enhancer of polycomb and the tip60 complex repress hematological tumor initiation … · research...

RESEARCH ARTICLE

Enhancer of Polycomb and the Tip60 complex represshematological tumor initiation by negatively regulating JAK/STATpathway activityAlessandro A. Bailetti1,*, Lenny J. Negrón-Pin ̃eiro

1, Vishal Dhruva1, Sneh Harsh1, Sean Lu1, Aisha Bosula1

and Erika A. Bach1,2,‡

ABSTRACTMyeloproliferative neoplasms (MPNs) are clonal hematopoieticdisorders that cause excessive production of myeloid cells. MostMPN patients have a point mutation in JAK2 (JAK2V617F), whichencodes a dominant-active kinase that constitutively triggers JAK/STAT signaling. InDrosophila, this pathway is simplified, with a singleJAK, Hopscotch (Hop), and a single STAT transcription factor,Stat92E. The hopTumorous-lethal [hopTum] allele encodes a dominant-active kinase that induces sustained Stat92E activation. Like MPNpatients, hopTummutants have significantly more myeloid cells, whichform invasive tumors. Through an unbiased genetic screen, we foundthat heterozygosity for Enhancer of Polycomb [E(Pc)], a componentof the Tip60 lysine acetyltransferase complex (also known as KAT5 inhumans), significantly increased tumor burden in hopTum animals.Hematopoietic depletion of E(Pc) or other Tip60 components in anotherwise wild-type background also induced blood cell tumors. TheE(Pc) tumor phenotype was dependent on JAK/STAT activity, asconcomitant depletion of hop or Stat92E inhibited tumor formation.Stat92E target genes were significantly upregulated in E(Pc)-mutantmyeloid cells, indicating that loss of E(Pc) activates JAK/STATsignaling. Neither the hop nor Stat92E gene was upregulatedupon hematopoietic E(Pc) depletion, suggesting that the regulationof the JAK/STAT pathway by E(Pc) is dependent on substrates otherthan histones. Indeed, E(Pc) depletion significantly increasedexpression of Hop protein in myeloid cells. This study indicates thatE(Pc) works as a tumor suppressor by attenuating Hop proteinexpression and ultimately JAK/STAT signaling. Since loss-of-functionmutations in the human homologs of E(Pc) and Tip60 are frequentlyobserved in cancer, our work could lead to new treatments forMPN patients.

This article has an associated First Person interview with the firstauthor of the paper.

KEY WORDS: JAK/STAT, Myeloproliferative neoplasms, Drosophila,E(Pc), Tip60, Melanotic tumors, Lysine acetyltransferases,Tumor suppressor

INTRODUCTIONMyeloproliferative neoplasms (MPNs) make up a group of clonaldisorders of the myeloid lineage, including polycythemia vera (PV),essential thrombocythemia (ET) and primary myelofibrosis (PMF).A gain-of-function mutation in the Janus tyrosine kinase JAK2(JAK2V617F) is the most prevalent mutation in MPNs and accountsfor ∼95% of PV cases and ∼60% of ET and PMF cases (Jones et al.,2005; Kralovics et al., 2005; Levine et al., 2005; James et al., 2005;Tefferi, 2016). The constitutively active JAK2V617F kinase isligand independent, and in animal models, its oncogenic potentialdepends on its downstream target STAT5, a member of the STATtranscription factor family (Walz et al., 2012; Yan et al., 2012; Sachset al., 2016). Treatments for MPN patients, including phlebotomy,aspirin and JAK2 inhibitors, are temporary and not curative (Tefferi,2016), highlighting the need for new treatments.

JAK2 and STAT5 are components of the conserved JAK/STATsignaling pathway, which regulates multiple developmental andimmunological processes, including hematopoiesis (Levy, 1999;Amoyel et al., 2014). The pathway is triggered when extracellularligands bind to cell-surface receptors, which activate receptor-associated JAKs. These kinases subsequently activate cytoplasmicSTAT dimers through phosphorylation of a highly conservedC-terminal tyrosine residue (O’Shea et al., 2002). PhosphorylatedSTAT dimers translocate to the nucleus, bind to specific sites ingenomic regulatory regions and alter target gene expression. InDrosophila, the JAK/STAT pathway is conserved but simplified,with three IL-6-like cytokines [Unpaired 1 (Upd1), Upd2 and Upd3],one Gp130-like cytokine receptor [Domeless (Dome)], one JAK[Hopscotch (Hop)] and one STAT (Stat92E) (Fig. 1A and Herreraand Bach, 2019). Activation of the Drosophila JAK/STAT pathwayinduces expression of target genes, such as Socs36E, chinmo, zfh1,upd2 and upd3 (Bach et al., 2007; Flaherty et al., 2010; Leathermanand Dinardo, 2008; Bazzi et al., 2018; Yang et al., 2015). Thesimplicity of the Drosophila JAK/STAT pathway represents an idealmodel system to study JAK/STAT signaling in vivo.

Similar to vertebrates, Drosophila hematopoiesis occurs in twotemporally distinct waves, the first during embryogenesis andthe second during larval stages (reviewed in Gold and Brückner,2015; Honti et al., 2014; Letourneau et al., 2016; Banerjee et al.,2019). In the embryo, multipotent hematopoietic progenitorscalled prohemocytes differentiate primarily into plasmatocytes,which function as macrophages in immunity, wound healing andtissue remodeling (Tepass et al., 1994; Wood and Jacinto, 2007).During larval stages, the embryonic plasmatocytes migrate toReceived 27 December 2018; Accepted 18 April 2019

1Department of Biochemistry and Molecular Pharmacology, New York UniversitySchool of Medicine, New York, NY 10016, USA. 2Helen L. and Martin S. KimmelCenter for Stem Cell Biology, New York University School of Medicine, New York,NY 10016, USA.*Present address: Program in Epithelial Biology and Department of Dermatology,Stanford University School of Medicine, Stanford, CA 94305, USA.

‡Author for correspondence ([email protected])

A.A.B., 0000-0001-7551-4813; E.A.B., 0000-0002-5997-4489

This is an Open Access article distributed under the terms of the Creative Commons AttributionLicense (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use,distribution and reproduction in any medium provided that the original work is properly attributed.

1

© 2019. Published by The Company of Biologists Ltd | Disease Models & Mechanisms (2019) 12, dmm038679. doi:10.1242/dmm.038679

Disea

seModels&Mechan

isms

https://doi.org/10.1242/dmm.040535https://doi.org/10.1242/dmm.040535mailto:[email protected]://orcid.org/0000-0001-7551-4813http://orcid.org/0000-0002-5997-4489

‘hematopoietic pockets’, microenvironments located in eachsegment of the larval body wall (Markus et al., 2009; Makhijaniet al., 2011). In pockets, the peripheral nervous system supports

resident (or sessile) embryonic plasmatocytes, which self-renewand proliferate (Leitao and Sucena, 2015; Petraki et al., 2015).As a result, the embryonically derived pool of plasmatocytes

Fig. 1. See next page for legend.

2

RESEARCH ARTICLE Disease Models & Mechanisms (2019) 12, dmm038679. doi:10.1242/dmm.038679

Disea

seModels&Mechan

isms

increases 30-fold during larval stages. Sessile plasmatocytes aregradually released into circulation beginning in the second larvalinstar (Makhijani et al., 2011). However, they can be mobilized enmasse in response to infection (Markus et al., 2009;Makhijani et al.,2011; Gold and Brückner, 2015). Additionally, in response toimmune challenge, for example parasitization by ovidepositorywasps, plasmatocytes can transdifferentiate into lamellocytes, largeflat cells that encapsulate objects too large to be phagocytosed(Markus et al., 2009; Honti et al., 2010; Stofanko et al., 2010;Avet-Rochex et al., 2010; Anderl et al., 2016).The second wave of hematopoiesis occurs in the larval lymph

gland, an organ which serves as a reservoir of prohemocytes, whichdifferentiate primarily into plasmatocytes during second and thirdlarval instars (Tepass et al., 1994; Lebestky et al., 2000; Mandalet al., 2004; Jung et al., 2005). However, under immune-challengedconditions, lymph gland prohemocytes can also differentiate intolamellocytes (Jung et al., 2005; Rizki, 1978). The lymph glanddisintegrates in early pupal stages, releasing mature hemocytes intocirculation (Grigorian et al., 2011). Lineage-tracing studies haveshown that the adult hemocyte pool is derived from both embryonicand lymph gland hematopoiesis (Holz et al., 2003).The hopTum mutation is a temperature-sensitive, gain-of-function

mutation in the Drosophila JAK caused by a G341E substitution,which results in sustained activation of Stat92E (Fig. 1B and Luoet al., 1995; Harrison et al., 1995). This mutation causes a ‘flyleukemia’ characterized by the presence of melanotic tumors, blackmasses comprised of aggregated plasmatocytes and lamellocytes.Importantly, lamellocytes are always observed in geneticbackgrounds harboring melanotic tumors (Zettervall et al., 2004;Minakhina and Steward, 2006). The hopTum mutation is X-linkedand is lethal in hemizygous males, and the melanotic tumors are

manifest in heterozygous females (Corwin and Hanratty, 1976;Hanratty and Ryerse, 1981). Melanotic tumors appear black as aresult of activation of the prophenol oxidase pathway, and thedysregulation of hematopoiesis in hopTum larvae shares featureswith the larval immune response triggered by wasp-egg infestation(Minakhina and Steward, 2006; Yang and Hultmark, 2016).

