Exposure to Therapeutic BTK Inhibitors Induces Phenocopying of Btk29A Mutants in the Fruit Fly Drosophila melanogaster

Background : Bruton’s tyrosine kinase (BTK) is a non-receptor type tyrosine kinase originally identified as the genetic signature responsible for X-linked agammaglobulinemia (XLA) when mutated. Its functional form is required for B lymphocyte maturation in both humans and mice, whereas loss-of-function causes a different form of developmental defect in the fruit fly, Drosophila melanogaster . Methods : Ibrutinib and other therapeutic inhibitors of BTK have been extensively used to successfully treat various leukemias and lymphomas. Btk29A type 2 is the ortholog of BTK in the fruit fly. We show that feeding wild-type flies an ibrutinib-containing diet induces phenocopying of Btk29A mutants, i.e., failure in the fusion of left and right halves of the dorsal cuticles, partial loss of wing tissues and dysregulation of germ cell production. Results : We have previously reported that Btk29A phosphorylates Drosophila Arm ( β -catenin), and ibrutinib reduces phosphorylation at Tyrosine142 of endogenously expressed β -catenin in Cos7 cells transfected with Btk29A type 2 cDNA. Conclusions : Thus, Drosophila is suitable for screens of novel BTK inhibitor candidates and offers a unique in vivo system in which the mode of action of BTK inhibitors can be examined at the molecular, cellular, and organismal levels.


Introduction
Bruton tyrosine kinase (BTK) was originally identified as a non-receptor tyrosine kinase whose deficiency is responsible for X-linked agammaglobulinemia (XLA) in humans [1][2][3]. Patients with XLA suffer from severe antibody deficiency and recurrent infections due to a differentiation defect causing the absence of mature B lymphocytes [4][5][6]. Conversely, inhibition of BTK has been shown to ameliorate various forms of leukemia and lymphoma [7,8], with the premier example being chronic lymphocytic leukemia (CLL) [9][10][11]. Ibrutinib is the first such inhibitor of BTK to be clinically applied [7,8,10,11]. However, there remain uncertainties as to how it alters signaling mediated by BTK in vivo, ultimately affecting development, physiology, and behavior at the organismal level. In addressing these questions, a genetically tractable in vivo animal model would be most useful.
Drosophila melanogaster (hereinafter Drosophila) is an outstanding model organism due to its amenability to experimental manipulations at the molecular, cellular, and organismal levels. The Btk29A type 2 isoform encoded by the gene Btk29A is the sole ortholog of human BTK in Drosophila: the other isoform, Btk29A type 1, has a unique N-terminus and distinct expression and is thus structurally and functionally separable from Btk29A type 2 [12]. Btk29A type 2 has been shown to play an essential role in dorsal closure in postembryonic development [13], a morphogenetic process to suture the left and right halves of the exoskeleton. Btk29A type 2 is also pivotal for stem cell niche functions, e.g., germ cell proliferation and differentiation, as well as follicle precursor migration in the ovary at the adult stage [13][14][15][16][17][18]. In the ovary, the Btk29A type 2 protein was shown to bind to β-catenin and to phosphorylate specific tyrosine residues of β-catenin to modulate Wnt4 signaling, thereby regulating germ cell development [19,20]. Furthermore, Btk29A type 2 was shown to mediate behavioral habituation in adult flies, which was abrogated by ibrutinib [18]. Thus, Drosophila will offer an ideal platform to evaluate the actions of ibrutinib on the BTK ortholog. Here we show that dietary ibrutinib recapitulates the effects of the Btk29A mutant on the morphogenesis of external structures and germ cell development, likely through its interference with β-catenin-Btk29A interactions.

Flies
Flies were raised on cornmeal-agar-yeast medium at 25°C. The Canton-Special (CS) strain was used to test the drug effects. The Btk29A ficP allele was isolated in our laboratory. The c587-Gal4 line was a generous gift from D. Drummond-Barbosa. The UAS-RNAi and UAS-Dicer2 strains were purchased from the Vienna Drosophila RNAi Center and Bloomington Stock Center, respectively.

