Wrapping glia regulates neuronal signaling speed and precision in the peripheral nervous system of Drosophila

Abstract

The functionality of the nervous system requires transmission of information along axons with high speed and precision. Conductance velocity depends on axonal diameter whereas signaling precision requires a block of electrical crosstalk between axons, known as ephaptic coupling. Here, we use the peripheral nervous system of Drosophila larvae to determine how glia regulates axonal properties. We show that wrapping glial differentiation depends on gap junctions and FGF-signaling. Abnormal glial differentiation affects axonal diameter and conductance velocity and causes mild behavioral phenotypes that can be rescued by a sphingosine-rich diet. Ablation of wrapping glia does not further impair axonal diameter and conductance velocity but causes a prominent locomotion phenotype that cannot be rescued by sphingosine. Moreover, optogenetically evoked locomotor patterns do not depend on conductance speed but require the presence of wrapping glial processes. In conclusion, our data indicate that wrapping glia modulates both speed and precision of neuronal signaling.

Fig. 1. Generation of a wrapping glia driver.

Confocal projection of larval filet preparations of the genotypes indicated. Representative images are shown taken from >10 animals analyzed for each genotype. a Third instar larva with the genotype [nrv2-stGFP]. Broad GFP expression is detected in the CNS. Note the restricted expression in the peripheral nervous system which corresponds to the wrapping glia (arrowheads). b Third instar larva of the genotype [hs-Flp; nrv2-Gal4/MCFO-2]. flp expression of the multicolor FlpOut construct was induced by 1 h 37 °C heat shock during first instar stage. Larvae were stained for HA (green), V5 (red), and FLAG (blue). c Same animal as in (a). Expression of stRed is observed only in the CNS and no expression is found in the wrapping glia. d–f Overlay of nrv2-GFP (green) and 90C03 > dsRed (red) expression. Note the complete overlap of dsRed (e) and GFP expression in the CNS (f). g Young third instar larva with the genotype [hs-Flp; nrv2-Gal4/MCFO-2; 90C03-Gal80]. flp expression was induced by 1 h 37 °C heat shock during first instar stage. Larvae were stained for HA (green), V5 (red), and FLAG (blue). h Living third instar larva of the genotype [nrv2-Gal4, UAS-CD8GFP; 90C03-Gal80/90C03-Gal80]. Note strong expression at the anterior tip of the larva. Scale bars are 250 µm (a–c, g, h) and 100 µm (d–f).

Fig. 2. FGF-receptor Heartless controls differentiation of the wrapping glia.

a, b Filet preparations of a third instar larvae with a genotypes as indicated stained for Repo expression (red), GFP (green), and HRP (blue). Representative images are shown taken from >10 animals analyzed for each genotype. b Note that upon expression of htl^DN, the wrapping glia appears thinner (arrows). Scale bar is 200 µm. c, d Electron microscopic images of segmental nerves taken at about 160 µm distance to the tip of the ventral nerve cord of wandering third instar larvae. Five or more specimens were fixed and embedded in a filleted form. Representative images are shown for each genotype. Scale bar is 2 µm. c Cross-section of a control nerve (white¹¹¹⁸). The nerve is surrounded by a neural lamella (nl). The perineurial glia (pg) and the subperineurial glia (spg) form the blood–brain barrier. The wrapping glia (wg) engulfs all axons (ax). The wrapping glia is highlighted by purple staining. d Cross-section of a nerve from an animal of the genotype [nrv2-Gal4; UAS-htl^DN]. Note that differentiation of the wrapping glia (purple shading) is compromised. A similar phenotype is seen in larvae of the genotype [nrv2-Gal4/UAS-htl^DN; 90C03-Gal80/90C03-Gal80], see Supplementary Fig. 2).

Fig. 3. Wrapping glia affects neuronal differentiation and behavior.

a Distribution of axonal size in 40 nm² bins of control larvae ([nrv2-Gal4, UAS-GFP; 90C03-Gal80/90C03-Gal80], 2046 axons of 26 nerves) and of larvae expressing htl^DN in wrapping glia ([nrv2-Gal4/UAS-htl^DN; 90C03-Gal80/90C03-Gal80]; 2555 axons of 22 nerves). The number of axons in the indicated size intervals is plotted for control (green) and htl^DN nerves (blue). b Same dataset as in (a). Relative changes in the number of axons in different axon size classes upon expression of htl^DN compared to control animals. Note the increase in the number of small diameter axons and the reduction of larger caliber axons. c The wrapping index of control larvae [white¹¹¹⁸] and larvae expressing htl^DN in the wrapping glia [nrv2-Gal4; UAS-htl^DN]. The wrapping index in control nerves is 18.5 (n = 27 nerves). The wrapping index is 7.5 upon heartless suppression (n = 22 nerves, p = 3.5E−11, t test). d The accumulated distance per minute is not changed by expression of htl^DN (genotypes as indicated). n = 68 control larvae and n = 71 larvae expressing htl^DN, p = 0.85, Wilcoxon rank-sum test. Box plots in (c, d) show median (horizontal line), boxes represent the first and third quartile, whiskers show standard deviation, individual points show outliers. e Bending angle distribution of control larvae and larvae expressing htl^DN in wrapping glia (genotypes as in (a)). A slight difference in the bending behavior can be detected (p = 8.4E–32, t test). f Quantification of the coiling phenotype. The relative distribution of 300 frames long movement clips (n = 263 30-s long video clips with 300 frames each) with 0, 1–30, or 31–60 frames showing coiling is shown. Green denotes control, blue denotes larvae expressing htl^DN in wrapping glia, genotypes as in (a).

