Adult Drosophila Lack Hematopoiesis but Rely on a Blood Cell Reservoir at the Respiratory Epithelia to Relay Infection Signals to Surrounding Tissues

Abstract

The use of adult Drosophila melanogaster as a model for hematopoiesis or organismal immunity has been debated. Addressing this question, we identify an extensive reservoir of blood cells (hemocytes) at the respiratory epithelia (tracheal air sacs) of the thorax and head. Lineage tracing and functional analyses demonstrate that the majority of adult hemocytes are phagocytic macrophages (plasmatocytes) from the embryonic lineage that parallels vertebrate tissue macrophages. Surprisingly, we find no sign of adult hemocyte expansion. Instead, hemocytes play a role in relaying an innate immune response to the blood cell reservoir: through Imd signaling and the Jak/Stat pathway ligand Upd3, hemocytes act as sentinels of bacterial infection, inducing expression of the antimicrobial peptide Drosocin in respiratory epithelia and colocalizing fat body domains. Drosocin expression in turn promotes animal survival after infection. Our work identifies a multi-signal relay of organismal humoral immunity, establishing adult Drosophila as model for inter-organ immunity.

Keywords: Drosocin; Drosophila melanogaster; Jak/Stat; NFκB; Stat92E; antimicrobial peptide; blood cell reservoir; fat body; hematopoiesis; hemocyte; hop; imd; innate immunity; macrophage; respiratory epithelia; tracheal air sacs; upd3.

Figure 1.. The respiratory epithelia provide the largest reservoir of hemocytes in adult Drosophila.

(A-C) Cryosections of adult Drosophila, HmlΔ-DsRednls hemocytes red, phalloidin green, respiratory epithelia (air sacs) blue. (A) Longitudinal section, anterior up; (B) cross section of head, dorsal up; (C) cross section of thorax, dorsal up. (D) Adult Drosophila, genotype HmlΔ-DsRed/+; btl-GAL4, UAS-GFP/+ ; hemocytes red, tracheal marker green; respiratory epithelia (air sacs) blue; dissection of head, anterior up; size bar 250μm. (E) Schematics of the the tracheal system and respiratory epithelia of the thorax and head (blue), hemocytes in red. Dashed lines indicate sections and dissected area shown in A-D. Note that model omits heart area, which is not visible in longitudinal section in (A). For full model, see Suppl. Fig. 1.

Figure 2.. Developmental changes of hemocytes.

(A-C) Lateral view of adult Drosophila, (A) Model, hemocytes red, (B,C) HmlΔ-GAL4, UAS-GFP (hemocytes, red pseudo color), (B) day 1 post eclosure, (C) day 6 post eclosure. (D) Larval fat body cells (Oil Red O pseudo-colored green) with associated hemocytes (HmlΔ-GAL4, UAS-GFP pseudo-colored red), dissection of abdomen. (E, F) cross sections of anterior abdomen, HmlΔ-DsRednls (hemocytes red), LipidTox (green), DAPI (blue); dashed box marks heart region; (E) day 1 post eclosure; (F) day 6 post eclosure. (G) External hemocyte quantification of fluorescently labeled hemocytes, time course. (H) Total hemocyte counts per animal, time course.

Figure 3.. Infection-induced changes of hemocytes and accumulation of particles at the respiratory epithelia.

(A-C) Lateral view of adult Drosophila; (A) Model, hemocytes red, (B, C) HmlΔ-GAL4, UAS-GFP (hemocytes red pseudo color), (B) no infection control, (C) Ecc15 injection. (D) External hemocyte quantification, dorsal thorax and anterior abdomen, controls and injected flies; p values of paired 2-tailed t test, *,**,***,or **** corresponding to p≤0.05, 0.01, 0.001, or 0.0001. (E) Total hemocyte counts, control and injected flies. Average and standard deviation; 2-tailed t test shows no statistically significant difference (NS). (F) Total hemocyte counts of flies split in two parts, head and thorax versus abdomen; control and injected flies. Flies were injected into thorax or abdomen at 5 days post eclosion and assayed at 12h and 6d post injection. Average and standard deviation. (G) qPCR expression levels of Hml and Crq from whole flies, −/+ infection, 48h post infection. 8 day old adults injected with E.coli at OD 2. (H-M) Injection of fluorescent microbeads (green pseudo color), HmlΔ-GAL4, UAS-GFP (hemocytes, red pseudo color), respiratory epithelia (air sacs, blue). (H, J, L) Injection in thorax, (H) external view, (J) head dissection, (L) thorax cross section. (I, K, M) Injection in abdomen, (I) external view, (K) head dissection, (M) thorax cross section. (N-Q) Injection of fluorescent E. coli bioparticles (green), HmlΔ-DsRed for head dissections or HmlΔ-DsRednls for cryosections (hemocytes, red), respiratory epithelia (air sacs, blue). (N, P) Injection in thorax, (N) head dissection, (P) thorax cross section. (O, Q) Injection in abdomen, (O) head dissection, (Q) thorax cross section. (R) Injection of pHrodo E. coli bioparticles into HmlΔ-DsRed adults; % of hemocytes positive for pHrodo bioparticles from head, thorax and abdomen 4 h post injection. Mean and standard deviation; one-way ANOVA shows no statistically significant difference (NS). (S) Dissected respiratory epithelia of the head from adults HmlΔ-DsRed (hemocytes, red) injected with pHrodo E. coli bioparticles (green) 4 hours post injection. (T-V) Examples of hemocytes with incorporated pHrodo bioparticles isolated from head, thorax, abdomen, corresponding to (R).

