Drosophila ßHeavy-Spectrin is required in polarized ensheathing glia that form a diffusion-barrier around the neuropil

Abstract

In the central nervous system (CNS), functional tasks are often allocated to distinct compartments. This is also evident in the Drosophila CNS where synapses and dendrites are clustered in distinct neuropil regions. The neuropil is separated from neuronal cell bodies by ensheathing glia, which as we show using dye injection experiments, contribute to the formation of an internal diffusion barrier. We find that ensheathing glia are polarized with a basolateral plasma membrane rich in phosphatidylinositol-(3,4,5)-triphosphate (PIP₃) and the Na⁺/K⁺-ATPase Nervana2 (Nrv2) that abuts an extracellular matrix formed at neuropil-cortex interface. The apical plasma membrane is facing the neuropil and is rich in phosphatidylinositol-(4,5)-bisphosphate (PIP₂) that is supported by a sub-membranous ß_Heavy-Spectrin cytoskeleton. ß_Heavy-spectrin mutant larvae affect ensheathing glial cell polarity with delocalized PIP₂ and Nrv2 and exhibit an abnormal locomotion which is similarly shown by ensheathing glia ablated larvae. Thus, polarized glia compartmentalizes the brain and is essential for proper nervous system function.

Fig. 1. Development of ensheathing glia.

Representative images are shown. a–d Dissected larval CNS of increasing age with the genotype [83E12-Gal4, UAS-CD8::GFP], stained for GFP (green), Repo (magenta) and neuronal membranes (anti-HRP, blue), anterior is up. The positions of the orthogonal section shown in a’–d’ is indicated by a dashed line. a, a’ In a stage 16 embryo ensheathing glial cells have not yet covered the neuropil (arrowheads). b, b’ First instar larval CNS. c, c’ Second instar larval CNS. d, d’ CNS of a wandering third instar larva. The arrows point towards dorsal protrusions of the ensheathing glia engulfing dorsal neurons. e MCFO labeling of ensheathing glia in third instar larva stained for the expression of V5 (green), HA (red), and FLAG and OLLAS epitopes (blue). flp expression was induced for one hour during first instar larval stage. Note that ensheathing glia tile the ventral nerve cord. f Two distinct ensheathing/wrapping glia cells cover the nerve root and part of the neuropil (arrowheads). g Ensheathing glia occupy specific territories in the neuropil. h, i Third instar larval nerve cord with the genotype [55B12-Gal4, 83E12-LexA, UAS-CD8::mCherry, LexAop-GFP]. All cortex glia cells are labelled by mCherry expression (magenta). Ensheathing glial cells are labelled by GFP expression (green). The dashed line indicates the position of the orthogonal view shown in (i). j Schematic view on a cross section through a hemineuromere indicating the position of the different glial cells. Astrocyte-like cells and ensheathing glial cells localize close to the neuropil and the axons connecting the neuropil with the periphery. The cortex glia covers lateral and ventral neuronal cell bodies. Scale bars a–i are 50 µm.

Fig. 2. Larval development of the ensheathing glia.

a, b Schematic view on a first instar larval central nervous system. All neuropil-associated glial cells were annotated in a serial section TEM volume. Yellow dots indicate the position of the nuclei of astrocyte-like cells, magenta dots indicate the positions of the ensheathing glial cell nuclei, green dots indicate the positions of the ensheathing/wrapping glial nuclei. The arrow points to a segmental position where only one instead of two ensheathing/wrapping glial cell was identified. c Representative section through the fourth abdominal hemineuromer of a L1 larva showing the position of three different ensheathing glial cells around the neuropil. d Representative section through the fourth abdominal neuromere of a third instar larval ventral nerve cord. Note that ensheathing glia completely cover the neuropil. It was not determined whether the red marked glial processes belong to the blue or green ensheathing/wrapping glial cell. Scale bar (c, d): 10 µm, (e–h): 5 µm, (i, j): 0.2 µm. e Ensheathing glia in a first and (f) in a third instar larva. g Ensheathing/wrapping glial cells in a first and (h) in a third instar larva. Note the thin processes that engulf cell bodies of dorsally located neurons (asterisk). i On the ventral face of the neuropil an ensheathing glial cell and a cortex glial cell form a two-layered sheath between cortex and neuropil. j A highly multilayered glial cell layer is found at the dorsal face of the neuropil.