Similar to MPN patients, the hopTum leukemic phenotypeoriginates in hematopoietic multipotent progenitors and can beadoptively transferred for multiple generations and up to 2 years(Hanratty and Ryerse, 1981). Tumors are not observed in thisgenotype until the middle of the third larval instar (Hanratty andRyerse, 1981; Lanot et al., 2001). In hopTum larvae, the stereotypicpattern of hematopoietic pockets is disrupted and sessileplasmatocytes are mobilized (Anderson et al., 2017). In hopTum

larvae, the lymph gland also prematurely histolyzes, releasing lymph-gland-derived plasmatocytes and lamellocytes into circulation(Hanratty and Ryerse, 1981; Lanot et al., 2001; Sorrentino et al.,2007; Anderson et al., 2017; Terriente-Félix et al., 2017). In hopTum

animals, the number of plasmatocytes and lamellocytes in circulationis dramatically increased, as result of upregulated plasmatocyteproliferation and massive induction of lamellocyte differentiation(Luo et al., 1995; Lanot et al., 2001; Silvers and Hanratty, 1984;Anderson et al., 2017; Bazzi et al., 2018). Sustained activation of theJAK/STAT pathway in plasmatocytes induces upd2 and upd3, whichencode pathway agonists (Fig. 1C). These ligands then activate theJAK/STAT pathway non-autonomously in body wall muscle (Yanget al., 2015;Bazzi et al., 2018;Yang andHultmark, 2017). JAK/STATactivation in muscle is required for the full maturation and function oflamellocytes in response to wasp infection, but the mechanism bywhich this occurs is not yet known (Yang et al., 2015).

By means of an F1 deficiency (Df) screen, we recently identified32 enhancer Dfs and 11 suppressor Dfs that modified the tumorburden of hopTum animals (Anderson et al., 2017). Df(2R)ED2219was among the strongest enhancers (Anderson et al., 2017). Here,we demonstrate that the gene uncovered by this Df and responsiblefor the enhancement of the hopTum leukemic phenotype is Enhancerof Polycomb [E(Pc)]. Despite its name, E(Pc) is not a component ofPolycomb group complexes but rather is part of the Tip60 complex,a member of the MYST family of lysine acetyltransferases (KATs).Although E(Pc) is not a catalytic subunit of Tip60, it is a criticalco-factor, and physical interactions between E(Pc) and Tip60regulate acetyltransferase activity (Searle et al., 2017; Xu et al.,2016). As KATs modify lysine residues in histones and otherproteins, these enzymes regulate numerous processes, includingprotein stability/turnover, chromatin remodeling and tumorigenesis(Sheikh and Akhtar, 2019). The Tip60 complex can act as a tumorsuppressor in human cancers. Mono-allelic loss of the human TIP60gene (KAT5) is a frequent event in mammary and head-and-neckcarcinomas and in human lymphoma (Gorrini et al., 2007; Zacket al., 2013). Tip60 expression is also significantly downregulated incolon and lung carcinoma (Lleonart et al., 2006). Furthermore, otherloss-of-function mutations (point mutations, deletions and othergenetic aberrations) in human KAT5, or in human EPC1 or EPC2[the human homologs of E(Pc)] are observed in numerous studies ofmyeloid and lymphoid malignancy in cBioPortal.org (Cerami et al.,2012; Gao et al., 2013). However, the causal factors that aredysregulated in human cancer upon downregulation of the TIP60complex are largely unknown.

Here, we show that E(Pc) heterozygosity significantly enhancesthe tumor burden in hopTum animals, similar to the effect of theenhancing deficiency Df(2R)ED2219 that uncovers it. Furthermore,we find that E(Pc) and Tip60 are required in the hematopoietic

Fig. 1. HopTum is a dominant-active kinase that causes hematopoietictumors. (A,B) Model of the Drosophila JAK/STAT pathway in wild-type (A) orhopTum (B) blood cells. In wild-type (A), an Upd cytokine binds to a dimeric cell-surface receptor, Dome. This induces the transactivation of associated Hoptyrosine kinases. These activated Hop proteins then phosphorylate the Domecytoplasmic domain. Inactive Stat92E proteins bind to the activated receptor,after which Stat92E becomes a substrate for Hop. Phosphorylated Stat92Edimers translocate to the nucleus, where they bind specific DNA sequencesand alter gene transcription. A well-established Stat92E target gene isSocs36E. (B) HopTum is a dominant-active kinase that activates Stat92E in aligand-independent manner. This mutation leads to sustained activation ofStat92E, sustained transcriptional responses by activated Stat92E andincreased expression of Stat92E target genes (bigger arrows). Brown circleswith ‘P’ indicate tyrosine phosphorylation events. (C) Model of dysregulatedhematopoiesis and melanotic tumor formation in hopTum larvae. Stat92Eactivation (pStat92E) in plasmatocytes leads to the induction of Upd2 andUpd3. These cytokines are released into the hemolymph and activate JAK/STAT signaling in larval muscle. This in turn is required for the differentiation oflamellocytes, a key cell type in the formation of melanotic tumors. Upd2 andUpd3 can act in an autocrine manner to increase proliferation ofplasmatocytes. The combination of ectopic differentiation of lamellocytes andthe expansion of the plasmatocyte population leads to the formation ofmelanotic tumors in larval stages (red arrowhead). (D) Graph of the tumorindex of hopTum adult females of the indicated genotypes. The tumor index forhopTum outcrossed to wild–type (hopTum/+, n=21 for D and n=46 for E) are thegray circles in D,E. This tumor index is significantly suppressed byheterozygosity for Stat92E (n=17 for D and n=25 for E) using the null alleleStat92E397 (D,E, purple circles). The tumor index of hopTum adult females isalso significantly suppressed when a dominant-negative version of Dome(DomeΔCyt, n=31) is mis-expressed in the hematopoietic compartment(D, yellow circles) or when hopTum females are also systemically heterozygousfor both upd2 and upd3 deletions (n=31) (E, orange circles). The genotypes forthe animals in panels D and E and all subsequent figures are listed in thesupplemental information. Error bars represent s.e.m. **P

compartment to repress myeloid lineage dysregulation and inhibitmelanotic tumor formation. The oncogenic potential of E(Pc) andTip60 depends on the presence of hop and Stat92E, as concomitantdepletion of either gene with E(Pc) or Tip60 severely perturbstumor formation. Furthermore, loss of E(Pc) or Tip60 leads tocell-autonomous increases in the activity of the Stat92E proteinand the expression of Stat92E target genes but does not alterexpression of the hop or Stat92E genes. Finally, depletion of E(Pc)significantly increases the levels of Hop protein expression. Ourmodel indicates that E(Pc) and the Tip60 complex [E(Pc)/Tip60] actas tumor suppressors by attenuating JAK/STAT signaling throughrepressing expression of the Hop protein.

RESULTSLoss of E(Pc) or Tip60 enhances hopTum tumorsAs mentioned above, plasmatocytes that harbor the HopTum proteinupregulate expression of the pathway ligands upd2 and upd3. Thesecytokines activate the JAK/STAT pathway non-autonomously in

body wall muscle, and we reasoned that they may also haveautocrine functions in plasmatocytes. To test this hypothesis, weblocked pathway activation in the hematopoietic compartment bymis-expressing a dominant-negative Dome receptor. We used thepan-hematopoietic driver HaHmlLT-Gal4, which is expressedstrongly in circulating and sessile hemocytes, the larval lymphgland and pericardial cells (Mondal et al., 2014; Anderson et al.,2017). We also observed occasional expression in small patches ofcells in the salivary gland and fat body (not shown), but it is notexpressed in larval muscle (Anderson et al., 2017). Indeed, when weinhibit pathway activation in plasmatocytes, we significantly impairthe formation of melanotic tumors in hopTum animals (Fig. 1D,yellow circles). These results support a model in which theproduction of Upd cytokines by plasmatocytes in hopTum larvaedrives the formation of melanotic tumors by autonomous and non-autonomous mechanisms (Fig. 1C). Consistent with theseobservations, reducing the systemic level of Stat92E or both upd2and upd3 significantly suppresses melanotic tumors in hopTum