Figure 4.. Contribution of the two hemocyte lineages to the adult blood cell pool.

(A) Timeline of the embryonic and lymph gland lineage, with major sites of hematopoiesis in the larva (hematopoietic pockets and lymph gland). Both lineages persist into the adult. (B-E) flipout-lacZ lineage tracing using srpHemo-GAL4. (B) Experimental genotype tub-GAL80^ts / srpHemo-GAL4, UAS-srcEGFP; UAS-Flp/ act>stop>nuc-lacZ. (C) Timeline of induction, hemocytes of the embryo were labeled in a 6h time window of Flp expression (grey box); blue bars mark time points of samples in (D-E). (D-D’) lacZ/βGal positive hemocytes in the pupa, x-gal staining (blue); (E) lacZ/βGal positive hemocytes in the dissected adult abdomen, x-gal staining (blue); note occasional labeling of larval fat body cells. (F-I) flipout-lacZ lineage tracing, HmlΔ-GAL4. (F) Experimental genotype UAS-Flp; HmlΔ-GAL4, UAS-GFP/ tub-GAL80^ts; act>stop>lacZ/ +. (G) timeline, induction at 0–48h AEL (grey box). Note that at this stage lymph gland hemocytes do not express HmlΔ-GAL4. Blue bars mark time points of samples examined and quantified in (H, I); (H) Thorax cross section of adult fly, genotype as in (F), lacZ/βGal positive hemocytes green (anti-βGal), air sacs and DAPI blue. (I) Fraction of lacZ/βGal positive hemocytes relative to all Crq positive hemocytes in late 3^rd instar larvae, and in the adult; ratio suggest contribution of the embryonic lineage to the adult blood cell pool (dashed line).

Figure 5.. Adult hemocytes do not expand; Srp marks active phagocytes in adult Drosophila.

(A) In vivo EdU incorporation. Percentage of EdU positive cells among hemocytes or control tissue in absence or presence of immune challenges as indicated; average and standard deviation. (B-F’) 2-color Fucci analysis of hemocytes, control (uninfected), sterile injury (PBS), and infection (M. luteus, E. coli); genotype is w¹¹¹⁸; HmlΔFucciOrange^G1; HmlΔFucciGreen^G2/S/M; (B-B’) embryonic-lineage hemocytes released from larvae, note green cells; (C-F’) imaging of Fucci hemocytes in adult flies, dorsal views of thorax and anterior abdomen, anterior up; note absence of green signal of HmlΔFucciGreen^G2/S/M. (G-I) Srp labels phagocytic plasmatocytes in the adult fly. (G) Srp and Hml positive hemocytes, young (3 days) and mature (11 days) adults.. (H) Thorax cross section of adult fly, genotype is HmlΔDsrednls/UAS-CD4-GFP; +/srpD-GAL4; respiratory epithelia (air sacs, blue). (I) Srp-Gal4, UAS-lifeact-GFP positive plasmatocytes (green) with red phagocytic vesicles, released ex vivo from adult fly; DAPI (blue). (J-M) In vivo phagocytosis assay, ex vivo examination of hemocytes. (J) % of hemocytes carrying blue beads, from young (3 days) and mature (11 days) adults. Genotype is HmlΔDsred/ UAS-stinger; +/ srpD-GAL4. (K, L, M) examples of labeled hemocytes as indicated. (N) Model, hemocyte production takes place during the larval stage (mainly in the hematopoietic pockets and the lymph gland), while in the adult hemocyte numbers decline over time.

Figure 6.. Hemocytes and Imd signaling are required for the induction of antimicrobial peptide genes including Drosocin.

