A dual role for the βPS integrin myospheroid in mediating Drosophila embryonic macrophage migration

Abstract

Throughout embryonic development, macrophages not only act as the first line of defence against infection but also help to sculpt organs and tissues of the embryo by removing dead cells and secreting extracellular matrix components. Key to their function is the ability of embryonic macrophages to migrate and disperse throughout the embryo. Despite these important developmental functions, little is known about the molecular mechanisms underlying embryonic macrophage migration in vivo. Integrins are key regulators of many of the adult macrophage responses, but their role in embryonic macrophages remains poorly characterized. Here, we have used Drosophila macrophages (haemocytes) as a model system to address the role of integrins during embryonic macrophage dispersal in vivo. We show that the main βPS integrin, myospheroid, affects haemocyte migration in two ways; by shaping the three-dimensional environment in which haemocytes migrate and by regulating the migration of haemocytes themselves. Live imaging revealed a requirement for myospheroid within haemocytes to coordinate the microtubule and actin dynamics, and to enable haemocyte developmental dispersal, contact repulsion and inflammatory migration towards wounds.

Keywords: Integrins; Macrophage; Migration.

Fig. 1.

Haemocytes in embryos mutant for myospheroid show defects in developmental dispersal. Lateral view of fixed WT (A–D) and mys maternal and zygotic (mys^M/Z) mutant embryos (E–H). Haemocytes were visualized by expression of the heterologous cell membrane marker CD2 driven by the srph-GAL4 driver and detected with an anti-CD2 antibody. (A,E) The initial phases of haemocyte migration from head mesoderm along the ventral nerve cord (VNC) are normal in mys mutant embryos (E; compare with A). (B,F) Mutant haemocytes fail to migrate along the length of the VNC by stage 13 of development (arrows in F). Mutant haemocytes can also be seen accumulating in the anterior region of the embryo (asterisk) and between the amnioserosa and yolk (arrowhead). (C,G) Haemocyte migration along the dorsal edge of the epidermis is also disrupted (arrows). (D,H) The migration phenotype is also observed at stage 15 (arrows in H).

Fig. 2.

Environmental and haemocyte specific requirements for βPS integrin. (A,B) Orthogonal projections of ventrally orientated stage 15 WT and mys mutant embryos with haemocyte-specific expression of GFP (green), injected with dextran dye (red) to reveal spatial constraints surrounding the haemocytes. Scale bars: 20 µm (ventral view); 5 µm (orthogonal projections). (A) Within injected WT embryos the dye permeates along the length of the embryo, indicating VNC–epithelial separation. (B) Within mys mutant embryos spreading of the dye becomes restricted, indicating incomplete VNC–epithelial separation (white arrowheads indicate restricted area of ventral midline). (i) In some embryos the absence of dye coincides with the distance reached by the lead haemocyte migrating from the anterior of the embryo along the ventral midline, whereas in others (ii) the lead haemocyte fails to reach the area where the dye becomes restricted (distance between lead haemocyte and spatial restriction indicated by line and asterisk), suggesting that spatial constraint is not the sole cause of disruption to haemocyte migration along the ventral midline. (C–E) Lateral view of fixed stage 13 embryos. Haemocyte myospheroid requirements were assessed by coexpressing UAS-CD2 (C) and either UAS-mysRNAi (D) or a dominant-negative version of the mys subunit, UAS-TorsoD/βcyt (E), under the control of the srph-GAL4 driver, and staining with an anti-CD2 antibody. Expression of either UAS-mysRNAi or UAS-TorsoD/βcyt phenocopies the haemocyte migration defects observed in mys mutant embryos. (F–K) Lateral views of fixed and stained stage 13 embryos. (F) mys^M/Z embryo; (G–J) grading of embryos into ‘classes’ based on level of haemocyte migration rescue; and (K) WT embryo. (L) Quantification of haemocyte migration rescue at stage 13 when UAS-mys is expressed in the midline under the control of the sim and slit GAL4 drivers, and/or in the haemocytes under the control of srph-GAL4.

Fig. 3.

