(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



Differential adhesion during development establishes individual neural stem cell niches and shapes adult behaviour in Drosophila [1]

['Agata Banach-Latapy', 'Institut Pasteur', 'Université Paris Cité', 'Cnrs', 'Structure', 'Signals In The Neurogenic Niche', 'Paris', 'Vincent Rincheval', 'Université Paris-Saclay', 'Uvsq']

Date: 2023-11

Neural stem cells (NSCs) reside in a defined cellular microenvironment, the niche, which supports the generation and integration of newborn neurons. The mechanisms building a sophisticated niche structure around NSCs and their functional relevance for neurogenesis are yet to be understood. In the Drosophila larval brain, the cortex glia (CG) encase individual NSC lineages in membranous chambers, organising the stem cell population and newborn neurons into a stereotypic structure. We first found that CG wrap around lineage-related cells regardless of their identity, showing that lineage information builds CG architecture. We then discovered that a mechanism of temporally controlled differential adhesion using conserved complexes supports the individual encasing of NSC lineages. An intralineage adhesion through homophilic Neuroglian interactions provides strong binding between cells of a same lineage, while a weaker interaction through Neurexin-IV and Wrapper exists between NSC lineages and CG. Loss of Neuroglian results in NSC lineages clumped together and in an altered CG network, while loss of Neurexin-IV/Wrapper generates larger yet defined CG chamber grouping several lineages together. Axonal projections of newborn neurons are also altered in these conditions. Further, we link the loss of these 2 adhesion complexes specifically during development to locomotor hyperactivity in the resulting adults. Altogether, our findings identify a belt of adhesions building a neurogenic niche at the scale of individual stem cell and provide the proof of concept that niche properties during development shape adult behaviour.

Funding: This work has been funded by a starting package from Institut Pasteur/ LabEx Revive (ANR-10-LABX-0073), a JCJC grant from Agence Nationale de la Recherche (NeuraSteNic, ANR-17-CE13-0010-01) to P.S. and a Projet Fondation ARC from the Association pour la Recherche contre le Cancer to P.S. and A.B-L. was supported by a post-doctoral fellowship from the LabEx Revive (ANR-10-LABX-0073). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Here, we investigated the cellular cues driving the correct establishment of a structurally sophisticated and functional CG niche around individual NSC lineages and their impact on neurogenesis. We found that CG are able to group together clonally related cells regardless of their individual identity. Further, we discovered that lineage information and individual encasing are mediated by the existence of multiple adhesion complexes within the niche. First, the cell adhesion protein Neurexin-IV is expressed and crucial in NSC lineages to maintain their individual encasing, through its interaction with Wrapper, a protein with immunoglobulin domains present in the CG. The loss of Neurexin to Wrapper interaction results in large, defined CG chambers containing multiple NSC lineages. In parallel, Neuroglian appears to form strong homophilic interactions between cells of the same lineages, binding them together by providing higher adhesion compared to the weaker interaction between CG and NSC lineages. In absence of Neuroglian, NSC lineages are clumped together in a random fashion. As such, differential adhesion is a core mechanism of NSC lineage encasing. Adherens junctions are also present in NSC lineages; however, they appear mostly dispensable for individual encasing. In addition, Neurexin-IV and Neuroglian adhesions are important for correct axonal projections in the developing CNS. Further, we demonstrated that the loss of Neurexin-IV/Wrapper and Neuroglian adhesions specifically during development causes a hyperactive locomotor behaviour in the adult. Our findings unravel a principle of NSC niche organisation based on differential adhesion and link the adhesive property of the niche and NSC lineages during development to adult neurological behaviour.

The reliable formation of such precise chequerboard structure implies that CG integrate proper cellular cues to encase specific cells, while navigating between a density of diverse cell types. However, the nature of these cues and the importance of such stereotyped encasing of NSC lineages on NSC activity and the function of neuronal progeny remained to be identified.