(A) Expression of Drosocin in hemocyte-ablated flies and controls. 5 day-old adult Drosophila untreated, injected with sterile PBS, or with E.coli in PBS (OD 6), 9.2 nl; genotypes are HmlΔ-GAL4, UAS-GFP/+ (control) or w; HmlΔ-GAL4, UAS-GFP/ UAS-rpr; UAS-hid/+ (hemocyte ablation); flies harvested at 6h and 24h post injection. Chart displays mean and standard error of the mean (SEM) of samples from a representative biological replicate experiment, using pools of 10 females per condition, and triplicate qPCR runs. Values are displayed relative to the RNA level induced by the sterile PBS injections in control flies. (B-D) Drosocin-GFP expression is restricted to the head and thorax. (B) Dro-GFP uninfected control; (C) Dro-GFP infected; (D) model of Dro-GFP expression (green), hemocytes (red), tracheal system (blue). (E) Location of fat body throughout the animal marked by ppl-GAL4, UAS-GFP (green). (F) Drosocin qPCR of head/thorax versus abdomen tissue. Flies were left untreated, or injected with sterile PBS, or E.coli in PBS (OD 6), 9.2 nl, and harvested at 6 and 24h post infection. Two-way ANOVA with Sidak’s multiple comparison test, *,**,***,or **** corresponding to p≤ 0.05, 0.01, 0.001, or 0.0001, respectively. (G-I) Expression of Drosocin in adult flies upon manipulation of Imd pathway activity. 5 day-old adult Drosophila untreated, injected with sterile PBS, or with E.coli in PBS (OD 6), 9.2 nl; flies harvested at 6h post injection. Charts display mean and confidence interval (CI) of samples from 3 averaged biological replicate experiments, using pools of 10 females per condition, and triplicate qPCR runs for each sample. Values of all charts are displayed relative to the average RNA level induced by the sterile PBS injections in control flies. Two-way ANOVA with Sidak’s multiple comparison test, *,**,***,or **** corresponding to p≤0.05, 0.01, 0.001, or 0.0001, respectively. Transgenes were inducibly expressed in hemocytes 24 hours before injections. (G) Drosocin RNA levels of control (HmlΔ-GAL4,UAS-GFP/+; tub-GAL80^ts/+) versus HmlΔ-GAL4,UAS-GFP/+; tub-GAL80^ts/UAS-imd RNAi; (H) control versus HmlΔ-GAL4,UAS-GFP/+; tub-GAL80^ts/UAS-PGRP-LC RNAi; (I) control versus HmlΔ-GAL4,UAS-GFP/UAS-imd; tub-GAL80^ts/+.

Figure 7.. The Drosocin response is localized to the reservoir of hemocytes at the respiratory epithelia and colocalizing fat body domains, and requires Upd3 signaling from hemocytes.