Fig. 4. Wrapping glial cells require innexins for their growth.

Third instar larval filet preparations expressing endogenously V5-tagged Ogre (a green), or endogenously V5-tagged Innexin2 (b–d green), or carrying a nrv2-GFP fusion (e–g, green). HRP (blue) labels neuronal membranes, (a, e–g) Innexin2 antibody staining is shown in red. More than ten animals were analyzed for each genotype. Representative images are shown. Scale bars are 200 µm. a In a control larva, Innexin1 (Ogre, green) and Innexin2 (Inx2, red) show overlapping localization. The boxed area is shown in higher magnification. b Upon panglial, RNAi mediated suppression of innexin2 expression no Inx2 can be detected in nerves. c Upon suppression of innexin2 expression only in the subperineurial glial cells some reduction in Inx2 can be observed. d Similar expression patterns are observed upon innexin2 suppression in wrapping glia. e Note that wrapping glial cells cover the entire width of the nerve. f Upon knockdown of ogre in wrapping glia, the differentiation of wrapping glia is affected (arrowheads). Mainly thin wrapping glial processes can be detected. Inx2 is seen in the peripheral nerves (arrows). g A similar phenotype is noted when innexin2 is suppressed in the wrapping glia (arrowheads). h Bending distribution of control [nrv2-Gal4; 90C03-Gal80/UAS-GFP^dsRNA] and ogre knockdown third instar larvae [nrv2-Gal4; 90C03-Gal80/UAS-ogre^dsRNA]. ogre knockdown larvae show increased bending (p = 1.36E−25, t test). i The accumulated distance per minute is slightly reduced in ogre knockdown animals (p = 1.6E−19, Wilcoxon rank-sum test, n = 88 control larvae and n = 69 larvae expressing ogre^dsRNA). Box plot shows median (red line), boxes represent the first and third quartile, whiskers show standard deviation, red points show outliers. j Same dataset as in (i). Quantification of the coiling phenotype. The relative distribution of 300 frames long movement clips (n = 259 30 s long video clips with 300 frames each) with 0, 1–30, or 31–60 frames showing coiling. Green denotes control, red denotes larvae expressing ogre^dsRNA in wrapping glia. Genotypes as in (h).

Fig. 5. Ablation of wrapping glia in abdominal nerves.

a Confocal projection of a control third instar larval filet preparation [nrv2-Gal4 / UAS-GFP; 90C03-Gal80, UAS-CD8Cherry/90C03-Gal80]. Expression of CD8Cherry is observed in wrapping glia along abdominal nerves, in a structure close to the mouth hooks (arrowhead) and in few glial cells in the brain (see c). Representative images are taken from >10 animals analyzed for each genotype. b Wrapping glia ablated third instar larva [nrv2-Gal4/UAS-hid; 90C03-Gal80, UAS-CD8Cherry/90C03-Gal80]. Note the absence of CD8Cherry expression (see d). Scale bars are 500 µm. c Some brain glial cells express CD8Cherry in control larvae. d No CD8Cherry expression can be detected following hid expression. e, f Futsch expression as indicated, genotypes as in (a, b). Note the elaborated branching morphology of peripheral nerves. Upon wrapping glial ablation (f) Futsch expression remains unchanged compared to (e). g, h Electron microscopic images of segmental nerves at about 160 µm distance to the ventral nerve cord of wandering third instar larvae. The specimens were fixed and embedded in a filleted form. The size of the different axons was calculated using Fiji (yellow lines). Neural lamella (nl), perineurial glia (pg), subperineurial glia (spg), wrapping glia (wg), axon (ax). Scale bar is 2 µm. Genotypes as in (a, b). g Cross-section of a control nerve. h Upon ablation of the wrapping glial cells, only possible remnants of the wrapping glia can be recognized (asterisk). Note that large caliber axons are usually found close to the glial cells of the blood-brain barrier (arrowheads). i The number of axons in the indicated size intervals is plotted for control (green) and wrapping glia ablated (black) nerves (n = 1901 control axons and 1915 axons in glia ablated nerves, genotypes as in (a, b). j Relative changes in the number of axons in different axon size classes upon ablation of the wrapping glia compared to control animals. Note the increase in the number of small diameter axons and the reduction of larger caliber axons.

