Hemocytes are essential for Drosophila melanogaster post-embryonic development, independent of control of the microbiota

ABSTRACT Proven roles for hemocytes (blood cells) have expanded beyond the control of infections in Drosophila. Despite this, the crucial role of hemocytes in post-embryonic development has long thought to be limited to control of microorganisms during metamorphosis. This has previously been shown by rescue of adult development in hemocyte-ablation models under germ-free conditions. Here, we show that hemocytes have an essential role in post-embryonic development beyond their ability to control the microbiota. Using a newly generated strong hemocyte-specific driver line for the GAL4/UAS system, we show that specific ablation of hemocytes is early pupal lethal, even under axenic conditions. Genetic rescue experiments prove that this is a hemocyte-specific phenomenon. RNA-seq data suggests that dysregulation of the midgut is a prominent consequence of hemocyte ablation in larval stages, resulting in reduced gut lengths. Dissection suggests that multiple processes may be affected during metamorphosis. We believe this previously unreported role for hemocytes during metamorphosis is a major finding for the field.

Survival was scored as pupae obtained from inoculated first instar larvae for Hml ∆ >rpr or Hml P2A >rpr (attP2). For Hml ∆ >Bax or Hml P2A >Bax (attP2) survival was scored relative to the internal controls Hml ∆ >Bal or Hml P2A >Bal as fraction of GFP negative per GFP positive pupae (see materials and methods section for details). Each dot represents an individual vial.
One-way ANOVA were performed.
(B) Eclosion rates from Hml ∆ >Bax or Hml P2A >Bax (attP2) animals reared at controlled density under conventional conditions, on food containing 5 mg/mL Ampicillin and 5 mg/mL Kanamycin or under germ-free conditions. Eclosion rates were scored relative to the internal controls Hml ∆ >Bal or Hml P2A >Bal as fraction of Cy+ per Cy-adults, which can produce fractions of expected >1 (see materials and methods section for details). Each dot represents an individual vial. One-way ANOVA were performed.
(C) Comparison of Hml P2A -GAL4 driver inserted in attP40 or attP2 based on controlled density cultures raised under conventional conditions. Larval survival was scored as pupae obtained from inoculated first instar larvae. Eclosion rates were scored as number adults obtained from pupae that were formed and pupariation timing as average over the day of pupariation for each pupae in one vial. Pupal lethality was scored by determining the fraction of all pupae in a vial that terminated development before pupal stage P8 (light pupae), during stage P8-P14 (pharates) or in P15 (failed eclosion). Each dot represents an independent experiment including 4 vials each. One-way ANOVA were performed.
(E) Pupal phenotypes of Hml P2A >rpr (attP2) animals. In early pupae (12h after pupariation formation) defective retraction of mouth hooks (white arrowhead) and lack of clearly visible anterior spiracles (yellow arrowhead) is evident. By 24 h after pupariation formation 95% of pupae show a large posterior air bubble (red arrowhead).
(F) Larval survival from controlled density cultures under conventional conditions. Survival was scored as pupae obtained from inoculated first instar larvae. Genotypes used here are identical with those in Figure 3F. One-way ANOVA were performed.
Data are mean±s.d.. Controls: attP2>rpr and attP2>Bax (A), attP2>Bax (B), yw>rpr (C & E), attP2>rpr (F). * p<0.03; ** p<0.002; *** p<0.0002; **** p<0.0001. ns, not significant. Development: doi:10.1242/dev.200286 Figure 4D. Dissection was carried out in Schneider's medium containing Ca 2+ (A) or in calcium-free PBS (B), which reduced the contraction of guts after dissection. In both conditions, Hml P2A >rpr (attP2) animals showed reduced midgut length compared to controls. This effect was observed despite the sexual dimorphism of midgut length in males and females and under conventional as well as germfree conditions. Number of guts are indicated. One-way ANOVA were performed. Box plot shows median values (middle bars) and first to third interquartile ranges (boxes); whiskers indicate minimum and maximum values. (C). Fluorescent micrographs of midgut regions close to the midgut/hindgut border from Hml P2A >rpr (attP2) and control WS-larvae reared at controlled density under conventional conditions and dissected in PBS. 3D reconstruction of DNA and Phalloidine staining consistently showed a slight deterioration of visceral musculature. In optical cross sections we observed no difference in staining the musculature on the basal side (white asterisk) and the brush border on the apical side of gut cells (yellow asterisk). Maximum projection of stacks visualized in 3D showed that enterocytes (large nuclei) and smaller cells that might include intestinal stem cells, enteroblasts and enterendocrine cells were present with no apparent differences.
(D) Electron micrographs of midgut regions close to the midgut/hindgut border from Hml P2A >rpr (attP2) and control WS-larvae reared at controlled density under germ-free conditions. Basement membrane (yellow arrowheads), peritrophic membrane (red arrowheads) and microvilli (black arrowheads) are comparable between the genotypes.