(A-A”) Dissected heads of genotype Drosocin-GFP/HmlΔ-DsRed (Drosocin-GFP green, hemocytes red), respiratory epithelia (air sacs, blue); (A’) Drosocin-GFP, white; (A”) respiratory epithelia, white. Note Drosocin-GFP expression is high in fat body and moderate in respiratory epithelia (arrowhead). (B-D) Tissue specific RNAi knockdown of Drosocin; overall Drosocin mRNA levels were quantified by qPCR. 6–7 day-old adult females were left untreated, injected with PBS, or E.coli in PBS (OD 6), 9.2 nl, and harvested 6 and 12h post infection. Charts display mean and SEM of samples from a representative biological replicate experiment, using pools of 10 females per condition, and triplicate qPCR runs. Values of all charts are displayed relative to the RNA level induced by the sterile PBS injections in control flies. (B) Drosocin RNAi silencing in hemocytes; (C) in respiratory system; (D) in fat body. (E-F’’’) Anatomy of fat body tissue lining the respiratory epithelia and hemocytes; HmlΔ-DsRednls (hemocytes, red), fat body (LipidTOX, large, distinct green cells), respiratory epithelia (air sacs, blue) (E) Sagital section of adult Drosophila. (E’) Closeup of region indicated in (E). (F) Thorax cross section. (F’-F’’’) Closeup of regions indicated in (F). (G-L) Expression of Drosocin in adult flies upon silencing or overexpression of upd3 and silencing of genes of the Jak/Stat pathway. 5 day-old adult Drosophila untreated, injected with sterile PBS, or E.coli in PBS (OD 6), 9.2 nl; flies harvested at 6h post injection. Charts display mean and CI of samples from 3 averaged biological replicate experiments, using pools of 10 females per condition, and triplicate qPCR runs for each sample. Values of all charts are displayed relative to the average RNA level induced by the sterile PBS injections in control flies. Two-way ANOVA with Sidak’s multiple comparison test, *,**,***,or **** corresponding to p≤0.05, 0.01, 0.001, or 0.0001, respectively. (G, H) Drosocin qPCR of whole flies, inducible transgene expression in hemocytes, (G) Genotypes are control (HmlΔ-GAL4,UAS-GFP/+; tub-GAL80^ts /+) versus HmlΔ-GAL4,UAS-GFP/+; tub-GAL80^ts /UAS-upd3 RNAi. (H) Control versus HmlΔ-GAL4,UAS-GFP/ UAS-upd3; tub-GAL80^ts /+. (I, J) Drosocin qPCR of whole flies, inducible transgene expression in tracheal system. (I) Genotypes are control (btl-GAL4, tub-GAL80^ts, UAS-GFP /+) versus btl-GAL4, tub-GAL80^ts, UAS-GFP / UAS-hop RNAi; (J) Genotypes are control versus btl-GAL4, tub-GAL80^ts, UAS-GFP / UAS-Stat92E RNAi. (K, L) Drosocin qPCR of whole flies, transgene expression in fat body. (K) Genotypes are control (ppl-GAL4, UAS-GFP / +) versus ppl-GAL4, UAS-GFP /+; UAS-hop RNAi/+; (L) Genotypes are control versus ppl-GAL4, UAS-GFP/+; UAS-Stat92E RNAi/+. (M) Model of communication between hemocytes, fat body and respiratory epithelia, in which hemocytes act as sentinels of infection. Gram-negative bacteria that accumulate together with hemocytes in the reservoir between respiratory epithelia and fat body; activation of Imd signaling through PGRP-LC on hemocytes triggers upd3 expression and Upd3 secretion. Upd3 activates Jak/Stat signaling in adjacent domains of the fat body and the respiratory epithelia, contributing directly or indirectly to the induction of Drosocin expression. Since we PGRP-LC/Imd signaling and Upd3/Jak/Stat signaling are required but not sufficient to induce Drosocin expression, additional events (AE) in parallel to these pathways are proposed that would provide sufficiency to trigger Drosocin induction.

Figure 8.. Drosocin silencing in respiratory epithelia and fat body decreases animal survival after infection.

(A-D) Survival assays. Adult female F1 progeny from crosses of GAL4 drivers with the following lines: UAS-Drosocin RNAi lines (D1, D2) or controls (yw (YW), w¹¹¹⁸ (W1118)). Mutant Rel^E20 (R) and spz^rm7 (S) as controls. Figure displays one out of 3 comparable biological replicate experiments; in each experiment, for each genotype and condition 40 to 60 females were assessed; p-values log-rank (Mantel-Cox) test, *,**,***,or **** corresponding to p≤0.05, 0.01, 0.001, or 0.0001, respectively; upper symbol for comparison w1118 vs. Dro RNAi D1; lower symbol for comparison of crosses of yw vs. Dro RNAi D2. 5 day old female flies were treated as follows and then incubated at 29°C. (A-D) Ubi-GAL4 crosses, (A) E. coli (OD6, 9.2 nl); (B) sterile PBS injection (9.2nl); (C) uninjected control; (D) E. cloacae (OD4, 9.2nl). (E-G) HmlΔ-GAL4 crosses, (E) E. coli (OD6, 9.2 nl); (F) sterile PBS injection (9.2nl); (G) uninjected control. (H-J) btl-GAL4 crosses, (H) E. coli (OD6, 9.2 nl); (I) sterile PBS injection (9.2nl); (J) uninjected control. (K-M) ppl-GAL4 crosses, (K) E. coli (OD6, 9.2 nl); (L) sterile PBS injection (9.2nl); (M) uninjected control. Drosocin knockdown with ubiquitous driver, or tracheal system- or fat body driver caused significantly reduced survival after gram-negative infection in all replicates. Drosocin knockdown showed a partially penetrant effect on survival after PBS injection: Tracheal system knockdown: survival significantly reduced for D1 in 0 out of 3 replicate experiments, for D2 in 2 out of 3 replicate experiments. Fat body knockdown: survival significantly reduced for D1 in 1 out of 3 replicate experiments, for D2 in 1 out of 3 replicate experiments. Ubiquitous Drosocin knockdown: survival was significantly reduced for D1 in 2 out of 3 replicate experiments, for D2 in 3 out of 3 replicate experiments. As expected, survival was not affected when Drosocin was silenced in hemocytes (E, F, G).