myospheroid is important for lateral migration, random migration and inflammatory migration. (A,B) Single projection of live imaging of haemocytes expressing the nuclear marker red stinger under the control of the srp-GAL4 promoter during lateral migration from the ventral midline at stage 13/14. (A) In WT embryos haemocytes migrate from the midline laterally along highly organised, segmental paths. (B) In the mys mutant these stereotyped lateral migrations are almost completely abolished. Scale bars: 25 µm. (C,D) Still images taken from live-cell imaging of haemocytes expressing mCherry–Moesin (to label F-actin) undergoing lateral migration. (C) In WT embryos haemocytes polarise, before rapidly migrating laterally (blue asterisk). (D) In mys integrin mutant embryos haemocytes often polarise but fail to migrate laterally (orange asterisk). Scale bars: 10 µm. (E) Tracking haemocytes undergoing lateral migration reveals that in mys mutant embryos a lower percentage of the haemocytes present on the midline at the start of imaging migrate laterally than in WT embryos (average decrease from 85.2% and 30.9%, P<0.05, n = 5 embryos per genotype). (F) The velocity of the laterally migrating haemocytes was significantly lower in the mutant embryos (average velocity of WT and mys haemocytes were 2.21±0.11 µm/minute and 1.19±0.13 µm/minute, respectively; P<0.0001. Median and interquartile range (IQR) plotted for n = 49 (WT) and n = 18 (mys) haemocytes. (G–I) Tracking of haemocytes expressing red stinger undergoing random migration at stage 15 (G,H) reveals (I) slower random migration velocity of mys mutant haemocytes at this stage (WT and mys haemocytes migrated at 1.8±0.57 µm/minute and 2.7±0.45 µm/minute, respectively). P<0.05; values are means ± s.e.m. for n = 69 (WT) and n = 85 (mys) haemocytes. Scale bars: 25 µm. (J,K). Stills taken from movies of haemocytes in a WT and mys mutant embryo, respectively, migrating to an epithelial wound (asterisk). Scale bars: 10 µm. (L) Monitoring the number of haemocytes at the wound every 5 minutes post wounding over a 60 minute time period indicates a small but significant reduction in the number of mys haemocytes present at early time points following wounding (P<0.05 at 10 and 15 minutes post wounding). (M) Tracking reveals a reduction in the velocity of haemocytes in mys mutant embryos compared with WT when migrating towards a wound (1.8±0.6 µm/minute and 3.0±0.3 µm/minute, respectively). P<0.01; median and IQR plotted for n = 31 (WT) and n = 20 (mys) haemocytes.

Fig. 4.

myospheroid plays a role in contact repulsion. (A,B) Stills taken from live-cell imaging of GFP-expressing haemocytes in WT and mys mutant embryos undergoing random migration at stage 15. (A) In WT embryos contacting haemocytes (blue asterisk) demonstrate contact repulsion, rapidly repolarising and migrating away from one another. (B) In mys mutant embryos haemocytes remain in contact, unable to undergo contact repulsion (orange asterisks). Scale bars: 10 µm. (C) Quantification of the time the lamellipodia of two haemocytes remain in contact. There is a dramatic increase in this time interval in mys mutant embryos (average time in contact for WT and mys haemocytes was 5.6 and 27.1 minutes, respectively), P>0.01, n = 89 (WT) and n = 43 (mys) contact events.

Fig. 5.

myospheroid is important in maintaining haemocyte microtubule dynamics. (A–F) Stills taken from live-cell imaging of haemocytes expressing Clip170–GFP, to label microtubules (MTs), and mCherry–Moesin, to label F-actin, migrating randomly at stage 15. The imaging reveals disruption of MT dynamics in haemocytes lacking functional myospheroid. (A) In WT haemocytes, loss of the MT arm coincides with repolarisation of the actin cytoskeleton. (B) In mys mutant haemocytes the MT arm often collapses within the actin protrusions. (C) Quantification of the number of MT arms formed in a haemocyte per hour reveals no significant difference between WT and mys mutant haemocytes. Values are means ± s.e.m. for n = 30 (WT) and n = 24 (mys) haemocytes. (D) Quantification of the interactions triggering MT arm loss upon cell–cell contact with MT arm alignment or with no MT arm alignment, or MT arm loss independent of cell-cell contact, in WT and mys haemocytes. Values are means ± s.e.m. for n = 30 (WT) and n = 24 (mys) total MT arm loss events. (E,F) Stills taken from rapid live-cell imaging of haemocytes expressing mCherry–Clip and Moesin–GFP under the control of a single copy of the srp-GAL4 driver to label only the MT + ends. (G) Tracking the MT tips revealed that in the absence of myospheroid, the MT protrusion rate was decreased (WT and mys protrusion rates were 0.17 and 0.15 µm/second, respectively, P<0.05. Values are means ± s.e.m. for n = 9 haemocytes for each genotype. (H) The distance to the leading edge reached by the MT tips in mys mutant haemocytes was not significantly different from that in WT (mean distance was 2.1 µm and 1.7 µm, respectively). Values are means ± s.e.m. for n = 9 haemocytes per genotype. Scale bars: 10 µm (A, B), 10 µm (E, F), 10 µm (D).

Fig. 6.

Haemocytes lacking myospheroid show altered actin dynamics. (A,B) Stills taken from live-cell imaging of haemocytes expressing LifeAct under the control of srp-GAL4 in WT and mys mutant embryos. Scale bars: 10 µm. The graphs beneath show the lamellipodial area of five haemocytes measured at 30 second intervals over a 30 minute time period. The large fluctuations in WT and mys mutant haemocytes indicate that the overall lamellipodial dynamics remain unchanged in the absence of integrin βPS. (C) This was confirmed when the average lamellipodial area change per haemocyte is compared with that in WT (n = 5 haemocytes per genotype). (D,E) Other actin-dependent structures within the haemocytes are affected in the mys mutant, with (D) an increase in the number of microspikes compared to WT (average 4.8 and 6.6, respectively, P<0.05) and (E) the number of filopodia per haemocyte (average 9.6 and 14.4, respectively, P<0.05). (F) Quantification reveals no difference in the lamellipodial area of WT and mys mutant haemocytes (n = 47 haemocytes for both genotypes).