( A ) Schematic of the Drosophila larval CNS depicting the localisation of the NSC lineages. The 2 main neurogenic regions are the CB, comprising 2 hemispheres, and the VNC. The OL corresponds to the precursor of the visual system and is organised as a neuroepithelial tissue, which will undergo conversion into NSCs. ( B ) Schematic of the Type I and Type II NSC lineages. Type II lineages are only found in the CB, with a number of 8 per hemisphere. Type I lineages populate both the CB and the VNC. Type I NSCs divide asymmetrically to give birth to a GMC, which itself divides once to generate 2 neurons or glial cells. Type II NSCs self-renew while generating one imINP. The imINP will mature in mINP, which will then divide to generate a GMC. ( C ) Confocal pictures representing the dorsal region of the larval CB (left panel) and ventral region of the VNC (right panel) 72 h after larval hatching (ALH72, at 25°C) labelled with markers for the CG membrane (Nrv2::GFP, green), glia nuclei (anti-Repo, yellow), NSC (anti-Dpn, grey), and neurons (anti-ElaV, magenta). ( D ) Schematic of the NSC niche (orthogonal view), made by the PG (brown), SPG (orange), CG (green), NSCs (grey), GMCs/INPs (blue) and N (light purple). One entire NSC lineage, composed of one NSC and its immature neuronal progeny, is encased within a seemingly continuous layer of CG membrane, forming a chamber. ( E ) Timeline of the encasing of NSC lineages by CG, parallelling NSC behaviour. At the beginning of larval stage (0 h after larval hatching, ALH0), NSCs are quiescent and not individually encased by the CG, whose network is not formed. Upon larval feeding, both NSCs and CG grow. Individual encasing of NSC (chamber formation) correlates with the time of its first division, the final step of NSC reactivation. The CG will then keep growing (extension) to adapt to the production of newborn secondary neurons and the increase in lineage size until the end of larval stage. ( F , G ) The CG chambers individually encase whole NSC lineages, including the NSC (NSC level) and its newborn neuronal progeny (neuron level) throughout its depth (orthogonal view). ( F - F” ) Confocal pictures of a VNC labelled with markers for the CG membrane (Nrv2::GFP, green), NSCs (anti-Dpn, grey), and neurons (anti-ElaV, light purple). The orthogonal view ( F ) shows the overall cellular organisation along the dorso-ventral axis of the VNC, with NSCs mostly localised ventral, and neuronal progeny below, more dorsal. Yellow boxes correspond to the close-up panels below. Dashed white lines indicate one NSC chamber. Horizontal white dashed lines indicate the planes of the ( F’ ) NSC and ( F” ) neuron levels. ( G - G” ) Schematics representing the respective pattern of NSC and CG membrane for the different views (orthogonal, NSC level, and neuron level) for a group of NSCs. Horizontal black lines in ( G ) indicate the planes of the ( G’ ) NSC and ( G” ) neuron levels. ALH, after larval hatching; CB, central brain; CG, cortex glia; CNS, central nervous system; GMC, ganglion mother cell; imINP, immature intermediate neural progenitor; INP, intermediate neural progenitor; mINP, mature intermediate neural progenitor; N, neuron; NSC, neural stem cell; OL, optic lobe; PG, perineurial glia; SPG, subperineurial glia; VNC, ventral nerve cord.

Fly NSCs are born during embryogenesis, during which they cycle to generate primary neurons in a first wave of neurogenesis. They then enter quiescence, a mitotically dormant phase from which they exit to proliferate through the activation of PI3K/Akt signalling in response to nutrition [ 8 , 9 ]. This postembryonic, second wave of neurogenesis generates secondary neurons that will make up 90% of the adult CNS and lasts until the beginning of pupal stage. NSCs finally differentiate or die by apoptosis after pupariation. Larval NSCs populate the different regions of the CNS, namely, the ventral nerve cord (VNC), the central brain (CB), and the optic lobe (OL) ( Fig 1A ). They nevertheless display distinct properties, mainly through different modes of division and expression of specific transcription factors ( Fig 1B ) [ 10 ]. Type I NSCs reside in the CB and VNC and divide asymmetrically to generate a smaller ganglion mother cell (GMC). GMCs further terminally divide to produce 2 neurons. Type II NSCs, found exclusively in the CB, represent a smaller population with only 8 cells per hemisphere [ 11 – 13 ]. Type II NSC self-renewal produces an intermediate neural progenitor (INP), which undergoes a limited number of asymmetric divisions to produce GMCs that will subsequently divide to give neurons.