Fig. 2. Heterozygosity forE(Pc), or the overlapping deficiencyDf(2R)ED2219, enhances hopTum tumorigenicity. (A) Graph of the tumor index of the indicatedgenotypes. The tumor index for hopTum outcrossed to wild-type (n=55) is the first set of gray circles. This value is significantly suppressed by concomitantheterozygosity for Stat92E mutation (n=34; purple circles). Df(2R)ED2219 heterozygosity (n=13) significantly enhances the tumor index (first set of green circles).Of the five Dfs that overlap with Df(2R)ED2219 [Df(2R)BSC703 (n=24), Df(2R)BSC336 (n=36), Df(2R)BSC304 (n=33), Df(2R)BSC358 (n=41), Df(2R)ED2222(n=17)], only Df(2R)ED2222 enhances the hopTum tumor index (second set of green circles). Of the four genes uncovered by Df(2R)ED2219 and Df(2R)ED2222[E(Pc), inv, en, tou], only alleles of E(Pc) enhanced the tumor index (third and fourth set of green circles). E(Pc)1 (n=25); E(Pc)w3 (n=36); inv30 (n=41); invKG04405

(n=26); invEenE (n=23); en1 (n=23); en4 (n=27); en54 (n=23); en59 (n=29); tou1 (n=50); tou2 (n=73); touKG02432 (n=28). Error bars represent s.e.m. *P

animals (Fig. 1E, orange circles; Luo et al., 1997; Hou et al., 1996;Yan et al., 1996; Bazzi et al., 2018).We previously performed a deficiency screen to identify dominant

modifiers of hopTum tumors (Anderson et al., 2017). One of the mostrobust enhancers in this screen was Df(2R)ED2219 (Fig. 2A). Todetermine the underlying gene responsible for this modification, wesurveyed five deficiencies that overlapped Df(2R)ED2219 (Fig. 2B).Of these deficiencies, onlyDf(2R)ED2222 significantly enhanced thehopTum tumor phenotype (Fig. 2A). The genomic region common toDf(2R)ED2219 and Df(2R)ED2222 uncovered four genes: E(Pc),invected (inv), engrailed (en) and toutatis (tou) (Fig. 2B). Using loss-of-function alleles for each candidate gene, we determined that onlyE(Pc) hypomorphic alleles significantly enhanced the hopTum tumorphenotype (Fig. 2A). These data strongly implicateE(Pc) as being thegene responsible for the enhancement caused by Df(2R)ED2219heterozygosity. Owing to male lethality, we did not screen Dfs on theX chromosome for interactions with hopTum, and as a result, we were

not able to determine whether the Dfs uncovering the X-linked geneTip60 interacted in the screen (Anderson et al., 2017). Instead, weused RNA interference (RNAi) and mis-expression of a dominant-negative Tip60 (see below).

We next assessed whether the E(Pc) global heterozygousphenotype was actually due to loss of E(Pc) in the hematopoieticcompartment. We significantly depleted E(Pc) or Tip60 using thehematopoietic driver HaHmlLT-Gal4 (Fig. S1). Hematopoieticdepletion of either E(Pc) or Tip60 significantly increased thehopTum tumor phenotype (Fig. 3A, green and pink circles), as didmis-expression of dominant-negative Tip60E431Q (Fig. 3B, yellowcircles). We previously established that mis-expression ofUAS-hopTum by HaHmlLT-Gal4 induced lamellocytes in a cell-autonomous manner (Anderson et al., 2017). It is important to notethat mis-expression of wild-type Stat92E alone does not induce JAK/STAT target genes (Ekas et al., 2010). To further confirm that theE(Pc)/Tip60 phenotype is autonomous to the hematopoietic lineage,

Fig. 3. E(Pc) works through the Tip60 complex to repress hopTum melanotic tumors. (A) In the endogenous hopTum background (gray circles, n=50),hematopoeitic depletion of E(Pc) using the GD12282 line [light green circles, labeled >E(Pc)-i 1, n=8] or using the JF03101 line [dark green circles, labeled>E(Pc)-i 2, n=18)] or Tip60 (pink circles, n=20) enhances the tumor phenotype. The purple circles represent the suppression of tumor burden by heterozygosityfor Stat92E mutation (n=11). (B) Increasing the dose of Tip60 by mis-expressing a wild-type version (>Tip60WT, n=34; dark red circles) significantly reduces thetumor burden in hopTum (endogenous allele, n=30; gray circles). By contrast, mis-expressing a dominant-negative Tip60 (>Tip60E431Q, n=4; yellow circles)significantly enhances the tumor burden in hopTum (endogenous allele) animals. We note that very few adult female hopTum/+; HaHmlLT>Tip60E431Q eclose,suggesting pupal lethality. The purple circles represent the suppression of tumor burden by heterozygosity for Stat92E mutation (n=24). (C) Mis-expression ofUAS-hopTum (n=64) in the hematopoietic compartment using the HaHmlLT-Gal4 driver induces melanotic tumors (gray circles). Concomitant depletion of E(Pc)using the GD12282 line [green circles, labeled >E(Pc)-i 1, n=15] or Tip60 (pink circles, n=36) significantly increases the tumor burden in these animals.(D) Increasing the dose of E(Pc) by supplying an additional genomic copy [g-E(Pc)-GFP, n=33] or by mis-expressing a wild-type E(Pc) tagged with GFP [>E(Pc)-GFP, n=10] (dark and light blue circles, respectively) significantly reduces the tumor burden in endogenous hopTum animals (gray circles, n=18), similar toheterozygosity for Stat92E mutation (purple circles, n=8). By contrast, heterozygosity for the E(Pc)1 loss-of-function allele (brown circles, n=8) significantlyenhances the tumor burden in endogenous hopTum animals. Error bars represent s.e.m. **P

we concomitantly depleted either gene while mis-expressing UAS-hopTum. Either manipulation resulted in a significant enhancement ofthe tumor burden (Fig. 3C, green and pink circles).We next addressedthe converse: whether increasing the dose of wild-type E(Pc) or wild-type Tip60 would suppress the hopTum tumor phenotype. Weincreased the E(Pc) genomic dose using a GFP-tagged genomiccopy [g-E(Pc)-GFP] or mis-expressed a wild-type version of eitherfactor. In all three scenarios, increasing the level of wild-typeE(Pc) orTip60 significantly suppressed the hopTum tumor phenotype and tothe same degree as halving the genetic dose of Stat92E (Fig. 3B,D).hopTum/Ymales normally die before adulthood (Fig. S2, gray bar), butthey eclose in considerable numbers when they are systemicallyheterozygous for Stat92E, or when Stat92E or hop arehematopoietically depleted (Fig. S2, purple, green and yellow bars,respectively). Remarkably, increasing hematopoietic expression ofE(Pc) or Tip60 also rescued hopTum/Y adult males (Fig. S2, blue andred bars). Since these surviving hopTum males, regardless ofmanipulation, still contain melanotic tumors, it is not clearmechanistically how hematopoietic depletion of hop or Stat92E, orhematopoieticmis-expression ofE(Pc) or Tip60, rescues them. Takentogether, these results indicate that E(Pc) and Tip60 act in thehematopoietic lineage to repress hopTum tumor formation and hopTum/Y male lethality.

Hematopoietic depletion of E(Pc) or Tip60 in wild–typeanimals leads to ectopic lamellocyte differentiation andmelanotic tumorsIn wild-type, healthy larvae, the pool of circulating hemocytescomprises primarily plasmatocytes; lamellocytes are not observed(Honti et al., 2014). However, after immune challenge, lamellocytesquickly differentiate frommultipotent progenitors or transdifferentiatefromplasmatocytes (Honti et al., 2010; Jung et al., 2005; Anderl et al.,2016). To assess the role of E(Pc) and Tip60 in hemocytedevelopment, we depleted either factor in an otherwise wild-typebackground using the HaHmlLT-Gal4 driver. In hemolymph bleedsfrom control larvae that only expressed the driver, we observedplasmatocytes but not lamellocytes (Fig. 4A,B). By contrast, in bleedsfrom animals hematopoietically mis-expressing hopTum, we foundan expansion of plasmatocytes and ectopic differentiation oflamellocytes (Fig. 4C,D; see the outlined cell, arrow, in 4F for anexample of a lamellocyte). Furthermore, in this background, weobserved the formation of clusters of lamellocytes and plasmatocytes(Fig. 4C,D), hereafter referred to as ‘microtumors’ (see Fig. 4D for anexample and Materials and Methods for a description ofmicrotumors). Strikingly, when we depleted E(Pc) or Tip60 or mis-expressed dominant-negative Tip60E431Q in wild-type hemocytes,we detected plasmatocytes, lamellocytes and microtumors inlarval bleeds and frank melanotic tumors in adults (Fig. 4E-K).To determine whether hematopoietic depletion of E(Pc) increasesthe number of larval hemocytes or the number of lamellocytes, orboth, we counted circulating hemocytes from control larvae(HaHmlLT>GFP/+), larvae with hematopoietic mis-expression ofhopTum (HaHmlLT>hopTum), or larvae with hematopoietic depletionof E(Pc) (HaHmlLT>E(Pc)RNAi). Control larvae had ∼3000estimated total circulating hemocytes and no lamellocytes, asexpected (Fig. 4L,M). By contrast, HaHmlLT>hopTum larvae hadsignificantly more estimated total circulating hemocytes (∼15,000)and 30% lamellocytes (Fig. 4L,M), consistent with a prior report(Zettervall et al., 2004). In HaHmlLT>E(Pc)RNAi, the estimatednumber of total hemocytes was significantly increased (∼5600)compared to controls, and the percentage of lamellocytes wasdramatically increased to 38% (Fig. 4L,M). These results

indicate that the E(Pc) RNAi tumors are caused primarily byectopic lamellocyte differentiation and secondarily by increasedplasmatocyte proliferation.