Fig. 3. Larval thoracic ensheathing glial cells can divide to generate adult brain ensheathing glia.

Representative images are shown. a Control third instar larvae with nuclei of ensheathing glia labelled [83E12-Gal4, UAS-Lam::GFP]. Nuclei are stained using DAPI. The boxed area is shown in high magnification in (b). c Adult control brain. The number of ensheathing glial nuclei is quantified in (m, n, n = 5 brains for all genotypes). d–f Expression of an activated FGF-receptor leads to an increase in the number of larval ensheathing glia in larval thoracic neuromeres (arrowheads). The white boxed area is shown in higher magnification (e). g–i Upon expression of string dsRNA in all ensheathing glia the number of larval ensheathing glial cells is slightly reduced. For quantification see (m). The boxed area in the larval CNS (g) is shown in high magnification in (h). i Expression of string dsRNA in all ensheathing glia reduces the number of ensheathing glial cells in the adult CNS. j–l A similar reduction in the number of ensheathing glial cells is observed following expression of fzr. The boxed area in (j) is shown in high magnification in (k). Scale bars larval CNS (a, b, d, e, g, h, j, k) are 50 µm, scale bars for adult CNS (c, f, i, l) are 100 µm. Quantification of the ensheathing glial cell number in five larval brains (m) and in five adult brains (n) using Imaris (unpaired t-test, two-tailed). The standard deviation is indicated. The optic lobes and the tract ensheathing glial cells were excluded in the quantification. For larva: control – string^dsRNA: *p = 0.0192; control – fzr: **p = 0.0013; control - λhtl: ****p =< 0.0001; for adult: control – string^dsRNA: ***p = 0.0002; control – fzr: **p = 0.0018; control - λhtl: **p = 0.0057. (O) Quantification of DAPI intensity in 30 larval abdominal, 30 larval thoracic, and 10 adult thoracic ensheathing glial and neuronal nuclei. Source data are provided as a Source Data file.

Fig. 4. Ablation of ensheathing glia causes compensatory growth and increased proliferation of astrocyte-like cells.

Representative images are shown. a–f Third instar larval brains stained for Rumpel (green, expressed by ensheathing glia), Nazgul (magenta, expressed by astrocyte-like cells) and HRP (blue, neuronal membrane marker). a Maximum projection of a confocal stack of a control larval CNS. The dashed line indicates the position of the orthogonal view shown in (c). b Dorsal view of a control ventral nerve cord. Note the regular positioning of the astrocyte-like cells (magenta). c Orthogonal section of astrocyte-like cells which are labelled using anti-Nazgul staining (white, arrowheads point to short protrusions). d, e Upon ablation of the ensheathing glia [UAS-rpr; 83E12-Gal4^AD; repo-Gal4^DBD, UAS-hid], Rumpel expression is greatly reduced. The dashed line indicates the position of the orthogonal view shown in (f). e Dorsal view of a larval ventral nerve cord following ensheathing glia ablation. The position and the morphology of the astrocyte-like cells appears disorganized. f Orthogonal section. Note, that astrocyte-like glial cells develop long cell processes that appear to encase the entire neuropil in the absence of ensheathing glia (arrowheads). g Dorsal view on a thoracic neuromere of an adult control fly. h Thoracic neuromere of an adult fly lacking ensheathing glial cells. The number of astrocyte-like cells increases. i Quantification of the number of astrocyte-like cells in the ventral nerve cord of adult flies (n = 6 brains, **p = 0.0018, unpaired t-test, two-tailed, standard deviation is indicated). j Upon ablation of the ensheathing glia, longevity is reduced (n = 200 mated females, ****p = 4.67365E-83, Log-rank test, two tailed, standard deviation is indicated). Scale bars are: a–f 50 µm, g, h 100 µm. Source data are provided as a Source Data file.