Fig. 6. Larval locomotion impairment upon wrapping glial ablation.

a, b Representative FIM images. a Tracks of 15 control third instar larvae [nrv2-Gal4/UAS-GFP; 90C03-Gal80, UAS-CD8Cherry/90C03-Gal80]. b Third instar larvae with ablated wrapping glial cells [nrv2-Gal4/UAS-hid; 90C03-Gal80, UAS-CD8Cherry/90C03-Gal80]. Larvae were imaged for 3 min at 10 frames/s. Upon ablation of wrapping glia larvae cover only a small part of the 20 cm × 20 cm arena and perform many turns. c Peristaltic body contractions over time. upper panel: control larva, lower panel: wrapping glia ablated larva, genotypes as in (a, b). Note the reduction in body size, and reduced and irregular peristaltic waves upon ablation of wrapping glia. d The peristalsis frequency is reduced upon wrapping glia ablation (p = 9.8E−17). e–i Box plots showing quantifications of the larval locomotion parameters. n = 91 control larvae, and n = 79 wrapping glia ablated larvae. All box plots show median (red line), boxes represent the first and third quartile, whiskers show standard deviation, red points show outliers. e Accumulated distance [cm/min], p = 9.1E−7. f Distance to origin [cm/min], p = 1.5E−25. g Number of turns/min, p = 2.3E−4. h Number of stops/min, p = 1.6E−35. i Number of head bends/10 s p = 8.3E−21. d–g Statistical analyses: Wilcoxon rank-sum test.

Fig. 7. Coiling behavior upon wrapping glia ablation.

a Graph showing all bending angles of control [nrv2-Gal4/UAS-GFP; 90C03-Gal80, UAS-CD8Cherry/90C03-Gal80] and wrapping glia ablated larvae [nrv2-Gal4/UAS-hid; 90C03-Gal80, UAS-CD8Cherry/90C03-Gal80] (p = 1.8E−244, t test). A bending angle of 180° corresponds to a non-bended larva. b Box plot shows bending strength [°] during stop phases. Control animals show a median bending angle of around 3°, while Hid expressing animals show a median bending angle of around 47° (p = 6.8E−37, Wilcoxon rank-sum test). Genotypes as in (a). n = 91 control larvae, and n = 79 wrapping glia ablated larvae. Box plot shows median (red line), boxes represent the first and third quartile, whiskers show standard deviation, red points show outliers. c Five frames of a movie showing coiling of a larva with ablated wrapping glia (genotype as in a). Representative example out of 150 coiling events analyzed. Scale bar is 1 mm. d Relative changes in coiling behavior. The following genotypes were analyzed: Control (green), wrapping glia ablation (black), genotypes as in (a), and larvae expressing htl^DN in wrapping glia (blue) [nrv2-Gal4, UAS-htl^DN; 90C03-Gal80/90C03-Gal80]. The proportion of 30 s long video clips (300 frames) showing no coiling, 1–30 frames, 30–60 frames, or more than 60 frames with coiling are plotted. Filled bars indicate animals that were fed with sphingosine. In control larvae coiling is observed in 0.097% of all frames (n = 657 video clips of 300 frames each). Upon feeding with sphingosine, coiling frequency increases in control animals to 0.356%, (n = 203, p = 0,0474). Upon wrapping glia ablation, coiling is noted in 1.507% of all frames (n = 557). Upon feeding with sphingosine coiling frequency does not change significantly to 2.166%, (n = 269, p = 0,2653). Larvae expressing htl^DN in wrapping glia show coiling in 0.774% of all frames (n = 263). Upon feeding with sphingosine, coiling frequency decreases significantly to 0.106% (n = 223, p = 0.0001) which is not distinguishable from the coiling frequency of control animals (0.097%, p = 0.4209). Statistical analyses: Wilcoxon rank-sum test.