We also determined that depletion of seven other Tip60components, including Brahma associated protein 55kD(Bap55) and SWI2/SNF2 family member domino (dom), from thehematopoietic compartment, caused ectopic differentiation oflamellocytes and the formation of microtumors (Fig. 4N-P andTable 1). These phenotypes are consistent with prior reports ofectopic lamellocyte differentiation upon E(Pc) knockdown inlymph gland prohemocytes and in dom mutant larvae (Owusu-Ansah and Banerjee, 2009; Braun et al., 1998). These data indicatethat the E(Pc)/Tip60 complex represses lamellocyte differentiationand tumor formation in wild-type animals.

E(Pc)actsupstreamof or in parallel tohop in tumor formationWe developed an assay in which we quantified the GFPfluorescence intensity of aggregated hemocytes in the larvalcirculatory system as a proxy for tumor formation (see Materialsand Methods). Bleeds from HaHmlLT>hopTum larvae hadsignificantly higher GFP intensity compared to control bleeds(Fig. 5A, purple triangles), and these aggregates consisted of GFP-labeled plasmatocytes and lamellocytes (Fig. 4D). Depletion ofStat92E or hop significantly decreased hemocyte aggregates inHaHmlLT>hopTum, demonstrating that these aggregates result fromincreased JAK/STAT activity (Fig. 5A, green circles and redtriangles, respectively). Furthermore, we showed that hematopoieticdepletion of Stat92E or hop significantly reduced the tumor burdenin animals heterozygous for the endogenous hopTum allele (Fig. 5B,green circles and red triangles, respectively). These resultsdocument the efficacy of the Stat92E and hop RNAi transgenesin suppressing both larval hemocyte aggregates and adultmelanotic tumors. As expected, hemocyte aggregation in controlanimals was minimal (Fig. 5C,E, blue circles), and depletion ofStat92E or hop in an otherwise wild-type background also producedminimal aggregation (Fig. 5C,E, respectively, purple triangles).Hematopoietic depletion of E(Pc) significantly increased hemocyteaggregation in larvae and tumor burden in adults (Fig. 5C-F,green circles). To test whether the E(Pc) RNAi hemocyteaggregation and tumor phenotypes depended on JAK/STATsignaling, we concomitantly depleted E(Pc) and Stat92E or E(Pc)and hop. Indeed, knockdown of either Stat92E or hop significantlysuppressed both phenotypes (Fig. 5C-F, red triangles). These dataindicate that tumor initiation caused by E(Pc) depletion isdependent on the JAK/STAT pathway.

JAK/STAT signaling is repressed by the E(Pc)/Tip60 complexOur results thus far raised the possibility that the E(Pc)/Tip60complex regulates activity of the JAK/STAT pathway. To addressthis issue, we assessed whether E(Pc) depletion could cell-autonomously increase Stat92E activation. We monitored Stat92Eactivity using an established Stat92E transcriptional reporter(10xStat92E-DsRed) containing regulatory sequences from aStat92E target gene, Socs36E, to drive expression of thefluorescent protein DsRed (Bach et al., 2007). Control hemocytesfrom uninfected wild-type animals displayed low levels of DsRed(Fig. 6A). As expected, this reporter was strongly upregulated inhemocytes from hopTum larvae (Fig. 6B). Upon E(Pc)hematopoietic depletion, we also observed a robust induction of10xStat92E-DsRed (Fig. 6C). Furthermore, the Stat92E target genesSocs36E, chinmo and zfh1 were significantly upregulated inhemocytes upon depletion of E(Pc) or Tip60 (Fig. 6D-F, green

6


Disea

seModels&Mechan

isms

http://dmm.biologists.org/lookup/doi/10.1242/dmm.038679.supplementalhttp://dmm.biologists.org/lookup/doi/10.1242/dmm.038679.supplementalhttp://dmm.biologists.org/lookup/doi/10.1242/dmm.038679.supplemental

and pink circles). The upd2 and upd3 genes were induced inhemocytes in the endogenous hopTum background as well as uponhematopoietic mis-expression of UAS-hopTum (Fig. 6G,H, white

and black circles), consistent with a recent report (Bazzi et al.,2018). Both upd2 and upd3 were also significantly upregulated inhemocytes depleted for E(Pc) or Tip60 (Fig. 6G,H, green and pink

Fig. 4. Hematopoietic depletion of E(Pc)/Tip60 induces precocious lamellocyte differentiation and tumor formation. (A-B′) Bleeds from control larvae(HaHmlLT>GFP) contain P1-positive plasmatocytes (red, A) but not L1-positive lamellocytes (red, B). (C-D′) Larvae with hematopoietic mis-expression ofUAS-hopTum have an increase in plasmatocytes (C) and an ectopic differentiation of lamellocytes (D). (E-J) Hematopoietic depletion of E(Pc) (E,F) or Tip60(G,H) or hematopoieticmis-expression of a dominant-negative Tip60E431Q (I,J) in an otherwisewild-type background induces lamellocyte differentiation. F-actin isblue in A-J. Experiments in A-J were performed at least three times with similar results. (K) Graph of tumor indices from adult females. Melanotic tumorsare not observed in control (HaHmlLT>GFP) animals (gray circles, labeled ‘Control’, n=20). E(Pc) or Tip60 depletion using the same driver increases the tumorburden (green and pink circles, respectively, n=33 for both genotypes). Mis-expression of Tip60E431Q also leads to melanotic tumors in wild-type animals (yellowcircles, n=47). Error bars represent s.e.m. *P

circles). Importantly, transcription of the Stat92E gene was notincreased in E(Pc)/Tip60-depleted hemocytes, ruling out the modelthat E(Pc)/Tip60 negatively regulates JAK/STAT activity byrepressing the Stat92E gene (Fig. 6I, green and pink circles).Taken together, these data demonstrate that, in wild-typehemocytes, E(Pc)/Tip60 negatively regulates JAK/STAT activity,thereby restricting lamellocyte differentiation.

Hematopoietic depletion of E(Pc) increases Hop proteinexpression in vivoOur genetic studies demonstrate that loss of E(Pc)/Tip60 (i.e.decrease in lysine acetylation) leads to an increase in JAK/STATsignaling. Inhibitors of lysine deacetylases (termed KDACi)significantly inhibited proliferation of MPN cells and reduceddisease burden in a preclinical mouse model of PV (Guerini et al.,2008; Akada et al., 2012). These data suggest that lysineacetyltransferases could regulate the level of Hop protein. Wetested this hypothesis by measuring Hop protein expression using anepitope-tagged hop-GFP-V5 expressed under endogenous regulatorysequences on a bacterial artificial chromosome in transgenic flies(Sarov et al., 2016). We isolated circulating hemocytes from controlhop-GFP-V5 larvae or from experimental hop-GFP-V5 larvae inwhich E(Pc) was hematopoietically depleted. Hop-GFP-V5 proteinwas normalized to expression of the 27 kDa GFP protein driven byHaHmlLT>GFP. In seven independent experiments, we observed asignificant, 2-fold increase in the level of Hop-GFP-V5 protein inE(Pc)-depleted hemocytes compared to controls (Fig. 7A, quantifiedin 7B). We assessed whether the increase in Hop protein expressionwas due to increased transcription of the hop gene. We performedqPCR analysis using four independent pairs of primers expandingexon junctions of the hop transcript. As expected, hemocytes from theendogenous hopTummutant did not have a significant increase in hoptranscription (Fig. 7C, white circles). By contrast, in the positivecontrol, hemocytes mis-expressing UAS-hopTum had a significantupregulation of hop expression (Fig. 7C, black circles). Importantly,E(Pc)- or Tip60-depleted hemocytes did not display any alteration inhop gene expression (Fig. 7C, green and pink circles). These resultsstrongly suggest that E(Pc) negatively regulates Hop protein levels inDrosophila blood cells. To determine whether the converse is true,we assessed whether KDACi treatment would reduce Hop proteinlevels.We transfected hemocyte-derived S2 cells with aMyc epitope-

tagged Hop (Hop-Myc-His), and then treated the cells with twodifferent KDACi [1 µM trichostatin A (TSA) or 3 mM sodiumbutyrate (NaBut)] for 16 h. In five independent experiments, weobserved a significant reduction in the level of Hop–Myc protein inKDACi-treated S2 cells compared to vehicle DMSO-treated controls(Fig. 7D, quantified in 7E). Hop protein was reduced by 29.5% with3 mM NaBut treatment and by 39.8% with 1 µM TSA treatment.Taken together, these data indicate that the activity of the lysineacetyltransferase E(Pc)/Tip60 directly or indirectly regulates the levelof Hop protein in Drosophila blood cells, thereby controlling thelevel of JAK/STAT pathway activity.