Stem cells are multipotent progenitors driving the growth and regeneration of the tissue they reside in through the generation of differentiated cells. Their localisation within the tissue is restricted to carefully arranged cellular microenvironments, or niches, which control their maintenance and activity in response to local and systemic cues [ 1 – 3 ]. The niches comprise the stem cell themselves, their newborn progeny, and a number of cells of various origins and roles that support stem cell decisions. The diversity of cellular shapes and roles requires a precise spatial organisation to enable proper niche function towards all and every stem cells. Within the central nervous system (CNS) in particular, a highly structured organ dependent on the tight arrangement of cellular connections, the neural stem cell (NSC) niches are anatomically complex microenvironments that must form within such constraint. They comprise multiple cell types such as neurons, various glial cells, vasculature and immune cells [ 4 , 5 ], which are precisely organised with respect to NSCs. While studies have focused on the identification of signalling pathways operating in an established niche and controlling neurogenesis [ 5 – 7 ], how the niche is first spatially built around NSCs, and the importance of its architecture on neurogenesis, from stem cell division to the integration of the newborn neurons, are poorly understood.

Results

Occluding junction components are expressed in NSC lineages These findings prompted us to investigate the potential presence and role of other adhesion complexes that could provide intralineage cohesion and differential adhesion to sort NSC lineages from CG. Occluding junctions (tight junctions in vertebrate and septate junctions in Drosophila) [47,48] primarily perform a permeability barrier function to paracellular diffusion. However, they can also provide some adhesion between the cells they link. Drosophila septate junctions are formed by the assembly of cell surface adhesion molecules that can interact in cis or trans, in an homologous or heterologous fashion, and which are linked to the intracellular milieu by supporting membrane or cytoplasmic molecules [47]. A core, highly conserved tripartite complex of adhesion molecules comprises Neuroglian (Nrg), Contactin (Cont), and Neurexin-IV (Nrx-IV). Nrg, the Drosophila homologue of Neurofascin-155, is an L1-type family transmembrane protein, containing several immunoglobulin domains and is mostly homophilic. Cont, homologous to the human Contactin, also contains immunoglobulin domains, is GPI anchored, and only performs heterophilic interactions. Nrx-IV, homologous to the human Caspr/Paranodin, is a transmembrane protein with a large extracellular domain containing laminin-G domains and EGF repeats [49] and is able to set up heterophilic interactions. Several cytoplasmic or membrane-associated proteins also participate in septate junction formation, such as the FERM-family Coracle (Cora), the MAGUK protein Discs large (Dlg1), and the integral membrane Na K-ATPase pump (ATPα). We decided to perform a preliminary characterisation of the expression and function of septate junction components in NSC lineages. We found that Nrx-IV, Nrg, Cora, Dlg1, and ATPα are all present in NSC lineages when the chamber is formed (ALH72) and Cora also showed a staining along the CG interface (S6A Fig). We were not able to assess Cont due to lack of access to working reagents. Multiple septate junction components are thus expressed in NSC lineages, localising between cells of the same lineage. We then probed the importance of such expression in the individual encasing of NSC lineages by CG. We knocked down nrx-IV, nrg, dlg1, cont, and ATPα in NSC lineages from ALH0 using specific RNAi lines (S2 Table). Larvae from dlg1, cont, and ATPα knockdown died at early larval stages. From ATPα knockdown, few larvae still reached late larval stage, displaying restricted irregularities in the CG network. In contrast, nrx-IV and nrg knockdowns mostly survived and resulted in altered encasing of individual NSC lineages. We thus decided to focus on Nrx-IV and Nrg functions in NSC lineages. Interestingly, they both perform nervous system–specific roles outside of the septate junction. Nrx-IV is required in the embryonic CNS for axonal wrapping by the midline glia [50–52]. Neuronal Nrg is important for axonal guidance and dendritic arborization of peripheral neurons [53–55], as well as for the function and axon branching of specific larval CB neurons [56,57]. For both proteins, their role in NSC lineages during larval neurogenesis is, however, poorly known.