Characterization of External Structures
Wings: Wings were removed from thoraces for observation and image acquisition using an Axio microscope (ZEISS, Oberkochen, Germany).
Thoraces: Dorsal thoraces were photographed under an M205FA microscope (Leica, Wetzlar, Germany). The distance between the left and right counterparts of homologous innermost bristles was measured by counting the number of intervening pixels on the computer screen with the aid of Photoshop software. Half of the value of the interbristle distance thus measured was defined as the distance between the bristle and the midline and used for comparisons among tested fly groups. Some of the flies fed an ibrutinib-containing diet during the larval stage showed disrupted dorsal structures. The proportion of flies exhibiting the dorsal open phenotype was calculated to quantify the ibrutinib action on the dorsal structure formation.

Histology
For the antibody staining, ovaries were dissected in Phosphate-Buffered Saline (PBS) and immersed in 4% paraformaldehyde in PBS for 40 minutes. The ovaries were then washed three times in PBS-T (PBS supplemented with 0.1% Triton-X), and the reaction was blocked for 1.5 h in PBS supplemented with 0.1% Triton and 0.5% Bovine Serum Albumin (BSA). Then, the tissue was incubated with a primary antibody for 3 h at room temperature or at 4°C overnight. The primary antibodies used in this study were mouse monoclonal 1B1 (1:10; Developmental Studies Hybridoma Bank) and rat anti-Vasa (1:10; Developmental Studies Hybridoma Bank). The fluorescence-conjugated secondary antibodies were from Molecular Probes and were used at a 1:200 dilution. All samples were mounted in 80% glycerol.

Transfections and Western Blotting
Cos7 cells (American Type Culture Collection (ATCC), Washington, DC, USA) were seeded overnight at 50% confluence in 6-well plates. Mycoplasma testing has been done for the cell lines used. Transfections were carried out by using polyethylenimine (PEI) (Polyscience, Warrington, PA, USA). Cells at 48 h post-transfection were harvested, washed twice with PBS in Eppendorf tubes, and re-suspended in 1 mL PBS. Cells were collected by low-speed centrifugation of 1250 rpm (500 g) at 4°C.
The collected cells were re-suspended in 100 µL of lysis-Radioimmunoprecipitation assay (RIPA) buffer containing cocktails of protease inhibitor and phosphatase inhibitors, and the sample was then boiled at 65°C for 2-3 min, vortexed, and centrifuged. The protein was subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and probed with the respective antibody. The anti-pY142 (ECM Biosciences, Versailles, KY, USA) antibody was diluted 1:1000, the anti-HA (Sigma-Aldrich, Burlington, MA, USA) antibody was diluted 1:2000, and the anti-β-catenin (BD Biosciences, Franklin Lakes, NJ, USA) antibody was diluted 1:2000.
Pharmacological inhibition was achieved with ibrutinib (PCI-32765; Pharmacyclics, Sunnyvale, CA, USA). The ibrutinib kept at -20°C was dissolved in dimethyl sulfoxide (DMSO) to produce a 10 mM solution, which was added to PBS to attain a final concentration of 10 µM for application onto cells. For each experiment, fresh dilutions were prepared. At 36-48 h post-transfection, cells were starved under serum-free conditions for 4 h, and inhibitors were added during the last 2 h period of starvation.
The band intensity was quantified with the aid of the ImageJ (version 1.53t, NIH, Bethesda, MD, USA) software run on a MacIntosh computer, and obtained data were transferred to Microsoft Excel (version 16, Microsoft, Redmond, WA, USA) and presented as column histograms. The bars represent the means and standard errors of two independent experiments. Data analysis was carried out using statistically available software Graph Prism software (version 6 for MacOS, Boston, MA, USA). Comparisons between groups were made using one-way analysis of variance (ANOVA). p values < 0.05 were considered significant. **p < 0.01, ***p < 0.001, n.s., statistically nonsignificant.