DISCUSSIONIn this study, we characterized the deficiency Df(2R)ED2219, whichsignificantly enhanced the tumor burden in hopTum animals (Andersonet al., 2017). After testing additional deficiencies that overlapped withDf(2R)ED2219, we found that only Df(2R)ED2222 recapitulated thisenhancement. These two overlapping deficiencies uncovered fourgenes, only one of which – E(Pc) – enhanced the hopTum tumorphenotypewhen heterozygous.We found that the enhancement causedby loss ofE(Pc)was due to its role in the hematopoietic system throughthe Tip60KAT complex. Furthermore, we showed that E(Pc) is criticalfor repressing the precocious differentiation of lamellocytes, that thisrepression is JAK/STAT dependent, and that the depletion of E(Pc)causes a substantial increase in JAK/STAT signaling by increasinglevels of the Hop protein. This leads to increased hemocyte productionof Upd2 and Upd3 by activated Stat92E, which then presumablytriggers JAK/STAT signaling non-autonomously in muscle andincreases lamellocyte differentiation by means of a currentlyunknown mechanism. As a result, melanotic tumors are significantlylarger in hopTum animals that are heterozygous for an E(Pc) mutationcompared with hopTum heterozygous for a wild-type allele.

A recent publication reported that E(Pc) repressed expression of theStat92E gene by directly binding regions near the transcription startsite in somatic cells of theDrosophila testis (Feng et al., 2017).Whilewe cannot rule out the possibility that the reduction of E(Pc)/Tip60causes precocious lamellocyte differentiation as a result of changes inhistone acetylation, we did not observe a change in Stat92Etranscripts upon depletion of E(Pc)/Tip60. Therefore, the E(Pc)/Tip60 loss-of-function phenotype in the hematopoietic compartmentis not a result of ectopic expression of the Stat92E gene. We doobserve cell-autonomous increases in the expression of Stat92E targetgenes upon E(Pc)/Tip60 depletion, arguing that increased Stat92Eactivity occurs upon loss of E(Pc)/Tip60 in hemocytes. Nevertheless,future work will be needed to determine whether E(Pc)/Tip60 acts onchromatin and/or histones at Stat92E target genes or at genes thatregulate pathway activity. Loss of E(Pc)/Tip60 appears to elicit cell-type-specific responses. For example, Tip60 interacts with thetranscription factor Myc to maintain Drosophila neural stem cells(Rust et al., 2018), and regulates expression of CyclinB or germlinedifferentiation genes in germline cells in the Drosophila ovary(McCarthy et al., 2018; Feng et al., 2018).

Our unbiased genetic screen identifiedmutations inE(Pc) as potentenhancers of hematopoietic tumors inDrosophila, indicating that thisphenotype is sensitive to the reciprocal activity of KATs and KDACs.We favor the interpretation that E(Pc)/Tip60 negatively regulates thelevels of Hop protein either directly or indirectly and that thisregulatory mechanism is a causal event in melanotic tumor formationin hopTum. Our model (Fig. 8) is supported by several lines ofevidence. First, numerous groups have reported that upregulation ofHopWT or HopTum protein is sufficient to induce melanotic tumors,indicating that Hop protein levels are causal to this oncogenic

Table 1. Hematopoietic depletion of Tip60 complex componentsinduces lamellocyte differentiation

Gene depleted Lamellocytes present?

Bap55 ✓Brd8 ✓MRG15 Not determineddomino ✓dPontin Not determinedGas41 ✓Ing3 ✓Nipped-A ✓reptin ✓

Tip60 complex components were depleted in the hematopoietic compartmentusing the HaHmlLT-Gal4 driver. The entire microscope slide containing thefixed and stained hemolymph was scanned by confocal microscopy.Lamellocytes were not observed in bleeds from HaHmlLT-Gal4 controls.However, in the genotypes indicated by the tick, lamellocytes were observed inat least several fields of the confocal images. Although less penetrant thanE(Pc) or Tip60 depletion, depletion of seven different components of the Tip60complex induced ectopic lamellocyte differentiation. EachRNAi linewas testedin at least three independent experiments.

8


Disea

seModels&Mechan

isms

phenotype (Harrison et al., 1995; Zettervall et al., 2004; Luo et al.,1995; Bazzi et al., 2018; Anderson et al., 2017). Second, depletion ofhop significantly reduces tumor formation caused by loss of E(Pc).Third, the level of Hop protein in purified larval blood cells issignificantly increased upon depletion of E(Pc). Fourth, the level ofHop protein in cultured S2 cells is significantly decreased upontreatment with KDACi. Lastly, the E(Pc)/Tip60 complex does not acton the hop gene: hematopoietic depletion of complex componentsdoes not increase hop transcripts. Currently, we do not know whetherthe E(Pc)/Tip60 complex regulates Hop protein expression at thelevel of translation of the hop mRNA, or at the level of Hop proteinstability and/or turnover. Future work will be needed to addressthese issues.As noted above, KDACi, which increase lysine acetylation, were

highly effective in inhibiting proliferation of MPN cells and inreducing disease burden in a preclinical PV mouse model (Gueriniet al., 2008; Akada et al., 2012). The chromatin-independent effects

of KDACi inMPN cells result in large part from deacetylation of thechaperone Hsp90, of which JAK2 is a client protein (Bali et al.,2005; Wang et al., 2009; LaFave and Levine, 2012). Treatment ofMPN cell lines with pan-KDACi panobinostat disrupts theinteraction between Hsp90 and JAK2V617F (Wang et al., 2009).Additionally, the Hsp90 inhibitor PU-H71 is effective at degradingJAK2V617F in MPN cell lines and primary patient samples(Marubayashi et al., 2010). Moreover, administration of PU-H71in preclinical PV and ET mouse models normalized blood counts,reduced allele burden and increased mean survival (Marubayashiet al., 2010). It is intriguing to speculate that a similar regulatorynetwork exists between Hsp83, the Drosophila ortholog of Hsp90,and Hop in Drosophila myeloid-like cells, but future experimentswill be needed to test this hypothesis. In sum, our results revealremarkable similarities between the regulation of JAK proteins byKATs and KDACs across species and highlight the power ofDrosophila as a low-complexity model for human MPNs.

Fig. 5. The E(Pc)-depletion phenotype isabrogated by concomitant knockdown ofStat92E or hop. (A,C,E) Graphs of larvalhemocyte aggregation. (B,D,F) Graphs of adulttumor indices. (A) Hemocyte aggregation issignificantly increased in HaHmlLT>hopTum

(purple triangles, n=8) compared to controls (bluecircles, n=8). Depletion of Stat92E (green circles,n=8) or hop (red triangles, n=9) inHaHmlLT>hopTum larvae significantly reduceshemocyte aggregation. (B) The adult tumor indexof endogenous hopTum/+ (blue circles, n=22) issignificantly suppressed when Stat92E397 isheterozygous (purple triangles, n=10), or whenStat92E or hop is hematopoietically depleted(green circles and red triangles, n=5 and 17,respectively). (C,D) Hematopoietic depletion ofE(Pc) [green circles, labeled E(Pc)-i 2, n=20 for Cand n=63 for D] in an otherwise wild-typebackground significantly increases hemocyteaggregation (C) or adult tumor burden (D)compared to control (blue circles, n=21 for C andn=40 for D). Concomitant depletion of Stat92Eand E(Pc) (red triangles, n=19 for C and n=17 forD) significantly reduces hemocyte aggregationcompared to depletion of only E(Pc). There is nosignificant difference between the control (bluecircles) and depletion of only Stat92E (purpletriangles, n=19 for C and n=31 for D). (E,F)Hematopoietic depletion of E(Pc) [green circles,labeled E(Pc)-i 1, n=21 for E and n=57 for F]significantly increases hemocyte aggregation (E)or adult tumor burden (F) compared to control(blue circles, n=20 for E and n=55 for F).Concomitant depletion of hop suppresseshemocyte aggregation due to E(Pc) depletion(red triangles, n=15 for E and n=69 for F). Therewas no significant difference between the control(blue circles) and depletion of only hop (purpletriangles, n=17 for E and n=90 for F). Error barsrepresent s.e.m. **P

MATERIALS AND METHODSFly stocks and husbandryThe following stocks were obtained from the Bloomington DrosophilaStock Center: Stat92E397; hopTum; Df(2R)ED2219; Df(2R)BSC703;

Df(2R)BSC336; Df(2R)BSC304; Df(2R)BSC358; Df(2R)ED2222; E(Pc)1;E(Pc)w3; inv30; invKG04405; invE enE; en1; en4; en54; en59; tou1; tou2;touKG02432; upd2Δ upd3Δ; E(Pc) TRiP RNAi line JF03101 [termed E(Pc)-i2]; Tip60 TRiP RNAi line HM05049; hop TRiP RNAi line GL00305;