A glia to NSC lineages adhesion through Nrx-IV and Wrapper is required for individual encasing The dual role of Nrx-IV within and outside septate junctions is sustained by the existence of alternative splicing [51]. Nrx-IV can be produced as a septate junction isoform (Nrx-IVexon3) and a neuronal isoform outside of SJ (Nrx-IVexon4). Nrx-IV role within the embryonic CNS is through its recruitment by and binding to its glial partner Wrapper, another member of the immunoglobulin family [50–52]. We thus sought to assess whether the role of Nrx-IV in NSC lineages encasing by CG was dependent on Wrapper. Previous studies had reported that regulatory sequences in the wrapper gene drive in the CG during late larval stages [41,59]. In situ hybridization against wrapper mRNA (HCR RNA-FISH; see Methods) detected wrapper expression in CG throughout larval development and chamber formation (Fig 7A). We then asked whether knocking down wrapper in the CG would recapitulate the encasing phenotype found under nrx-IV loss of function in NSC lineages. We first checked the efficiency of RNAi knockdown in the CG using RNA FISH and found that it was specifically wiping out wrapper signal in the CG (S7A and S7B Fig), while preserving its known expression in the midline glia (see arrowheads in S7A Fig). Driving this RNAi line in the CG reproduced the highly characteristic pattern of large yet defined CG chambers (Fig 7B, observed at the neuron level) found during nrx-IV knockdown in NSC lineages (compare with Fig 6B). A similar result was obtained with a second RNAi line against wrapper (S7C Fig). We further quantified the loss of individual encasing by marking NSC lineages with Raeppli-NLS while driving wrapper RNAi in the CG and found that 25% of the NSC lineages was sharing a CG chamber with at least another one (Fig 7C and 7D). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 7. A CG to lineage interaction through Nrx-IV and Wrapper is required for individual encasing of NSC lineages by the CG. (A) Representative confocal pictures of the localisation of wrapper mRNA at ALH0, ALH24, ALH48, and ALH72 at 25°C. n ≥ 6 VNCs for all time points. wrapper mRNA (magenta) is detected through RNA FISH. CG membrane is visualised by cyp4g15-mtd::Tomato (green), NSCs are labelled with anti-Dpn (grey), and neurons are labelled with anti-ElaV (blue). Yellow arrowheads indicate examples of colocalisation between CG membrane and wrapper mRNA signal. (B) Representative confocal pictures of the thoracic VNC for a condition in which wrapper is knocked down by RNAi (cyp > wrapper RNAi, BDSC line 29561) in the CG (driver line Nrv2::GFP, tub-GAL80ts; cyp4g15-GAL4) at the NSC level, at the neuron level, and in orthogonal view. Larvae are dissected after 68 h at 29°C from ALH0. Control, cyp >—(x w1118) (n ≥ 10 VNCs) and cyp > wrapper RNAi (n ≥ 10 VNCs). CG membrane is visualised by Nrv2::GFP (green), and NSCs are labelled with anti-Dpn (grey). (C) Representative confocal picture of the thoracic VNC for a condition in which wrapper is knocked down by RNAi (BDSC line 29561) from ALH0 in CG, while NSC lineages are marked with the multicolour lineage tracing Raeppli-NLS (blue, white, orange, and red). Raeppli-NLS is induced at ALH0 using hs-Flp and under the control of the LexA/LexAop system (dpn enhancer), while wrapper RNAi is under the control of the GAL4/UAS system (dpn > raeppli-NLS + cyp > wrapper RNAi). Larvae are dissected after 72 h at 29°C from ALH0. See S1 Table for detailed genetics, timing, and conditions of larval rearing. CG membrane was visualised with Nrv2::GFP (green). Dashed white lines highlight examples of NSC lineages encased individually, and dashed yellow lines examples of NSC lineages encased together. (D) Quantification of the percentage