Fig. 8. Ablation of wrapping glial cells reduces action potential conduction speed.

a–c Conduction speed in motor axons. a Schematic representation of the experimental setup. One electrode (extra1) was placed at the seventh or eighth abdominal nerve close to the ventral nerve cord. The membrane potentials exiting the CNS during fictive crawling were recorded by a second electrode (extra2) placed anterior to the affected segment. The time interval (∆t) between identifiable depolarization phases was put in relation to the distance between the two electrodes (∆d) to calculate the conduction speed. b Exemplary recording traces of the motor unit. Genotypes are as in Fig. 7. c The mean conduction speed in control motor axons is 0.196 m/s (n = 16 larvae). Upon wrapping glia ablation, the mean conduction speed is 0.133 m/s (n = 14 larvae, p = 0.0000125). A similar conduction velocity (0.131 m/s, n = 10 larvae, p = 0.00000374) is observed in motor axons of larvae expressing htl^DN in the wrapping glia. d–f Conduction speed in sensory axons. d To determine the conduction speed in sensory axons, membrane potentials were recorded following mechanical stimulation of the innervated segment (extra1, extra2). The conduction speed was calculated by determining the temporal delay of identifiable sensory spikes between the recording sites and the distance between the recording sites. e Exemplary recording traces of the sensory unit. Genotypes are as above. f Sensory axons show a 34% reduced conduction speed compared to motor axons (c). The mean conduction speed in control sensory axons is 0.129 m/s (n = 16 larvae). Upon ablation of the wrapping glia, the mean conduction speed is 0.079 m/s (n = 14 larvae, p = 0.00000587) corresponding to decrease in conduction speed by 38.5%. In sensory axons of larvae expressing htl^DN in the wrapping glia a slightly reduced conduction velocity of 0.110 m/s (n = 10 larvae, p = 0.0229) is observed. Both plots (c, f) show the mean and standard deviation. Statistical analyses: Wilcoxon rank-sum test and t test.

Fig. 9. Schematic representation of the Goro circuit.

a Control situation. The rolling response is triggered when a noxious signal is perceived by the mdIV neurons and transmitted via the abdominal nerves to the ventral nerve cord. Here, the mdIV neuron synapses onto the Basin-4 neurons (1) which project to the A05q neurons (2) which innervate the Goro neurons (3) that act as the command neurons orchestrating the activity of motor neurons (4). Additional pathways exist which involve neurons located in the brain hemispheres. b Upon ablation of wrapping glial cells, mdIV neuron activation results in ephaptic crosstalk and uncontrolled activation of motor axons. Likewise, upon ablation of wrapping glial cells, Goro neuron activation results in ephaptic crosstalk (flash) and activation of a sensory input affecting the rolling response.

Fig. 10. Modulation of the Goro circuit.

a Upon optogenetic activation of the mdIV neurons in third instar control larvae [ppk-LexA, nrv2-Gal4/UAS-GFP; LexAop-csChrimson, nrv2-Gal4/90C03-Gal80] rolling is induced after 210 ms (n = 38 larvae). Upon ablation of the wrapping glia in third instar larvae [ppk-LexA, nrv2-Gal4/UAS-hid; LexAop-csChrimson, nrv2-Gal4/90C03-Gal80] no rolling response is induced but larvae cramp after 80 ms (n = 28 larvae; p = 7E−10). Upon expression of htl^DN [ppk-LexA, nrv2-Gal4/UAS-htl^DN; LexAop-csChrimson, nrv2-Gal4/90C03-Gal80] a rolling response is initiated as in control but with a delay of 120 ms (n = 32 larvae, p = 1.9E−5). b In control larvae optogenetic activation of the Goro neurons [Goro-LexA/UAS-GFP; LexAop-csChrimson, nrv2-Gal4/90C03-Gal80] induces rolling after 475 ms (n = 27 larvae). Upon ablation of wrapping glia [Goro-LexA/UAS-hid; LexAop-csChrimson, nrv2-Gal4/90C03-Gal80] the rolling response is induced at about the same time (n = 43 larvae, p = 0.212). A slightly significant delay in the rolling response can be observed when comparing to larvae expressing htl^DN [Goro-LexA, nrv2-Gal4/UAS-htl^DN; LexAop-csChrimson, nrv2-Gal4/90C03-Gal80] (n = 26 larvae, p = 0.013). c Control larvae show 5 rolls (n = 23 larvae, median value). Upon ablation of wrapping glia 9 rolls are induced (median value, n = 33 larvae, p = 0.000074). Larvae expressing htl^DN in wrapping glia show 6 rolls which is not significantly different from the control (n = 20 larvae, p = 0.48). Genotypes as in (b). d The orientation of rolling upon Goro activation changes when wrapping glia are ablated but not when wrapping glia differentiation is impaired. Control larvae (n = 27) and larvae expressing htl^DN (n = 20) do not change orientation of rolling reaction (p = 0.43). In contrast upon wrapping glia ablation larvae change orientation during every other roll (n = 39 larvae, p = 0.005). Genotypes as in (b). Box plots: median (horizontal line), boxes represent first and third quartile, whiskers show standard deviation, individual points show outliers. All statistical analyses: t test.