The Immune Phenotype of Three Drosophila Leukemia Models

Abstract

Many leukemia patients suffer from dysregulation of their immune system, making them more susceptible to infections and leading to general weakening (cachexia). Both adaptive and innate immunity are affected. The fruit fly Drosophila melanogaster has an innate immune system, including cells of the myeloid lineage (hemocytes). To study Drosophila immunity and physiology during leukemia, we established three models by driving expression of a dominant-active version of the Ras oncogene (Ras^V12 ) alone or combined with knockdowns of tumor suppressors in Drosophila hemocytes. Our results show that phagocytosis, hemocyte migration to wound sites, wound sealing, and survival upon bacterial infection of leukemic lines are similar to wild type. We find that in all leukemic models the two major immune pathways (Toll and Imd) are dysregulated. Toll-dependent signaling is activated to comparable extents as after wounding wild-type larvae, leading to a proinflammatory status. In contrast, Imd signaling is suppressed. Finally, we notice that adult tissue formation is blocked and degradation of cell masses during metamorphosis of leukemic lines, which is akin to the state of cancer-dependent cachexia. To further analyze the immune competence of leukemic lines, we used a natural infection model that involves insect-pathogenic nematodes. We identified two leukemic lines that were sensitive to nematode infections. Further characterization demonstrates that despite the absence of behavioral abnormalities at the larval stage, leukemic larvae show reduced locomotion in the presence of nematodes. Taken together, this work establishes new Drosophila models to study the physiological, immunological, and behavioral consequences of various forms of leukemia.

Keywords: Genetics of Immunity; Ras; hemocyte; insect immunity; nematodes; oncogene.

Figure 1

Establishing Drosophila leukemia by expressing Ras^V12 alone and in combination with knockdown of tumor suppressors in hemocytes. (A) Larval hemocytes from control crosses expressing eGFP (w; hml (Δ)-GAL4 UAS-eGFP × w¹¹¹⁸, which is abbreviated to HFP/w). (B–D) The number of hemocytes increases drastically when a dominant active form of Ras (UAS-Ras^V12) alone (HR) or in combination with RNAi lines of tumor suppressors scribble (scrib, HRS) or lethal giant larvae [(l(2)gl, HRL] is expressed using the same driver as in (A). (A–D) show phase contrast exposures while (A’–D’) show the corresponding images under fluorescence, revealing eGFP-expressing hemocytes. (E) Quantification of late third-instar larval hemocytes for the genotypes shown in (A–D). (F) Leukemia lines are viable in the prepupal stage (12 hr of metamorphosis) at 25°. No sign of larval lethality is seen in the leukemic lines. White arrow indicates the normal prepupa. (G) Quantification of eclosion rates at 25° for all lines. Expression of Ras^V12 alone in hemocytes (HR) does not interfere with eclosion rates at 25°; however, reduced eclosion is observed with HRS and HRL. ANOVA followed by Tukey’s multiple comparison test was performed. The same and different letters above columns on the graphs indicate nonsignificance [a to a for instance, see (G)] and significance [a to b for example, see (E) and (G)], respectively. Data represent the means with SD; P < 0.0001 and P < 0.01 was found for (E) and (G), respectively. Bar in (A–D), 50 μm.

Figure 2

Leukemic lines at higher temperatures die and contain hemocyte aggregates at the pupal stage. (A) At 29°, HR failed to eclose and larval/pupal mortality was observed for HRS and HRL. (B) Control HFP/w pupae. Opening of pupal cage at 46 hr after puparium formation (APF) at 29° showed fewer hemocytes and formation of adult structures (outlined in phase contrast). (C–E) Opening of pupae from leukemic lines showed clusters of hemocytes and an absence of adult structures. The dashed line in (C) outlines the vacuole-like structure (perhaps due to drying out). Green arrows point toward GFP-positive hemocytes. Scale bar, 20 μm.