of NSC lineages non-individually encased from (C). Control, dpn > raeppli-NLS (n = 10 VNCs) and dpn > raeppli-NLS + cyp > wrapper RNAi (n = 12 VNCs). Data statistics: generalised linear model (Binomial regression with a Bernoulli distribution). Results are presented as box and whisker plots. (E) Representative confocal pictures of thoracic VNCs for which combined NSC lineage and CG drivers (wor + cyp >; driver line Nrv2::GFP, wor-GAL4/CyO; cyp4g15-GAL4, tub-GAL80ts) were used to perform double RNAi knockdown against nrx-IV and RFP; wrapper and mCherry; and nrx-IV and wrapper. nrx-IV RNAi, BDSC line 32424; wrapper RNAi, VDRC line 105314. Larvae are dissected after 68 h at 29°C. CG membrane is visualised by Nrv2::GFP (green), and NSCs are labelled with anti-Dpn (grey). (F) Quantification of the CG membrane volume per NSC from (E). See Methods for details. nrx-IV + RFP RNAis (n = 8 VNCs), wrapper + mCherry RNAis (n = 9 VNCs), nrx-IV + wrapper RNAis (n = 9 VNCs). Data statistics: Kruskal–Wallis H test with Dunn’s multiple comparisons test. p = 0.0029 for the Kruskal–Wallis H test on grouped dataset. P values from Dunn’s multiple comparisons test are displayed on the graph. Results are presented as box and whisker plots. (G) Representative confocal picture of a thoracic VNC for a condition in which wrapper is overexpressed in NSC lineages from ALH0 (wor > wrapper, driver line Nrv2::GFP, wor-GAL4; tub-GAL80ts). Larvae are dissected after 68 h at 29°C. CG membrane is visualised by Nrv2::GFP (green), and NSCs are labelled with anti-Dpn (grey). (H) Quantification of the percentage of NSCs non-individually encased from (G). Control, wor >—(x w1118) (n = 7 VNCs), wor > wrapper (n = 7 VNCs). Data statistics: generalised linear model (Binomial regression with a Bernoulli distribution). Results are presented as box and whisker plots. For all box and whisker plots: whiskers mark the minimum and maximum, the box includes the 25th–75th percentile, and the line in the box is the median. Individual values are superimposed. The data underlying this figure’s quantifications can be found in S1 Data. ALH, after larval hatching; CG, cortex glia; FISH, fluorescent in situ hydridization; Nrx-IV, Neurexin-IV; NSC, neural stem cell; VNC, ventral nerve cord. https://doi.org/10.1371/journal.pbio.3002352.g007 These results suggest that Nrx-IV in NSC lineages interact with Wrapper in the CG for ensuring NSC individual encasing. To strengthen this relationship, we performed genetic interactions between nrx-IV and wrapper, comparing CG phenotype between nrx-IV only knowdown, wrapper only knockdown, and double nrx-IV and wrapper knockdown in both NSC lineages and CG (see Methods, S1 Table, and S7D Fig for expression of the combined driver lines; single knockdowns were dose compensated). At a qualitative level, we first observed that all 3 combinations displayed the characteristically large and defined CG chambers associated with nrx-IV loss of function in NSC lineages and wrapper loss of function in CG (Fig 7E). As we could not add a clonal analysis tool to these already complex genotypes, we decided to quantify the volume of CG membrane as a proxy for the density of encasing (i.e., grouped NSC lineages means larger chambers and, conversely, less CG membrane per NSC lineage). We first confirmed the validity of our proxy by assessing whether it was able to detect nrx-IV phenotype. To do so, we measured CG volume between a control condition, shg RNAi in NSC lineages, and nrx-IV RNAi in NSC lineages and found that nrx-IV knockdown resulted in a significantly lower CG volume compared to control, while shg knockdown did not show a decrease (S7E Fig). We then used the same approach to assess NSC encasing upon double knockdown of nrx-IV and wrapper and