Macrophages Use Distinct Actin Regulators to Switch Engulfment Strategies and Ensure Phagocytic Plasticity In Vivo

Abstract

Macrophages must not only be responsive to an array of different stimuli, such as infection and cellular damage, but also perform phagocytosis within the diverse and complex tissue environments found in vivo. This requires a high degree of morphological and therefore cytoskeletal plasticity. Here, we use the exceptional genetics and in vivo imaging of Drosophila embryos to study macrophage phagocytic versatility during apoptotic corpse clearance. We find that macrophage phagocytosis is highly robust, arising from their possession of two distinct modes of engulfment that utilize exclusive suites of actin-regulatory proteins. "Lamellipodial phagocytosis" is Arp2/3-complex-dependent and allows cells to migrate toward and envelop apoptotic corpses. Alternatively, Diaphanous and Ena drive filopodial phagocytosis to reach out and draw in debris. Macrophages switch to "filopodial phagocytosis" to overcome spatial constraint, providing the robust plasticity necessary to ensure that whatever obstacle they encounter in vivo, they fulfil their critical clearance function.

Keywords: Drosophila; Filopod; Lamellipod; Phagocytosis; actin; cytoskeleton; efferocytosis; engulfment; hemocyte; macrophage.

Graphical abstract

Figure 1

Macrophage Phagocytosis Is Extremely Robust (A) Top: schematic of the developmental migration of macrophages (M$\text{Φ}$, green) on the ventral side of the Drosophila embryo between the CNS (gray) and the overlying epithelium (embryonic stage 12). During their dispersal, macrophages clear corpses, which can be visualized through the injection of fluorescent Annexin V (red). Bottom: in vivo imaging of dispersing macrophages (stage 12) expressing LifeAct-GFP (F-actin, green) engulfing Annexin-V-labeled debris (red). The dashed box is magnified in subsequent panels. Arrow highlights engulfed corpse. (B) GFP-tagged Arp3 (subunit of Arp2/3 complex) and Dia and Ena (green) co-localize with LifeAct-mCherry (F-actin, red) at phagocytic cups (in contrast to microtubles [CLIP170]). Arrows highlight actin-rich phagocytic cups (all embryonic stage 12). (C) Control and scar and arp3 mutant macrophages expressing LifeAct-GFP (F-actin, green) in stage 15 embryos globally expressing mCherry-moesin (F-actin, red). mCherry-labeled corpses (arrows) accumulate in macrophages via engulfment. (D) Quantification of mCherry-positive corpses/macrophages (control = 6.11 ± 0.24, dia = 5.13 ± 0.18, ena = 5.66 ± 0.20, scar = 11.61 ± 0.52, and arp3 = 12.51 ± 0.44 [mean corpse/cell ± SEM]; n = 100 cells [>10 stage 15 embryos]/genotype). Error bars represent 95% confidence intervals [CIs], and asterisks indicate statistical significance versus control mean (ANOVA, p < 0.05). ns, p > 0.05. All scale bars represent 10 μm. s.12/15 denotes embryonic stage.

Figure 2

Loss of Arp2/3 Complex Activity Promotes Filopodial Phagocytosis (A) Left: UV irradiation (blue) of individual epithelial cell (green) induces cell death. The extruded corpse can be visualized through injection of fluorescent Annexin V (red), which is then engulfed by macrophages. Right: series of stills before and after UV irradiation of an individual epithelial cell (^∗), delineated through global expression of GFP-moesin (F-actin, green). The dying cell is rapidly labeled with Annexin V (red) (embryonic stage 15). Time is in minutes. (B) Macrophages expressing LifeAct-GFP (green) engulfing UV-induced corpses (^∗). All but scar and arp3 mutant macrophages are capable of enveloping dying cells with their lamellipods. Instead, scar and arp3 mutant macrophages use filopods to draw small pieces of material (^∗) back into the cell body (double-headed arrows). FP4mito (FP4) inhibits Ena. All embryonic stage 15. Time is in minutes. (C) Percentage of phagocytic events classified as lamellipodial, filopodial, or ambiguous in different genotypes in response to UV-induced cell death (control = 5.0 ± 5.0, scar = 51 ± 8.72, and arp3 = 60 ± 18.71 [mean percentage of filopodial phagocytosis ± SEM; five stage 15 embryos/genotype]). Asterisks indicate significantly increased filopodial phagocytosis compared to control (ANOVA, p < 0.05). (D) Peak phagocytic cup size (μm²) of different genotypes during engulfment of UV-induced apoptotic corpses. y axis represents percentage of phagocytic cups that reached a peak size within a delineated range of areas (n ≥ 15 phagocytic events [five stage 15 embryos]/genotype). FP4mito (FP4) inhibits Ena. Asterisks indicate significantly reduced peak phagocytic cup size compared to control (ANOVA, p < 0.05). (E) Mean cell speed (μm/min) for each hour of development from embryonic stage 12 for indicated genotypes. Both mutants are significantly slower than the control (^∗) at all time points (ANOVA, p < 0.05, three embryos/genotype). Error bars represent SEM. (F) Percentage of phagocytic events classified as filopodial for each hour of development from embryonic stage 12 for indicated genotypes. Both mutants exhibit significantly higher frequencies of filopodial phagocytosis than the control (^∗) at all time points (ANOVA, p < 0.05, three embryos/genotype). Error bars represent SEM. All scale bars represent 10 μm. s.12-15 denotes embryonic stage.