Fig. 6. E(Pc)/Tip60 autonomously represses Stat92E activity. (A) P1-positive plasmatocytes (blue) from bleeds of control larvae display a low level of Stat92Eactivity as monitored by 10xSTAT-DsRed (red). (B) P1-positive plasmatocytes (green) from bleeds of hopTum larvae display a high level of Stat92E activity asmonitored by 10xSTAT-DsRed (red). (C) P1-positive plasmatocytes (green) from larvae hematopoietically depleted for E(Pc) display high Stat92E activity asmonitored by 10xSTAT-DsRed (red). Scale bars: 50 μm. Experiments in A-C were performed at least three times with similar results. (D-I) qPCR on mRNAs fromcirculating hemocytes from these larvae: control (gray circles), endogenous hopTum/+ (white circles); UAS-hopTum (black circles); E(Pc) RNAi 1 (light greencircles); E(Pc) RNAi 2 (dark green circles); Tip60 RNAi (pink circles). (D-F) Transcription of Stat92E target genes, Socs36E (D), chinmo (E) and zfh1 (F) aresignificantly increased in hemocytes from endogenous hopTum, UAS-hopTum, E(Pc) RNAi (using either RNAi construct) and Tip60 RNAi compared to thecontrol. For D and E, n=12 (control); n=9 (hopTum/+); n=9 (UAS-hopTum); n=10 [E(Pc)-i 1]; n=10 [E(Pc)-i 2]; n=9 (Tip60-i). For F, n=7 (control); n=6 (hopTum/+);n=9 (UAS-hopTum); n=5 [E(Pc)-i 1]; n=6 [E(Pc)-i 2]; n=5 (Tip60-i). (G,H) Transcription of upd2 (G) and upd3 (H) are significantly upregulated in hemocytesin endogenous hopTum, UAS-hopTum, E(Pc) RNAi (using either RNAi construct) and Tip60 RNAi compared to the control. For G and H, n=9 (control); n=7(hopTum/+); n=6 (UAS-hopTum); n=6 [E(Pc)-i 1]; n=7 [E(Pc)-i 2]; n=6 (Tip60-i). (I) Transcription of theStat92E gene is not increased in any of the genotypes tested.Rather, Stat92E mRNA is significantly decreased in UAS-hopTum, E(Pc) RNAi (using either RNAi construct) and Tip60 RNAi, but not endogenous hopTum,compared to the control. For I, n=11 (control); n=7 (hopTum/+); n=8 (UAS-hopTum); n=9 [E(Pc)-i 1]; n=9 [E(Pc)-i 2]; n=8 (Tip60-i). Error bars represent s.e.m.*P

Bap55 TRiP RNAi line HMS04015; and the dom TRiP RNAi linesHMS00192, HMS01855 and HMS02208.

The following RNAi stocks were acquired from the Vienna DrosophilaRNAi Center: E(Pc)GD12282 (v35268) [termed E(Pc)-i 1]; Stat92EGD4492

(v43866); Bap55GD11955 (v24703); Brd8GD8354 (v41530), Brd8KK107830

(v110618); dMRG15GD11902 (v43802); dMRG15KK107689 (v107689);domGD1420 (v7787); dPontinKK101103 (v105408); Gas41GD4100 (v12616);Gas41KK101151 (v106922); Ing3GD11989 (v52510); Ing3KK107543 (v109799);Nipped-AGD15595 (v40789); reptinGD4651 (v19021); and reptinKK105732

(v103483).The UAS-DomeΔCyt stock was a gift of James Castelli-Gair Hombria,

Centro Andaluz de Biología del Desarrollo (Brown et al., 2001). The E(Pc)-genomic-GFP, UAS-E(Pc) and UAS-E(Pc)-GFP stocks were gifts from DrXin Chen, Johns Hopkins University (Feng et al., 2017). The UAS-Tip60WT

and UAS-Tip60E431Q stocks were gifts of Dr Felice Elefant, DrexelUniversity (Lorbeck et al., 2011). The Hand-Gal4, HmlΔ-Gal4, UAS-

FLP.JD1, UAS-2xEGFP; Gal4-Act5C (FRT.CD2) (referred to asHaHmlLT-Gal4) stock was a gift from Dr Utpal Banerjee, University of California, LosAngeles (Mondal et al., 2014). We also used UAS-hopTum (Anderson et al.,2017). 10x-Stat92E-DsRed was a gift of Dr Martin Zeidler, SheffieldUniversity, UK.

Tumor indicesTumor indices were scored as described in Anderson et al. (2017). Wecrossed hopTum females to males that carried deficiencies or alleles. For eachexperiment, we set up in parallel: (1) a cross of hopTum/FM7 virgins to OreR

males, the progeny of which were used as a baseline control, and (2) a crossof hopTum/FM7 virgins to Stat92E397/TM6B, Tb males, the progeny ofwhich were used tomark suppression of the tumor phenotype.We scored themelanotic tumors in adult F1 progeny according to howmany quarters of theadult abdominal segments they encompassed. For instance, if a tumorcovered one quarter of a segment, it was given a score of 0.25, whereas a

Fig. 7. E(Pc) depletion in the hematopoieticcompartment increases Hop proteinexpression. (A) Western blot with anti-GFP oranti-V5 immunoprecipitates for the Hop-GFP-V5protein (top panel) or GFP from total cell lysates(bottom panel) from control and E(Pc)-depletedhemocytes. Hop-GFP-V5 is ∼150 kDa and GFP is27 kDa. This experiment was performed seventimes independently, with similar results.A representative western blot is shown here.(B) Graph of Hop-GFP-V5 protein normalized tothe GFP input in the total cell lysate for control andE(Pc)-depleted hemocytes. Hop-GFP-V5 proteinis significantly increased in E(Pc)-depletedhemocytes compared to control hemocytes. N=7independent experiments. (C) qPCR on mRNAsfrom circulating hemocytes from these larvae:control (gray circles, n=8), endogenous hopTum/+(white circles, n=6); UAS-hopTum (black circles,n=5); E(Pc) RNAi 1 (light green circles, n=6); E(Pc)RNAi 2 (dark green circles, n=6); Tip60 RNA-i(pink circles, n=5). Expression of the hop gene isnot increased in hemocytes from endogenoushopTum or depleted E(Pc) and Tip60. As expected,when hopTum is mis-expressed, hop mRNA issignificantly increased. (D) S2 cells weretransfected with Act5c-hop-cMyc-His and treatedwith 3 mM NaBut and 1 µM TSA for 16 h. Cellswere lysed 48 h post-transfection. Western blottingof Hop-Myc-His immunoprecipitates (top blot) wasperformed using an anti-Myc antibody and theinput (bottom blot) using an anti-actin antibody.This experiment was performed five timesindependently, with similar results. Arepresentative western blot is shown here.(E) Graph of normalized Hop-Myc-His for eachtreatment. NaBut and TSA treatments significantlyreduced Hop-Myc-His expression by 29.5% and39.8%, respectively. N=5 independentexperiments. Error bars represent s.e.m.**P

tumor that covered one entire abdominal segment was given a score of 1.0.Each individual progeny was given a tumor index (TI), which is the sum ofall tumor sizes per animal. The TI of each genotype corresponds to theaverage of all individual TIs in that genotype. TIs were graphed withstandard error bars by GraphPad Prism 7. The minimum sample size of eachgenotype was 15, and all crosses were repeated at least three times.

AntibodiesFor immunohistochemistry, we used mouse anti-P1 (for plasmatocytes at1:10) or mouse anti-L1 (for lamellocytes at 1:10), both gifts from Dr IstvanAndó (Kurucz et al., 2007). We used ToPro (Thermo Fisher Scientific) tolabel DNA at 1:1000 and phalloidin (Thermo Fisher Scientific) to labelF-actin at 1:25. Fluorescent secondary antibodies were obtained fromJackson ImmunoResearch and used at 1:200. For immunoprecipitation, weused 1 µl/sample of mouse anti-V5 monoclonal (Thermo Fisher Scientific,#R960-25) or 1 µl/sample of concentrate mouse anti-myc monoclonal[clone 9E10, Developmental Hybridoma Studies Bank (DHSB)]. Forwestern blotting, we used mouse anti-V5 monoclonal at 1:1000, rabbit anti-GFP polyclonal (Invitrogen, #A6455) at 1:1000, 9E10 (DHSB) at 1:1000,or mouse anti-actin (Millipore Sigma, #MAB1501) at 1:5000. We used goatanti-mouse IgG secondary antibody DyLight™ 680 conjugated (Rockland,#610-144-002) and goat anti-rabbit IgG secondary antibody DyLight™ 680conjugated (Rockland, #610-145-002), both at 1:10,000 dilution.