uncovered that it was significantly lower than upon the individual knockdown of either, which displayed a similar decrease (Fig 7F). These data pinpoint a greater effect of the combined nrx-IV and wrapper knockdowns than their individual contribution on NSC lineage encasing by the CG, showing that the genotype of one affects the phenotype of the other. This demonstrates an epistatic interaction between nrx-IV and wrapper in this context. We then wondered how the Nrx-IV to Wrapper interaction would fit in the differential adhesion hypothesis. If CG to NSC lineage adhesion is indeed weaker than intralineage adhesion (Fig 2F, panel II.2; A(LNSC-CG) < A(LNSC)), overexpression of Wrapper in NSC lineages should not affect the sorting between NSC lineages and CG, since A(LNSC) would still be superior to A(LNSC-CG). However, if Nrx-IV to Wrapper interaction is stronger than the sum of intralineage adhesions, then forcing its establishment within the lineage would favour the random grouping of NSC lineages together. We found that misexpressing wrapper in NSC lineages from larval hatching (ALH0), while successful (as confirmed by staining with an antibody that can also detect endogenous Wrapper; see S7F Fig), resulted in very little alteration of CG encasing of individual NSC lineages (Fig 7G and 7H). These data plead in favour of a CG to NSC lineage adhesion through Nrx-IV and Wrapper being weaker than the sum of intralineage adhesions. Taken together, our results suggest that Nrx-IV in NSC lineages partners with Wrapper in the CG, outside of a septate junction function. This interaction is essential to produce individual encasing of NSC lineages by CG, and in its absence, NSC lineages are randomly grouped in well-defined, larger chambers.

Intralineage adhesion through Nrg drives cell sorting and individual encasing by CG Like Nrx-IV, the dual role of Nrg in and outside of septate junction comes from differential splicing [60]. Nrg comes in 2 isoforms, with the same extracellular domain but different intracellular parts (Fig 9A). While the short isoform, Nrg167, localises in the septate junction of epithelial tissues, the long isoform, Nrg180, is expressed in neurons of the developing central and peripheral nervous systems [55,56,60]. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 9. Individual encasing of NSC lineages relies on strong intralineage adhesion through Nrg. (A) Schematic depicting the 2 isoforms for Nrg, Nrg167, and Nrg180. Only the intracellular C-terminal part differs and will modulate downstream signalling. (B) Representative confocal close-up picture of the localisation of the Nrg180 isoform in NSC lineages, in a thoracic VNC at ALH72 at 25°C. n = 8 VNCs. All Nrg isoforms are monitored through a Nrg::GFP protein trap (magenta) and the Nrg180 isoform is detected with a specific antibody (BP104, light blue). (C) Representative confocal close-up picture of the respective localisations of the Nrg167 and Nrg180 isoforms in NSC lineages, in a thoracic VNC at ALH72 at 25°C. n = 6 VNCs. The Nrg167 isoform is visualised by a protein trap in the nrg gene leading to the preferential expression of this isoform (Nrg167::GFP, yellow). The Nrg180 isoform is detected with a specific antibody (BP104, light blue). The dashed white line highlights the perimeter of the NSC devoid of BP104 signal. (D) Representative confocal picture of a thoracic VNC and close-up for nrg180 overexpression in the CG from ALH0 (cyp > nrg180, driver Nrv2::GFP, tub-GAL80ts; cyp4g15-GAL4). Larvae are dissected after 68 h at 29°C. CG membrane is visualised by Nrv2::GFP (green), NSCs are labelled with an anti-Dpn (grey), and Nrg180 is detected with a specific antibody (BP104, magenta). (E) Representative confocal pictures of thoracic VNCs for nrg167 