Figure 3

Dia and Ena Underlie Arp2/3-Complex-Independent, Filopodial Phagocytosis (A) GFP-tagged Dia or Ena (green) co-localize with LifeAct-mCherry (F-actin, red) at filopods and phagocytic cups (arrows) of scar mutant macrophages. Microtubules (CLIP170) are present in filopods, but not phagocytic cups. Scale bar represents 10 μm. s.15 denotes embryonic stage 15. (B) Quantification of mCherry-positive corpses/macrophage (control = 6.11 ± 0.24, FP4mito = 6.24 ± 0.25, ena; FP4mito = 5.9 +0.22, dia; FP4mito = 5.85 ± 0.24, and scar = 11.61 ± 0.52, n = 100 cells [>10 stage 15 embryos]/genotype]; dia, scar = 9.75 ± 0.51, scar; FP4mito = 9.18 ± 0.66, dia, scar; and FP4mito = 8.07 ± 0.47, n ≥ 50 cells [≥10 stage 15 embryos]/genotype). Error bars represent 95% CIs, and asterisks indicate statistical significance versus control or scar mutant mean (ANOVA, p < 0.05). ns, p > 0.05. FP4mito (FP4) inhibits Ena activity. (C) Percentage of phagocytic events classified as filopodial for each hour of development from embryonic stage 12 for indicated genotypes. scar mutant macrophages exhibit significantly higher frequencies of filopodial phagocytosis than the control at all time points (^∗). ns indicates that none of the other genotypes significantly differ from the control at any time point with the exception of dia; FP4mito, which has significantly reduced filopodial phagocytosis at 1 h (boxed asterisk) (ANOVA, p < 0.05; three embryos/genotype). Error bars represent SEM. (D) Engulfment rate per macrophage for each hour of development after embryonic stage 12 for indicated genotypes. scar mutants have significantly higher engulfment at indicated (^∗) time points compared to the control. ns indicates that none of the other genotypes significantly differ from the control at any time point. (ANOVA, p < 0.05; three embryos/genotype). Error bars represent SEM. (E) Schematic illustrating roles of actin regulators in lamellipodial and filopodial phagocytosis.

Figure 4

Wild-Type Macrophages Utilize Filopodial Phagocytosis to Overcome Spatial Restriction (A) Control macrophages expressing LifeAct-GFP (F-actin, green) in stage 12 embryos globally expressing mCherry-moesin (F-actin, red). Control macrophages utilize lamellipodial phagocytosis (top) or filopodial phagocytosis (bottom) to engulf mCherry-positive debris. (B) Percentage of phagocytic events classified as lamellipodial, filopodial, or ambiguous. Filopodial phagocytosis is significantly more common in dispersing (stage 12) rather than dispersed (stage 15), macrophages (20.41% ± 3.39% versus 2.50% ± 2.50%; mean ± SEM, five stage 15 embryos/genotype). Asterisk indicates statistical significance (ANOVA, p < 0.05). (C) Control (top) and scar (bottom) stage 15 macrophages expressing LifeAct-GFP (F-actin, green) engulfing E. coli (pHrodo, red) through lamellipodial or filopodial phagocytosis (arrows), respectively. Boxed region enlarged in intervening panels. Time is in seconds; scale bars represent 10/1 μm. (D) Laser ablation (^∗) of the embryo (stage 15) leads to the inflammatory recruitment of macrophages expressing LifeAct-GFP (F-actin, green) to the wound edge (dashed line), prompting lamellipodial and filopodial phagocytosis of necrotic debris (arrows). Dotted boxes are magnified in in adjacent panels. (E–G) Injection of dextran (red) and Annexin V (white) into interstitial space where macrophages (LifeAct-GFP [F-actin], green) reside demonstrates relationship between mode of engulfment and spatial constriction. (E) Dispersing macrophages (stage 12) are surrounded by less dextran-filled space compared to dispersed macrophages (stage 15). Note limited Annexin-V-labeled corpses at stage 15 due to their prior clearance. (F) Dashed box in (E). Lamellipodial phagocytosis of corpse (Arrow) by macrophage in a region of high dextran/low spatial constraint (embryonic stage 12). Dashed line denotes the cell outline. (G) Filopodial phagocytosis of a corpse (arrow) by a macrophage in a region of low dextran/high spatial constraint (embryonic stage 12). Dashed line denotes the cell outline. (H and I) Normalized (background-subtracted) fluorescence intensity plots across (H) lamellipodial phagocytosis (white line in Figure 4F) or (I) filopodial phagocytosis (white line in Figure 4G). Color-coded as in (F) and (G). A.U., arbitrary units. (J) Significantly less dextran (mean fluorescence intensity per μm², normalized to total dextran signal) surrounds filopodial compared to lamellipodial phagocytic events. A.U., arbitrary units. Asterisk indicates statistical significance (t test, p < 0.05, six stage 12 embryos). Unless otherwise indicated, all scale bars represent 10 μm. s.12/15 denotes embryonic stage.