Hemocyte isolation and immunohistochemistryWandering third-instar larvae were washed in 1× phosphate buffered saline(1× PBS). The larvae were then dissected into pap pen wells drawn onSuperfrost Plus microscope slides (Thermo Fisher Scientific, catalog #1255034). To isolate the hemocytes in ‘bleeds’, the larval cuticle waspunctured using fine forceps, and the hemolymph was allowed to extrudefrom the hemocoel into 30 μl 1× PBS. In subsequent steps, microscopeslides were kept in humidified chambers. Hemocytes were allowed to settleonto the slide for 30-45 min. Hemocytes were fixed by the addition of12.5 μl of 16% paraformaldehyde (PFA) into the 30 μl 1× PBS for a finalconcentration of 4% PFA and incubated for 10 min. The fixative solutionwas removed manually, and samples were washed twice for 10 min each in1× PBS-T (0.01% Triton X-100 in 1× PBS). The following steps were

performed with mild agitation. Hemocytes were blocked in 10% normalgoat serum (NGS; Vector Laboratories, S-1000) in 1× PBS-T for 1 h atroom temperature or overnight at 4°C. Primary antibodies were diluted inblocking solution and incubated overnight at 4°C. Samples were washedtwice with 1× PBS-T for 10 min. Secondary antibodies were diluted inblocking solution and incubated for 2 h at room temperature. Hemocyteswere washed twice with 1× PBS-T for 10 min. Samples were mounted inVectaShield (Vector Laboratories, H-1000). Images of the samples werecaptured using a Zeiss LSM510 confocal microscope at 20×, 40× or 63×magnification.

Hemocyte countsTo count circulating hemocytes, staged, third-instar larvae were washed in1× PBS and then the hemolymph was bled to 20 μl 1× PBS, and 10 μl of thetotal volume was transferred onto a hemocytometer. The total number ofcells were counted, multiplied by the original volume (20 µl) and theaverage number of hemocytes per larva was calculated. At least 15 larvae ofeach genotype were counted, and the counting was carried out in triplicates.We plotted the average number of hemocytes/larva and the percentage oflamellocytes using GraphPad Prism7. We used the Student’s t-test todetermine statistical significance.

MicrotumorsMicrotumors were classified as aggregates of hemocytes that containlamellocytes (as defined by F-actin, L1 and/or morphology), that were notmelanized and that were at least 50 µm in size.

Hemocyte aggregation assayThe hemocyte aggregation assay is essentially a procedure to quantifymicrotumors from the entire hemolymph of a single larva. Wandering third-instar larvae were washed in 1× PBS. Hemocytes from a single larva werebled into 5 μl 1× PBS in black resin dissection dishes and allowed to settlefor 30 min in a humidifier chamber. Images were taken using a NikonD5100 camera mounted on a Nikon SMZ 1500 dissecting microscope withUV X-cite 120 at 5× magnification. ImageJ was used to measure the GFPintensity in each sample. We determined that the GFP intensity within thecells does not change between genotypes. GFP intensities were normalizedto the control, with the control value set at 1. The relative GFP intensity wasplotted on the y-axis as ‘hemocyte aggregation’. These values were graphedand analyzed using GraphPad Prism 7 and statistical analysis was assessedby two-way ANOVA.

Hemocyte isolation for qPCR and immunoprecipitationLarvae were washed in 1× PBS. Hemocytes were collected into 100 µl 1×PBS droplet/well on dissection plates. These cells were then transferred intoEppendorf tubes and kept on ice. To separate hemocytes from thehemolymph, bleeds were centrifuged at 4°C at 1500 rpm (200 g) for10 min, and the supernatant was discarded. mRNA or protein extraction wasperformed on hemocyte pellets following the procedure below.

mRNA extraction and qPCR from hemocytesmRNA was extracted from hemocyte pellets by homogenizing them intoTRIzol (Thermo Fisher Scientific), followed by chloroform extraction andethanol precipitation. The mRNAs were diluted into H2O. A total of0.5-1 µg mRNA was treated with DNase (Ambion) as per themanufacturer’s instructions. The mRNA was reverse transcribed using aMaxima reverse transcriptase kit (Thermo Fisher Scientific). qPCR wasperformed using a SYBR Green PCR Master Mix kit (Thermo FisherScientific), using primers from Integrated DNA Technologies (describedbelow) in a CFX96 Touch™ Real-Time PCR Detection System, and usingthe Bio-Rad CFXManager 3.1. We used 95°C for denaturing for 10 s, 65°Cfor primer annealing for 30 s, and 72°C for primer extension for 10 s. Thesesteps were repeated 40 times. ΔΔCt for each gene was calculated relativeto the expression of the respective gene from the HaHmlLT>GFP controlcross and normalized to the Rpl15 endogenous control. Normalized relativegene expression was plotted in GraphPad Prism7 and Student’s t-test wasused for statistical analysis per gene for each genotype compared to thecontrol genotype.

Fig. 8. Model of JAK/STAT being regulated by signaling by E(Pc)/Tip60.E(Pc)/Tip60 suppresses lamellocyte differentiation and tumor formation.E(Pc)/Tip60 represses JAK/STAT signaling by inhibiting the protein expressionof Hop through a direct or indirect (dotted lines) mechanism. See text fordetails.

12


Disea

seModels&Mechan

isms

PrimersPrimers for the indicated genes (Table S1) were manufactured by IntegratedDNATechnologies using the following settings: primers must expand exonjunctions, 50% CG, product 100 to 200 bp in length, and meltingtemperature of 65°C.

Immunoprecipitation of hemocytesAfter centrifugation (see mRNA extraction section above), hemocytepellets were lyzed using 100 µl of an NP-40 lysis buffer (10 mM HEPES,10 mM KCl, 1 mM EDTA, 100 µM EGTA, 1 mM NaOV, 10 mMβ-glycerophosphate, 100 mM NaF, 1.05× cOmplete protease inhibitor, 1%NP-40). Cellswere lysed by pipetting up and down, incubated on ice for 7 minand centrifuged at 4°C for 10 min at 10,000 rpm (6149 g). The supernatantwas transferred to a new tube. To isolate sufficient quantities of hemocytes, wedissected larvae each day, flash freezing the hemocyte lysate and storing it at−20°C until enough larvaewere collected to perform an immunoprecipitation(typically 20-55 larvae). Hemocyte lysates were then pooled together andlysate buffer was added to bring the total volume to 500 µl. We ran 4% of thetotal lysate for the input control. A total of 500 µl of lysate buffer was added tothe rest of the lysate and used for immunoprecipitation.

For each sample, we added 20 µl of a 50:50 slurry of Protein A-Sepharosebeads (GE Healthcare, #17-5280-01) in lysis buffer, and 1 µl of mouseanti-V5 monoclonal per sample, and incubated at 4°C overnight withagitation. Samples were then centrifuged at 4°C for 1 min at 7000 rpm(3013 g), and the flow through was discarded. The samples were washedfour times with wash buffer (10 mM HEPES, 10 mM KCl, 1 mM EDTA,100 µM EGTA, 1% NP-40) and centrifuged at 4°C for 1 min at 7000 rpm.After the last wash, all excess supernatant was removed using a 50 µlHamilton syringe. Immunoprecipitated proteins were eluted by adding 20 µlof 2× Laemmli [containing β-mercaptoethanol (βME)] in lysis buffer andcentrifuging for 1 min at 7000 rpm at room temperature. Samples wereboiled for 5 min and centrifuged for 1 min at 7000 rpm at room temperature.The eluate was removed using a Hamilton syringe and transferred to a newtube for SDS-PAGE and western blotting.

S2 cell culture, transfection and KDACi treatmentS2 cells were grown in Schneider media (10% FBS and 1% Pen/Strep). S2cells were transfected using the Qiagen Quick-Start Protocol (Qiagen,#301427). Briefly, 2.4×106 S2 cells were seeded in 10 mm plates with9.6 ml of Schneider media a day prior to transfection. We transfected 12 μgof Act5c-hop-myc-his plasmid (Ekas et al., 2010) using the Qiagen protocol(96 µl Enhancer and 60 µl Effectene in 3.6 ml Schneider media). At 32 hpost-transfection, KDACi were added to the cultures for 16 h of treatment asdescribed (Klampfer et al., 2004). After a total of 48 h of transfection, theexperiment was terminated. S2 cells were collected by centrifuging them ina 15 ml conical tube for 3 min at 1000 g. S2 cells were lysed using 1 ml of1%NP-40 lysis buffer (10 mMHEPES, 10 mMKCl, 1 mMEDTA, 100 µMEGTA, 1 mM NaOV, 10 mM β-glycerophosphate, 100 mM NaF, 1.05×cOmplete protease inhibitor, 1% NP-40) and incubated on ice for 7 min andcentrifuged at 4°C for 10 min at 10,000 rpm. The supernatant wastransferred to a new tube. 5% of the total lysate was saved in samplebuffer with βME (250 mM Tris, pH 6.8, 25% glycerol, 2% SDS, 5% βME)for input control. Input samples were boiled for 5 min and centrifuged at15,000 rpm (17,530 g) for 1 min before freezing or loading them onto anSDS-PAGE gel (see below). The rest of the lysate was used forimmunoprecipitation as described above for hemocytes using 1 µl of9E10 (mouse anti-myc monoclonal) per sample. TSA was a gift from DrDavid Levy (NYU School of Medicine, USA) and NaBut was a gift from DrDanny Reinberg (NYU School of Medicine, USA).