Fig. 5. Ensheathing glial cells form an internal barrier around the neuropil.

a Schematic depiction of the dye injection experiments. Fluorescently labelled 10 kDa dextran was injected into the left hemisphere of a wandering third instar larva. Diffusion of the dye was monitored at two positions. In the neuropil (N) and in the cortex (C) harboring all neuronal cell bodies. b, b’ Stills of a representative movie of a larval brain lobe with the genotype [n = 5; 83E12-Gal4, UAS-CD8::GFP] following dye injection, time is indicated. c Quantification of dye diffusion rate in larvae of the genotype [83E12-Gal4, UAS-CD8::GFP] (blue line) and [UAS-rpr; 83E12-Gal4^AD, repo-Gal4^DBD, UAS-hid] (red line) (n = 5, standard deviation is shown, *p = 0.0181; unpaired t-test, two-tailed). d, d’ Stills of a representative movie of a larval brain lobe with the genotype [n = 5, UAS-rpr; 83E12-Gal4^AD; repo-Gal4^DBD, UAS-hid] following dye injection. Scale bars are 50 µm. Source data are provided as a Source Data file.

Fig. 6. Ensheathing glial cells show polarized plasma membrane domains.

Representative images are shown. a Projection of a confocal stack of a third instar larval ventral nerve cord coexpressing PH-AKT-GFP and PH-PLCδ-mCherry in ensheathing glial cells (EG) [83E12-Gal4, UAS-PH-AKT-GFP, UAS-PH-PLCδ-mCherry]. The dashed white line indicates the position of the orthogonal section shown in (b). b Orthogonal section showing the dorsal aspect of the neuropil. The boxed area is shown in higher magnification in (c–e). The dashed areas were subsequently used for quantification of GFP and mCherry localization. f Quantification of GFP/mCherry distribution in larval and adult ensheathing glia. The ratio of red (PLCδ-mCherry shown in magenta) and green (PH-AKT-GFP) fluorescent pixels of the areas indicated in (c) is plotted. The mean and standard deviation is shown. g Schematic view of a dorsal ensheathing glial cell. The neuropil facing domain is characterized by a high PIP₂ content, whereas PIP₃ is concentrated on the baso-lateral domain, where Nrv2 is predominantly localized, too. The blue dot indicates the position of a neuron. h Coexpression of Nrv2::GFP and 83E12-Gal4, UAS-mCherry in the ventral nerve cord of third instar larva. The boxed area is shown in higher magnification in (i, j). Note the preferential localization of Nrv2 at the basolateral cell domain of the ensheathing glia (arrows). Only little Nrv2 is found at the apical domain (open arrow head). Additional expression of Nrv2 is seen in the blood–brain barrier (filled arrowhead). k Quantification of polarized Nrv2::GFP localization. The mean and standard deviation is shown. Scale bars in a, b, h: 50 µm; in b–e, i, j: 25 µm. Source data are provided as a Source Data file.

Fig. 7. The ensheathing glial cells are flanked by extracellular matrix.

Representative images are shown. Expression of Integrin α and different extracellular matrix components as detected by using gene trap insertion lines. a–e Endogenously YFP-tagged Integrin α encoded by the inflated gene. f–j Endogenously GFP-tagged heparan sulfate proteoglycan Dally protein encoded by dally::GFP. k–o Endogenously tagged Collagen IV protein encoded by viking::GFP. p–t Endogenously tagged Perlecan encoded by trol::GFP. The different developmental stages are indicated. a, i, m, s The boxed areas are shown in the bottom row. Glial cells are in magenta (anti-Repo staining), neuronal membranes are in blue (anti-HRP staining) and the GFP-tagged proteins are in green (anti-GFP staining). a, b, e During late larval stages, prominent Inflated expression is seen at the blood–brain barrier (arrows) and weak expression is seen around the neuropil at the position of the ensheathing glia (arrowhead). c, d In adults, Integrin α is still found at the blood–brain barrier (arrows) and becomes more prominent around the neuropil (arrowheads). Note that strongest Inflated expression is detected in larval brains. No Inflated expression can be detected in the abdominal neuromeres (asterisk). f, g During the larval stage Dally is expressed at the blood–brain barrier (arrows) and in the CNS cortex (asterisks). A slightly stronger localization of Dally can be detected at the position of basal ensheathing glia processes (arrowhead). h–j In adults, Dally is enriched at the position of the ensheathing glia (arrowheads). k, l In the larval CNS Collagen IV is found at blood–brain barrier (arrows) but not within the nervous system. m–o Collagen IV is detected at the neuropil-cortex interface in adult stages (arrowheads in m, n, boxed area is shown in (o)). p, q Trol is detected at the larval blood–brain barrier (arrows) but not around the neuropil. r–t In adults, prominent Trol localization is seen at the position of the ensheathing glia (arrowheads). The boxed area in (s) is shown at higher magnification in (t). Scale bars are: larval CNS 50 µm, adult CNS 100 µm except for e: 25 µm; j, o, t: 50 µm.

