Systematic screen of chemotherapeutics in Drosophila stem cell tumors

Abstract

Here we report the development of an in vivo system to study the interaction of stem cells with drugs using a tumor model in the adult Drosophila intestine. Strikingly, we find that some Food and Drug Administration-approved chemotherapeutics that can inhibit the growth of Drosophila tumor stem cells can paradoxically promote the hyperproliferation of their wild-type counterparts. These results reveal an unanticipated side effect on stem cells that may contribute to tumor recurrence. We propose that the same side effect may occur in humans based on our finding that it is driven in Drosophila by the evolutionarily conserved Janus kinase-signal transducers and activators of transcription (JAK-STAT) pathway. An immediate implication of our findings is that supplementing traditional chemotherapeutics with anti-inflammatories may reduce tumor recurrence.

Keywords: Drosophila intestinal stem cell; cancer stem cell; drug screening; whole-animal screening.

Fig. 1.

Characterization of the screenable stem cell tumor model. (A) Diagram of intestinal stem cell (ISC) lineage showing polyploid enterocytes (EC) and diploid enteroendocrine cells (EE). (B) Genotype of the screenable tumor model showing the esg-Gal4 transcription factor driving the expression of GFP, luciferase, and RAF^gof UAS-linked transgenes. (C) WT ISCs and RAF^gof ISC tumors. ISCs are labeled by esg-Gal4 driving UAS-GFP (green). Nuclei are visualized with the DNA dye DAPI (blue). WT ISCs express dpERK (27) (Upper: red cytoplasmic staining) and Delta (28) (Lower: red membrane staining). Expression of UAS-RAF^gof with the esg-Gal4 driver increases dpERK and proliferation of ISC-like Delta-expressing cells. (D) Fate of WT and RAF^gof intestinal fragments injected into WT hosts (n > 100). The esg+ cells within the injected intestines are marked with GFP (green). (E) Measurements of luciferase activity from individually dissected flies shows that gut tumors contribute about 66% of the total luciferase activity in each animal, which correlates with the amount that is absent from colchicine-treated animals (F and G).

Fig. 2.

Screen of 88 FDA-approved oncology drugs identifies two classes of drugs that inhibit Drosophila RAF^gof tumors. (A) Schematic of drug screen. (B) Replicate whole-animal luciferase assays from flies fed either DMSO controls or drugs identified as hits from the screen. Bleomycin did not score as a hit and is included as a negative control. Each bar represents the average of 12 biological replicates; error bar = 1 SD; P < 0.001 by rank-sum analysis. (C) Confocal images of posterior midguts dissected from RAF^gof flies treated with the compounds that scored as hits, with bleomycin included as a negative control. (D) Confocal images of posterior midguts dissected from WT flies fed with drugs that scored as hits from the drug screen. (Upper) Drugs that have no effect and are termed class I drugs. (Lower) Drugs that induce an increase in esg+ cells (green) and are termed class II drugs.

Fig. 3.

Class II drugs increase stem cell proliferation. (A) PH3 staining (red) after treatment with the class II drug bortezomib. PH3 is evident in the nuclei of esg+ cells (green), visible in the surface view, and is absent from the nuclei of polyploid EC cells, visualized 1 µM below subsurface. (B) Box plot showing the number of PH3+ cells/gut: the class I drugs are labeled in blue, the class II in red. Note: vinblastine, vincristine, and paclitaxel arrest the cell cycle in M-phase and are therefore expected to increase PH3+ cells regardless of their effect on proliferation.

Fig. 4.

JAK-STAT mediates the side effect of class II drugs. (A) Confocal images of posterior midguts dissected from WT flies fed with class I and class II drugs. Expression of Upd-3 is detected by Upd-3 Gal4 driving UAS-GFP (false-colored red to distinguish from GFP-labeled esg+ cells in other panels). Upd-3 is induced by class II drugs, not by class I drugs. (B) Reduction of JAK-STAT signaling in ISCs by expression of pathway repressors reduces the proliferation response to the strong class II drug bortezomib.

Fig. 5.

Model of tumor recurrence. The finding that class II drugs have opposite effects on RAF^gof tumors and WT ISCs suggests an additional mechanism for tumor recurrence that includes active signaling from the microenvironment (white arrows) induced by class II drugs.