur for a small fraction of the IPL length.
Based on the figure panels cited by the reviewer, their confusion at the extent of the ectopic SAC projection phenotype is likely due to looking primarily at cross-sections. We should have done a better job drawing the reader’s attention to the en-face panels (e.g. Figure 8E; Figure 10C,D). This vantage point provides the best view of the overall anatomy of the SAC projections; it is hard to appreciate how extensive the IPL errors are without looking at these panels. This is because the amount of ectopic arbor present in a cross-sectional view is highly dependent on how the sectioning angle falls relative to the ectopic arbor fascicles. For example, a perpendicular cut through a long fascicle will leave only a tiny cross-sectional footprint. We have now rewritten the relevant portion of the Results section to draw attention to the whole-mount panels:
“The second type of SAC error in mature Megf10–/– IPL was dendrite mistargeting to ectopic IPL strata (Figure 8A,B,E). Both ON and OFF SACs were affected; in each case ectopic arbors were mostly found in IPL regions inappropriately close to the soma layers (Figure 8A,B). En-face images of mutant IPL revealed that ectopic OFF arbors formed a patchy but extensive fascicle network connecting many of the cells (Figure 8E,F; 78.5 ± 3.5% of SACs participated in the network). (subsection “SAC errors persist to adulthood in Megf10 mutants”).
8) The authors use Six3-Megf10 cko and Chat-Megf10 cko to argue for separate time windows for laminar formation and somatic spacing. However, the data indicate that these two processes occur in overlapping temporal windows. First, irregular soma spacing with gaps in MEGF10 null retinas is very obvious in Figure 5A, indicating that mosaic spacing is already affected at P0. Second, earlier deletion of megf10 in Six3-cko mice causes more sever deficit in somatic spacing than in Chat-megf10 cko mice, indicating that somatic mosaics formation requires early megf10 signaling.
The reviewer is absolutely correct: The time windows are “distinct,” but certainly not separate. As we wrote in the original submission, “… MEGF10 appears to act at distinct, albeit partially overlapping times, to control different aspects of SAC development, each of which are regulated by contact with homotypic neighbors.” (subsection “Formation of SAC IPL sublayers” of original submission; subsection “MEGF10 as the signal mediating SAC homotypic recognition” of new submission). We regret that the manuscript was not written clearly enough for this point to come through in the previous version. To clarify this point we have added a model figure (Figure 9C) summarizing all of the results of our conditional deletion experiments. As shown in the model, MEGF10 is continuously required until at least P3-5 for mosaic spacing, whereas its requirement in laminar targeting has ended by P3. We have also updated the text to more accurately reflect this point. For example, we now write that the two processes have “separable temporal requirements,” rather than “different” temporal requirements as in the first submission (subsection “MEGF10 as the signal mediating SAC homotypic recognition”).
9) Please explain the rationale of the cross-correlation analysis in Figure 8—figure supplement 1. Why is there a strong correlation between the locations of SAC somas and the corresponding locations in the IPL layer? Given the morphology of SACs, it is unexpected to see such a correlation. In addition, statistical test needs to be performed for Figure 8—figure supplement 1C to claim a difference between mutant and control groups.
In the interest of manuscript clarity, we have removed this cross-correlation analysis. The analysis is very abstract, and as the reviewer points out, the connection to anatomy and mechanism is opaque. It is nowhere near as definitive as the other “mosaic vs. IPL error” analysis that we performed (Figure 9—figure supplement 1). Furthermore, the rhetorical point supported by this analysis is rather peripheral to the overall thrust of the manuscript. Finally, the analysis requires a lot of explanation and is hard for the reader to grasp. For all these reasons, we decided to remove this from the paper.
Given that this has been removed, we have now revised our conclusions about the interactions between mosaic phenotypes and IPL errors. We now write: “While mosaic spacing errors do not account for the Megf10 mutant ectopic IPL phenotype, we cannot exclude the possibility that the placement of IPL arbor gaps might be at least partly explained by soma position.” (Discussion section).
- Subsection “Early SAC projections target neighboring SAC somata”: Please clarify the statement "At P0-1, other IPL layers do not yet exist.……" Do the authors mean that there are no neurites in the IPL other than those of SACs?
We apologize for the lack of clarity. What we meant to say was that neurites are present in the IPL but not stratified. We revised this sentence to read: “Because SACs stratify early – before any other cell type investigated to date (Figure 1; Kay and Sanes, 2013; Stacy and Wong, 2003) – they are unlikely to form strata by following pre-existing laminar cues.” (subsection “Early SAC projections target neighboring SAC somata”).
What are the dashed lines of different depths in Figure 5E? How do the authors determine the border of IPL and INL given that SACs are the only amacrine cell type in early development? Delayed migration towards IPL and/or aberrant neurite growth can obscure the classification of IPL borders.