Fig. 8. ß_H-Spectrin shows a polar distribution in ensheathing glia.

Representative images are shown. a Schematic view of the karst locus. Transcription is from left to right. Seven different karst mRNAs are generated from two promoters (P1, P2) as indicated (kst-RA - RH). The ß_H-Spectrin proteins PE and PG differ in a C-terminal exon indicated by brackets. The position of two MiMIC insertions and the EP insertion used for gain of function experiments is indicated, the blues line denotes the position of the peptide used for immunization. b Confocal view of the surface of a third larval instar brain of a kst^{MiMIC03134::GFP} animal that expresses mCherry in all ensheathing glia [83E12-Gal4, UAS-CD8::mCherry]. The dashed line indicates the position of the orthogonal view shown in (c, d). Expression at the blood–brain barrier forming subperineurial glia (arrows) and at the apical domain of the ensheathing glia (arrowheads) can be detected. For quantification see (k). e Similar view as shown in (b) of a control larva stained for Kst expression using a polyclonal antiserum. The dashed line indicates the position of the orthogonal view shown in (f). f Localization of Kst protein is weakly seen in the blood–brain barrier (arrow) and the ensheathing glia (arrowhead). g Orthogonal view of a kst^{MiMIC03134::GFP} animal expressing kst^dsRNA in the ensheathing glia (EG) [83E12-Gal4, UAS-kst^dsRNA]. Expression of Karst in ensheathing glia cannot be detected anymore (asterisk). Expression in the blood–brain barrier is unaffected (arrow). h, i The Trojan Gal4 element in kst^MiMIC03134 directs Gal4 expression in the kst P2 pattern. Lamin::GFP localizes to ensheathing glial nuclei (arrows), some cells of the blood–brain barrier (arrowheads) and some cells in the position of tracheal cells (asterisks). j Quantification of Karst localization in different membrane domains of ensheathing glial cells, quantification as described in Fig. 6. Scale bars are 50 µm except d: 10 µm. Source data are provided as a Source Data file.

Fig. 9. The ß_H-Spectrin cytoskeleton is required for glial polarity.