Bruton's Tyrosine Kinase Inhibitor Feeding
Fly food was softened by heating and then kept for 3 min to cool it down to~40°C before mixing with concentrations of BTK inhibitors, followed by cooling again to room temperature. DMSO (the final concentration: 0.38%) as a control was similarly added to food following the same protocol. Parental flies were transferred to food containing DMSO or a BTK inhibitor for 5 days so that their offspring would develop in food containing DMSO or the BTK inhibitor. Adult offspring were transferred 5 days after emergence to analyze their phenotypes. To determine the concentrations of ibrutinib for oral administrations in phenotypic analyses, female flies were allowed to lay eggs on the culture medium (2.6 mL) supplemented with either 200, 100, 75, 50, 25, 20, 15, or 10 µL of the 50 µM drug solution and the offspring were subjected to preliminary analysis: eggs deposited on the medium containing 200 or 100 µL of the drug solution failed in hatching; larvae were unable to complete development in the medium with 75 or 50 µL of the drug solution; adults emerged when grown on the

Results
Visual inspection of the notum of adult flies suggested that feeding ibrutinib throughout the larval stage results in the widening of the distance between left and right homologous bristles along the midline (Fig. 1A-D,G-J), a phenotype also observed in Btk29A ficP mutants (Fig. 1E,F,P). Measurements of the distance between the midline and innermost bristles in the control and test groups of flies demonstrated that feeding ibrutinib significantly widens this distance (Fig. 1P, Supplementary Fig. 1), suggesting a defect in dorsal closure in the ibrutinib-fed flies. Failures in dorsal closure were also manifested as an incomplete fusion of bilateral tergites in Btk29A ficP mutants (Fig. 1K-M), which was phenocopied by feeding an ibrutinib-containing diet to flies of control genotypes (Fig. 1N,O,Q). It remains an open question as to whether flies with a severe thorax phenotype concordantly have a severe dorsal phenotype. In addition, we noted that a substantial proportion of ibrutinibfed flies display distorted wing structures, particularly a loss of certain regions near the wing margin (Fig. 2C cf. Fig. 2A,B). Similar wing phenotypes were observed with an equivalent frequency in Btk29A ficP mutant flies that were not treated with ibrutinib (Fig. 2C,E).
We then examined, in wild-type flies, the possible effects of ibrutinib feeding on oogenesis, during which Btk29A mutations are known to induce overproduction of cystoblasts (see below), which are directly derived from germ stem cells (GSCs); a cystoblast differentiates into cystocytes that transform into an oocyte and nurse cells after four rounds of divisions [21,22]. This process of germ cell proliferation takes place in the anterior tip of an ovariole called the germarium, which contains one to three GSCs, several cystoblasts, and tens of cystocytes in wild-type flies [21]. Feeding ibrutinib caused underproduction of cystoblasts (Fig. 3E,F,K and L cf. Fig. 3A,G), as identified by the presence of spectrosomes (recognized as round structures rich in actin), often leading to a complete lack of cystoblasts in a substantial proportion of germaria (Fig. 3M). This is in sharp contrast to the germaria of Btk29A ficP mutants, where supernumerary cystoblasts were usually observed ( Fig. 3B,H,M). However, when Btk29A was knocked down with Btk29A RNAi in escort cells that support the maintenance of GSCs in the germarium, some ovaries of the manipulated flies carried an empty germarium (Fig. 3D,J,M). We infer that GSCs failed to reproduce themselves as the result of a strong bias toward differentiating into cystoblasts, resulting ultimately in the empty germarium. Conversely, when a GSC was maintained even upon accelerated cystoblast differentiation, the number of cystoblasts would increase. Because the Btk29A ficP mutation removes the type 2 Btk29A isoform, sparing the type 1 isoform, it might be that the type 1 isoform partially compensated for the lack of type 2 so that the GSC was maintained and produced more cystoblasts. In contrast, the UAS-Btk29A RNAi transgene likely knocked down both type 1 and type 2 isoforms, leading to contrasting outcomes, i.e., loss of the GSCs and cystoblasts.
Finally, we examined whether ibrutinib affects molecular interactions between Btk29A and β-catenin. As an experimental system, we chose the primate cell line Cos7 rather than flies because the former express no endogenous Btk, simplifying the interpretation of the assay result, while the latter has two Btk29A isoforms under complex control by other kinases, which complicates the analysis of Btk29Aβ-catenin interactions. Indeed, our previous experiment demonstrated that fly Btk29A phosphorylates endogenous β-catenin at its Y142 when transfected into Cos7 cells [19]. Btk29A carrying an HA tag was transfected into Cos7 cells, and the cell lysates were subjected to immunoblotting with an anti-pY142 β-catenin antibody. The tyrosine residue at position 142 (Y142) is one of the major Btk29A-dependent phosphorylation sites of β-catenin in vivo [19]. The results showed that incubation of cells with 10 µM ibrutinib reduced the amount of a Y142-phosphorylated population of endogenously expressed β-catenin (Fig. 4A), and this reduction was statistically significant (Fig. 4B). We suggest that ibrutinib interferes with Btk29A-β-catenin interactions, thereby manifesting defects in oogenesis and other morphogenetic processes in vivo. It remains to be examined whether the phosphorylation of Armadillo Y150 homologous to β-catenin Y142 is inhibited by ibrutinib in the fly ovary.