The dashed lines denote the borders of the IPL in each image. We had neglected to note this in the figure legend, but this has now been corrected (this panel is now Figure 6D). As we mentioned above in response to reviewer 2: IPL borders were defined in various ways depending on the experiment: (1) labeling of cell nuclei; (2) tdTomato fluorescence from unrecombined cells in mTmG mice, which fills the IPL; (3) immunofluorescence against GAD65, which also fills the IPL; (4) autofluorescence signal intensity differences between soma layers and IPL. These details have been added to subsection “Image acquisition and processing”.
https://doi.org/10.7554/eLife.34241.034Article and author information
Author details
Christopher V Wright
Funding
National Eye Institute (EY024694)
- Jeremy N Kay
Pew Charitable Trusts
- Jeremy N Kay
E. Matilda Ziegler Foundation for the Blind
- Jeremy N Kay
McKnight Endowment Fund for Neuroscience
- Jeremy N Kay
Alfred P. Sloan Foundation
- Jeremy N Kay
Whitehall Foundation
- Greg D Field
Research to Prevent Blindness (Unrestricted grant to Duke University)
- Thomas A Ray
- Suva Roy
- Christopher Kozlowski
- Jingjing Wang
- Jon Cafaro
- Greg D Field
- Jeremy N Kay
National Eye Institute (EY026344)
- Thomas A Ray
National Eye Institute (EY024567)
- Greg D Field
National Eye Institute (EY5722 to Duke University)
- Thomas A Ray
- Suva Roy
- Christopher Kozlowski
- Jingjing Wang
- Jon Cafaro
- Samuel W Hulbert
- Greg D Field
- Jeremy N Kay
The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.
Acknowledgements
For financial support we thank the National Eye Institute (EY024694 to JNK; EY026344 to TAR; EY024567 to GDF; EY5722 to Duke University); Pew Charitable Trusts, E Matilda Ziegler Foundation, McKnight Endowment Fund for Neuroscience, Alfred P Sloan foundation (JNK); the Whitehall Foundation (GDF); and Research to Prevent Blindness Unrestricted Grant (Duke University). We thank Sidney Kuo and Greg Schwartz for suggesting use of the Gjd2-GFP mouse line; members of the Kay lab and Cagla Eroglu for their comments on the manuscript; Megan Stogsdill and Ari Pereira for mouse colony support; and X Duan and J Sanes (Harvard) for Kcng4Cre mice.
Ethics
Animal experimentation: All animal experimental procedures were reviewed and approved by the Institutional Animal Care and Use Committee of Duke University (protocol A005-16-01).
Reviewing Editor
- Jeremy Nathans, Johns Hopkins University School of Medicine, United States
Version history
- Received: December 11, 2017
- Accepted: March 29, 2018
- Accepted Manuscript published: April 3, 2018 (version 1)
- Version of Record published: May 2, 2018 (version 2)
Copyright
© 2018, Ray et al.
This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.
Metrics
-
- 3,434
- Page views
-
- 543
- Downloads
-
- 33
- Citations
Article citation count generated by polling the highest count across the following sources: Scopus, Crossref, PubMed Central.
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)
Formation of retinal direction-selective circuitry initiated by starburst amacrine cell homotypic contact
eLife 7:e34241.
https://doi.org/10.7554/eLife.34241
Further reading
-
- Cell Biology
- Neuroscience
The amyloid beta (Aβ) plaques found in Alzheimer’s disease (AD) patients’ brains contain collagens and are embedded extracellularly. Several collagens have been proposed to influence Aβ aggregate formation, yet their role in clearance is unknown. To investigate the potential role of collagens in forming and clearance of extracellular aggregates in vivo, we created a transgenic Caenorhabditis elegans strain that expresses and secretes human Aβ1-42. This secreted Aβ forms aggregates in two distinct places within the extracellular matrix. In a screen for extracellular human Aβ aggregation regulators, we identified different collagens to ameliorate or potentiate Aβ aggregation. We show that a disintegrin and metalloprotease a disintegrin and metalloprotease 2 (ADM-2), an ortholog of ADAM9, reduces the load of extracellular Aβ aggregates. ADM-2 is required and sufficient to remove the extracellular Aβ aggregates. Thus, we provide in vivo evidence of collagens essential for aggregate formation and metalloprotease participating in extracellular Aβ aggregate removal.
-
- Computational and Systems Biology
- Neuroscience
The cerebellar granule cell layer has inspired numerous theoretical models of neural representations that support learned behaviors, beginning with the work of Marr and Albus. In these models, granule cells form a sparse, combinatorial encoding of diverse sensorimotor inputs. Such sparse representations are optimal for learning to discriminate random stimuli. However, recent observations of dense, low-dimensional activity across granule cells have called into question the role of sparse coding in these neurons. Here, we generalize theories of cerebellar learning to determine the optimal granule cell representation for tasks beyond random stimulus discrimination, including continuous input-output transformations as required for smooth motor control. We show that for such tasks, the optimal granule cell representation is substantially denser than predicted by classical theories. Our results provide a general theory of learning in cerebellum-like systems and suggest that optimal cerebellar representations are task-dependent.