Representative images are shown. a Orthogonal view of the dorsal aspect of a third instar control ventral nerve cord stained for ensheathing glial (EG) morphology [83E12-Gal4, UAS-CD8::GFP]. b Third instar larval brain of a zygotic karst null mutant larvae with the genotype [kst^MiMIC13613 / kst^MiMIC13613, 83E12-Gal4, UAS-CD8::GFP]. The absence of all ß_H-Spectrin protein affects the morphology of the dorsally located ensheathing glial cells. c–e Third instar larval ventral nerve cord with reduced kst expression in ensheathing glia [83E12-Gal4, UAS-kst^dsRNA, UAS-PH-AKT-GFP, UAS-PH-PLCδ-Cherry]. d PH-PLCδ-Cherry binds to PIP₂ and shows an even distribution between apical and basolateral plasma membrane domains. For quantification see (i). Quantification was performed as described in Fig. 6. e PH-AKT-GFP binds to PIP₃ and is distributed in a polarized manner. For quantification see (i). f–h Third instar larval ventral nerve cord with increased kst expression in ensheathing glia [83E12-Gal4, kst^{P{EPgy2}EY01010}, UAS-PH-AKT-GFP, UAS-PH-PLCδ-Cherry]. Note, the variable phenotype noted upon karst overexpression. Ensheathing glia with almost normal morphology (open arrowhead) is next to hyperconvoluted ensheathing glia (asterisk). Quantification of PIP₂ and PIP₃ localization is shown in (j). i, j Quantification was performed as described in Fig. 6. The mean and the standard deviation is shown. k Quantification of the distribution of Nrv2::GFP using an unpaired t-test, two-tailed. *p_{control-kstdsRNA} = 0.0474; ****p_{control-kstEP} < 0.001. The mean and the standard deviation is shown. l–n In karst mutant larvae, Nrv2 is found in ensheathing glia cell processes (open arrowhead) that face the neuropil. The dashed line indicates the neuropil ensheathing glia boundary. o–q Upon ensheathing glial specific kst knockdown, Nrv2 localizes to the basolateral cell domains, for quantification see (k). Filled arrowhead indicates the blood–brain barrier. r–t Upon overexpression of Karst, polarized localization of Nrv2 is less apparent in ensheathing glial cell processes (normal morphology: open arrowheads, hyperconvoluted morphology: asterisk), for quantification see (k). Scale bars are 50 µm. Source data are provided as a Source Data file.

Fig. 10. Karst is required in ensheathing glia for normal locomotor behavior.

a Exemplary locomotion tracks of control wandering third instar larvae with the genotype [83E12-Gal4^AD, repo-Gal4^DBD, UAS-GFP^dsRNA]. Larvae move in long run phases. b Upon kst knockdown in ensheathing glia (EG) only [83E12-Gal4^AD, repo-Gal4^DBD, UAS-kst^dsRNA], larvae do not explore the tracking arena. c The insertion of a Trojan Gal4 element into the first coding exon of kst-RE(RG) generates an PE and PG isoform specific mutant (see Fig. 8). Larvae with the genotype [kst^TrojanGal4 / Df(3 L)ED208] show a comparable locomotion phenotype as RNAi knockdown larvae. d Locomotion phenotype of zygotic karst null mutant larvae with the genotype [kst^MiMIC13613 / Df(3 L)ED208]. e Locomotion phenotype observed following specific ablation of the ensheathing glia [UAS-rpr; 83E12-Gal4^AD; repo-Gal4^DBD, UAS-hid]. f–h Quantification of peristalsis efficiency, velocity and bending distribution between control larvae, kst knockdown and ensheathing glia ablated animals. Box plots show median (line), boxes represent the first and third percentiles, whiskers show standard deviation, diamonds indicate outliers. i–k Quantification of the same parameters comparing wild type Canton S with kst null larvae and a kst^Trojan control strain that carries the Trojan insertion in a non-productive orientation [kst^{MiMIC03134-Trojan-C} in trans to the Df(3 L)ED208] with the isoform specific kst^{PE, PG} mutant [kst^{MiMIC03134-TrojanGal4} / Df(3 L)ED208. Quantification Wilcoxon rank-sum test, two tailed, n = 50. Peristalsis Efficiency [mm/wave]: GFP^dsRNA – kst^dsRNA: ****p = 2.65E-32 GFP^dsRNA – ablation; ****p = 3.22E-30; velocity [mm/sec]: GFP^dsRNA – kst^dsRNA: **** p = 3.81E-11; GFP^dsRNA – ablation: ****p = 1.54E-05; Peristalsis Efficiency [mm/wave]: Canton S – kst^MiMIC / Df: ****p = 3.98E-20; Canton S – Trojan ^C/ Df: ****p = 4.61E-13; Trojan^C – Trojan^Gal4 / Df: ****p = 1.22E-13; velocity [mm/sec]: Canton S – kst^MiMIC / Df: ****p = 1.15E-06; Canton S – Trojan^C/ Df: ****p = 1.21E-03; Trojan^C – Trojan^Gal4 / Df: ****p = 5.94E-13. Source data are provided as a Source Data file.