Discussion
In the present work, we showed that ibrutinib faithfully reproduced some of the Btk29A mutant phenotypes in Drosophila when orally administered during the larval and adult stages, an observation consistent with the view that BTK and its ortholog are the major targets for ibrutinib in both humans and flies. It is also reminded that ibrutinib does not distinguish between the two Btk29A isoforms because they share the same kinase domain. Our results also point to the possibility that Drosophila, with its genetic tractability and rich collections of molecular resources, may provide an ideal in vivo experimental system for elucidating the mode of action of therapeutic drugs that target BTK. Drosophila will be particularly effective when used to screen chemicals that potentially act on BTK for the development of novel therapeutic drugs to control CLL because a large number of flies can be easily obtained for rounds of assays, and due to their short life cycle, life-long chronic effects of compounds can be evaluated in only a few weeks. Visible phenotypes such as a collapse in the dorsal closure and wing formation and other external structures [23,24] will help to promptly reveal the effects of compounds without dissection of tissues or any other experimental handling, making it possible to enrich promising candidate molecules within a short period of time.
While our phenotypic and biochemical data presented in this and other works suggest functional homology between human BTK and fly Btk29A [25,26], one caveat is that we have no direct evidence that ibrutinib exerts its effects on BTK/Btk29A through the same mechanism in hu-mans and flies. In inhibiting human BTK, irreversible inhibitors, including ibrutinib, bind to the ATP pocket, in which Cystine481 plays an important role in inhibitor-BTK interactions [11,27]. However, fly Btk29A harbors a cystine-to-serine replacement at the corresponding position. Irreversible inhibitors, including ibrutinib, were found to be less effective for patients with CLL carrying the cystine-to-serine substitution in BTK. It remains to be clarified how ibrutinib modulates the Btk29A functions in flies. It might be envisaged that fly Btk29A has a higher order structure, distinct from that of human BTK, which compensates for the effect of cystine-to-serine replacement that reduces the binding of ibrutinib and other irreversible inhibitors.

Conclusions
Feeding wild-type flies an ibrutinib-containing diet induces phenocopying of Btk29A mutants. Thus, Drosophila is suitable for screens of novel BTK inhibitor candidates and offers a unique in vivo system in which the mode of action of BTK inhibitors can be examined at the molecular, cellular, and organismal levels.