Figure 3

Phagocytosis, hemocyte recruitment to wounds, and wound sealing are not impaired in leukemic lines. (A and B) Ex-vivo phagocytosis of Texas Red–conjugated E. coli (K-12 strain). A representative image of the HFP/w and HRS-leukemia lines is shown after phagocytosis. The red arrow indicates phagocytosed E. coli. (C) Quantification of the phagocytosis index. All leukemic lines show wild-type levels of phagocytic capacity. (D–F) Hemocyte migration to the wound edge in leukemic situation is comparable to wild type. Dashed line in phase-contrast and the GFP channel encircle the wound edges. (E and F) The increased number of hemocytes in the leukemic lines leads to a more blurry appearance in both GFP and merge channels. (G) Wounds in leukemic lines are sealed equally efficiently as in wild type. Bar in (A) and (B) is 10 μm, and in (D–F) is 50 μm. Data represent means of SD; Student’s t-test on the normalized data was performed in (C). ANOVA followed by Tukey’s multiple comparison test was performed in (G). n.s., not significant.

Figure 4

Toll and Imd signaling are regulated in opposite ways in leukemic larvae. Compared to control (HFP/w), expression (A) is found upregulated while Dpt A and Cec A1 (B and C) mRNA levels are downregulated in leukemic larvae compared to untreated larvae (normalized to Rpl32 after RT-qPCR, all levels are shown relative to nontreated controls, which are set to 1). Upon wounding, Drs is also upregulated in control larvae to levels similar to nonwounded leukemic lines [HFP/w in (A)]. However, no further Drs induction is observed after wounding leukemic larvae. In contrast, Dpt A and Cec A1 levels are induced after wounding in all lines. Data represent means of SD; Student’s t-test: ** P < 0.01, * P < 0.05. n.s., not significant.

Figure 5

Survival of septic wounded leukemic larvae is similar to wild type. (A) Aseptic wounding: control wild-type larvae show a slight increased survival compared to leukemic larvae; however, this is not significant. (B–D) Septic wounding with both Gram-negative (E. coli and Erwinia carotovora carotovora) and Gram-positive bacteria (S. aureus). None of the septic wounds caused significant increases in mortality for the leukemic lines compared with controls. Survival was monitored until eclosion. Log-rank test on GraphPad 6.0 was performed to determine statistical significance for the survival curves. (A) P = 0.1182, (B) P = 0.2074, (C) P = 0.3682, and (D) P = 0.3491.

Figure 6

Leukemic lines combined with tumor suppressor knockdowns are more susceptible to nematode infection. (A) Schematic diagrams of the expression pattern of the two different GAL4 drivers used. Bx-GAL4 expression is in the salivary glands and wing discs, whereas hml (Δ)-GAL4 expression is in the larval lymph glands and in both circulating and sessile hemocytes. (B) Upon infection with entomopathogenic Heterorhabditis bacteriophora, the combination of Ras85D^V12 expression and knockdown of tumor suppressors in hemocytes showed increased mortality compared to larvae expressing either construct alone (UAS- Ras85D^V12 and UAS-RNAi of tumor suppressors) or in salivary glands and wing discs. Comparisons were made with respective controls (Bx-GAL4/w and HFP/w). Mortality was normalized to controls which were set to 1. Data represent means of SD; Student’s t-test: ** P < 0.01, * P < 0.05. (C) Knocking down of different RNAi lines using two GAL4 driver lines (Bx-GAL4 and HFP) and expression of Ras85D^V12 alone with HFP only. When compared to respective controls, none of them showed significant mortality upon nematode infection. L, lymph gland; n.s., not significantm, S, salivary gland; W, wing disc.

Figure 7

Larval locomotion is impaired in HRS compared to control (HFP/w) in the presence of nematodes. (A) In the absence of nematodes no differences were observed in the distance covered by leukemic lines and control crosses whereas in the presence of nematodes HRS lines showed reduced mobility. (B) Bending frequencies do not differ in the absence of nematodes whereas in the presence of nematodes HRS larvae show reduced bending compared to controls. Each dot represents the mean value for a replicate and the middle line represents the mean of the replicates. Error bar represents SEM; sample size was at least 115 larvae. Fisher’s LSD test: ** P < 0.01, * P < 0.05.