and nrgGPI overexpression in the CG from ALH0 (cyp > nrg167 and cyp > nrgGPI, respectively, driver Nrv2::GFP, tub-GAL80ts; cyp4g15-GAL4). Larvae are dissected after 68 h at 29°C. CG membrane is visualised by Nrv2::GFP (green), and NSCs are labelled with an anti-Dpn (grey). (F) Quantification of the percentage of NSCs non-individually encased from (D, E). Control, cyp >—(x w1118) (n = 10 VNCs), cyp > nrg180 (n = 11 VNCs), cyp > nrg167 (n = 7 VNCs), and cyp > nrgGPI (n = 6 VNCs). Data statistics: generalised linear model (Binomial regression with a Bernoulli distribution) for nrg180. For nrg167 and nrgGPI, there is no variance and, thus, statistics cannot be applied (NA, non-applicable). Results are presented as box and whisker plots. (G) Representative confocal picture of a thoracic VNC and close-up for Nrg180 overexpression in the NSC lineages from ALH0 (wor > nrg180, driver Nrv2::GFP, worniu-GAL4; tub-GAL80ts). Larvae are dissected after 68 h at 29°C. CG membrane is visualised by Nrv2::GFP (green), NSCs are labelled with an anti-Dpn (grey), and Nrg180 is detected with a specific antibody (BP104, magenta). (H) Representative confocal pictures of thoracic VNCs for nrg167 and nrgGPI overexpressed from ALH0 in NSC lineages (wor > nrg167 and wor > nrgGPI, respectively, driver line Nrv2::GFP, wor-GAL4; tub-GAL80ts). Larvae are dissected after 68 h at 29°C. CG membrane is visualised by Nrv2::GFP (green), and NSCs are labelled with anti-Dpn (grey). (I) Quantification of the percentage of NSCs non-individually encased from (G, H). Control, wor >—(x w1118) (n = 8 VNCs), wor > nrg180 (n = 8 VNCs), wor > nrg167 (n = 7 VNCs), and wor > nrgGPI (n = 6 VNCs). Data statistics: generalised linear model (Binomial regression with a Bernoulli distribution). p = 1.93 × 10−113 for the grouped dataset. P values for individual comparisons test are displayed on the graph. Results are presented as box and whisker plots. For all box and whisker plots: whiskers mark the minimum and maximum, the box includes the 25th–75th percentile, and the line in the box is the median. Individual values are superimposed. The data underlying this figure’s quantifications can be found in S1 Data. ALH, after larval hatching; CG, cortex glia; Nrg, Neuroglian; NSC, neural stem cell; VNC, ventral nerve cord. https://doi.org/10.1371/journal.pbio.3002352.g009 We first determined which isoform is expressed in NSC lineages during the larval stage, taking advantage of isoform-specific tools. Staining of Nrg::GFP CNS (ALH72) with an antibody (BP104) specifically recognising the Nrg180 isoform [60] revealed that Nrg180 localises in the membranes of all cells from NSC lineages, but not in known septate junctions within the tissue (yellow arrowheads) (Figs 9B and S9A). In neurons, Nrg180 was not only found in the cell body but appeared also enriched in their axonal bundle, a localisation reported previously [61]. We then took advantage of an Nrg::GFP fusion shown in other tissues to preferentially target the Nrg167 isoform (called Nrg167::GFP; [53,55]). Nrg167::GFP also appeared enriched between cells of the same NSC lineage, where it colocalised with BP104 staining, except on the NSC perimeter, devoided of BP104 (Figs 9C and S9B; see dashed white line for lack of BP104). In contrast, only Nrg167 is detected in septate junctions. We then wondered whether nrg knockdown was able to lower the levels of both isoforms. nrg knockdown in NSC lineages completely depleted the BP104 signal (S9C and S9D Fig; mean of 0.1 normalised to control). Nrg167::GFP levels also were strongly decreased upon nrg knockdown (S9E and S9F Fig; mean of 0.25 normalised to control). While it appears in a lower fashion than Nrg180, it might be due to a higher stability of the Nrg167::GFP fusion, while endogenous Nrg180 was detected with an antibody. Taken together, these data suggest that the 2 isoforms of Nrg are expressed, and can be efficiently knocked down, in NSC lineages. Mostly homophilic interactions (between the same or different isoforms) have been reported for Nrg. We thus wondered whether an Nrg to Nrg interaction within the NSC lineages could fulfil the role of an intralineage adhesion stronger than a CG to NSC adhesion. Since nrg knockdown in CG (S8C Fig) did not recapitulate nrg knockdown in NSC lineages, the homophilic Nrg interactions between CG and NSC lineages, if existing, are not involved in individual encasing. We then assessed the relevance of intralineage Nrg interactions in the differential adhesion hypothesis (Fig 2F, panel II.2). If such adhesion is stronger than the CG to NSC lineage interaction, then expressing Nrg in CG would force CG to interact with each other. Strikingly, misexpressing Nrg180 in CG from larval hatching (ALH0) resulted in altered CG morphology and loss of individual encasing of NSC lineages (Fig 9D). CG membranes displayed local accumulation as well as unusual curvature, and NSCs were not separated from each other by CG anymore but were rather found grouped close to each other. Overexpressing Nrg167 in CG (from ALH0) produced an even more dramatic phenotype, with localised, compact globules of CG membranes and the complete lack of individual encasing of NSC lineages (Fig 9E). Interestingly, overexpression of an NrgGPI construct in which the transmembrane and cytoplasmic domains are replaced by a GPI anchor signal [62] also resulted in aggregated CG and clustered NSC lineages (Fig 9E). This shows that intracellular signalling through the divergent C-terminal domain is not required for this sorting of CG and NSC lineages, but rather that adhesion through the extracellular part mediates this effect. The quantification of NSC encasing upon expression of the different Nrg isoforms in CG confirmed our interpretation (Fig 9F). Altogether, these results demonstrate that providing Nrg homophilic interactions in the CG is sufficient to segregate them from the whole population of NSC lineages, which they normally bind to through a weaker Nrx-IV to Wrapper interaction. This further suggests that Nrg homophilic adhesions between cells of the same NSC lineage are responsible for keeping these cells together and excluding the CG. If Nrg interactions are indeed responsible for providing binding between cells of the same NSC lineage, including the stem cell, one consequence is that NSCs could bind to each other. Interestingly, Nrg appears expressed in NSCs only after their encasing (see Fig 8A). This fits the idea that early on, when NSCs are not encased yet and separated from other NSCs by the CG, A (NSC-NSC) is kept low. As such, a precocious expression of Nrg in NSCs would be predicted to lead to their grouping (and further the grouping of their neuronal lineages) in a CG chamber. Strikingly, expressing either of the 3 Nrg isoforms from ALH0 resulted in multiple, larger, and well-defined CG chambers containing several NSCs, all in a similar fashion (Fig 9G–9I). As expressing NrgGPI also led to the grouping of NSC lineages, it implies that the adhesive role of Nrg is responsible for such effect. This contrasts with the lack of effect of misexpressing Wrapper in NSC lineages (also from ALH0; see Fig 7G and 7H), showing that not all adhesion complexes can lead to A(NSC-NSC) high enough to group NSCs together. These results suggest that a proper timing in establishing intralineage adhesion through Nrg is instrumental in ensuring the individual encasing of NSC lineages by CG.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3002352

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.