Western blottingWe cast by hand 8%, 12% or 15% standard SDS-PAGE gels. Proteins wereseparated by electrophoresis using 1× SDS-Tris running buffer (0.3% Tris-base, 1.44% glycine, 0.1% SDS, 8.5 pH). Proteins were transferred to0.45 µm nitrocellulose membrane (Bio-Rad, #1620115) by western blottingat 300 mAmp for 2 h at 4°C in 1× transfer buffer (0.3% Tris-base, 1.44%glycine, 20% methanol). Each subsequent step was performed with mildagitation. Membranes were blocked with 5% non-fat dairy milk in 1× TBS-

T [0.242% Tris-base, 0.8% NaCl (Fisher Scientific, #BP358-1), 0.1%Tween (Fisher Scientific, #BP337-500)] for 1-2 h at room temperature orovernight at 4°C. Excess block was removed by quickly washing themembrane in 1× TBS-T. Membranes were blotted with respective primaryantibodies in 1× TBS-T for at least 1 h at room temperature or overnight at4°C. Membranes were then washed three times for 10 min with 1× TBS-T.Secondary antibodies were diluted in 1× TBS-T and incubated for at least1 h at room temperature. Membranes were then washed at least three timesfor 10 min with 1× TBS-T and imaged using a Li-Cor Odyssey scanner withthe Odyssey Infrared Imaging System Application Software, Version 3.0.We used ImageJ to calculate the density of the protein bands on westernblots. We normalized the intensity of the Hop-GFP-V5 band to that of theGFP band. We normalized the intensity of the Hop-Myc-His band to that ofthe actin band for the S2 cell experiments.We plotted the relative intensity ofthe Hop-GFP-V5 bands or the Hop-Myc-His bands using GraphPad Prism7.To determine statistical significance, we used Student’s t-test for the in vivoHop-GFP-V5 experiment and one-way ANOVA for the in vitro Hop-Myc-His experiment.

Male survivalWe quantified the number of adult hopTum/Y males and adult hopTum/+females for all genotypes. Survival was calculated as the number of eclosedhopTum males divided by the total number of eclosed animals bearing thehopTum chromosome.

AcknowledgementsWe thank Dr Chen, Dr Banerjee, Dr Elefant, the Bloomington Drosophila StockCenter and the Vienna Drosophila Resource Center for stocks; and Dr Andó,Dr Ryoo, Dr Levy and Dr Reinberg for antibodies and reagents. We are grateful tomembers of the Bach, Treisman and Ryoo labs for insightful discussions.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsConceptualization: A.A.B., E.A.B.; Methodology: A.A.B., E.A.B.; Validation: A.A.B.,L.J.N.-P., V.D., S.H., S.L., A.B.; Formal analysis: A.A.B., L.J.N.-P., V.D., S.H., S.L.,A.B., E.A.B.; Investigation: A.A.B., L.J.N.-P., V.D., S.H., S.L., A.B., E.A.B.;Resources: E.A.B.; Writing - original draft: A.A.B., E.A.B.; Writing - review & editing:A.A.B., E.A.B.; Supervision: E.A.B.; Project administration: A.A.B., E.A.B.; Fundingacquisition: A.A.B., E.A.B.

FundingThis study was supported by R01GM085075 from the National Institutes of Health (toE.A.B.), a National Science Foundation Graduate Research Fellowship (to A.A.B.)and the Jack Kent Cooke Foundation Continuing Graduate Scholarship (to A.A.B.).

Data availabilityAll data are available upon request.

Supplementary informationSupplementary information available online athttp://dmm.biologists.org/lookup/doi/10.1242/dmm.038679.supplemental

ReferencesAkada, H., Akada, S., Gajra, A., Bair, A., Graziano, S., Hutchison, R. E. and

Mohi, G. (2012). Efficacy of vorinostat in a murine model of polycythemia vera.Blood 119, 3779-3789. doi:10.1182/blood-2011-02-336743

Amoyel, M., Anderson, A. M. and Bach, E. A. (2014). JAK/STAT pathwaydysregulation in tumors: a Drosophila perspective. Semin. Cell Dev. Biol. 28,96-103. doi:10.1016/j.semcdb.2014.03.023

Anderl, I., Vesala, L., Ihalainen, T. O., Vanha-Aho, L.-M., Andó, I., Rämet, M. andHultmark, D. (2016). Transdifferentiation and proliferation in two distincthemocyte lineages in Drosophila melanogaster Larvae after Wasp infection.PLoS Pathog. 12, e1005746. doi:10.1371/journal.ppat.1005746

Anderson, A. M., Bailetti, A. A., Rodkin, E., De, A. and Bach, E. A. (2017). Agenetic screen reveals an unexpected role for Yorkie signaling in JAK/STAT-dependent hematopoietic malignancies in Drosophila melanogaster. G3 7,2427-2438. doi:10.1534/g3.117.044172

Avet-Rochex, A., Boyer, K., Polesello, C., Gobert, V., Osman, D., Roch, F., Auge,B., Zanet, J., Haenlin, M. and Waltzer, L. (2010). An in vivo RNA interferencescreen identifies gene networks controlling Drosophila melanogaster blood cellhomeostasis. BMC Dev. Biol. 10, 65. doi:10.1186/1471-213X-10-65

13


Disea

seModels&Mechan

isms

http://dmm.biologists.org/lookup/doi/10.1242/dmm.038679.supplementalhttp://dmm.biologists.org/lookup/doi/10.1242/dmm.038679.supplementalhttp://dmm.biologists.org/lookup/doi/10.1242/dmm.038679.supplementalhttps://doi.org/10.1182/blood-2011-02-336743https://doi.org/10.1182/blood-2011-02-336743https://doi.org/10.1182/blood-2011-02-336743https://doi.org/10.1016/j.semcdb.2014.03.023https://doi.org/10.1016/j.semcdb.2014.03.023https://doi.org/10.1016/j.semcdb.2014.03.023https://doi.org/10.1371/journal.ppat.1005746https://doi.org/10.1371/journal.ppat.1005746https://doi.org/10.1371/journal.ppat.1005746https://doi.org/10.1371/journal.ppat.1005746https://doi.org/10.1534/g3.117.044172https://doi.org/10.1534/g3.117.044172https://doi.org/10.1534/g3.117.044172https://doi.org/10.1534/g3.117.044172https://doi.org/10.1186/1471-213X-10-65https://doi.org/10.1186/1471-213X-10-65https://doi.org/10.1186/1471-213X-10-65https://doi.org/10.1186/1471-213X-10-65

Bach, E. A., Ekas, L. A., Ayala-Camargo, A., Flaherty, M. S., Lee, H., Perrimon,N. and Baeg, G.-H. (2007). GFP reporters detect the activation of the DrosophilaJAK/STAT pathway in vivo. Gene Expr. Patterns 7, 323-331. doi:10.1016/j.modgep.2006.08.003

Bali, P., Pranpat, M., Bradner, J., Balasis, M., Fiskus, W., Guo, F., Rocha, K.,Kumaraswamy, S., Boyapalle, S., Atadja, P. et al. (2005). Inhibition of histonedeacetylase 6 acetylates and disrupts the chaperone function of heat shockprotein 90: a novel basis for antileukemia activity of histone deacetylase inhibitors.J. Biol. Chem. 280, 26729-26734. doi:10.1074/jbc.C500186200

Banerjee, U., Girard, J. R., Goins, L. M. and Spratford, C. M. (2019). Drosophilaas a genetic model for hematopoiesis. Genetics 211, 367-417. doi:10.1534/genetics.118.300223

Bazzi, W., Cattenoz, P. B., Delaporte, C., Dasari, V., Sakr, R., Yuasa, Y. andGiangrande, A. (2018). Embryonic hematopoiesis modulates the inflammatoryresponse and larval hematopoiesis in Drosophila. eLife 7, e34890. doi:10.7554/eLife.34890

Braun, A., Hoffmann, J. A. andMeister, M. (1998). Analysis of the Drosophila hostdefense in domino mutant larvae, which are devoid of hemocytes. Proc. Natl.Acad. Sci. USA 95, 14337-14342. doi:10.1073/pnas.95.24.14337

Brown, S., Hu, N. and Hombria, J. C.-G. (2001). Identification of the firstinvertebrate interleukin JAK/STAT receptor, the Drosophila gene domeless. Curr.Biol. 11, 1700-1705. doi:10.1016/S0960-9822(01)00524-3

Cerami, E., Gao, J., Dogrusoz, U., Gross, B. E., Sumer, S. O., Aksoy, B. A.,Jacobsen, A., Byrne, C. J., Heuer, M. L., Larsson, E. et al. (2012). The cBiocancer genomics portal: an open platform for exploring multidimensional cancergenomics data. Cancer Discov. 2, 401-404. doi:10.1158/2159-8290.CD-12-0095

Corwin, H. O. and Hanratty, W. P. (1976). Characterization of a unique lethaltumorous mutation in Drosophila. Mol. Gen. Genet. 144, 345-347. doi:10.1007/BF00341734

Ekas, L. A., Cardozo, T. J., Flaherty, M. S., Mcmillan, E. A., Gonsalves, F. C. andBach, E. A. (2010). Characterization of a dominant-active STAT that promotestumorigenesis in Drosophila. Dev. Biol. 344, 621-636. doi:10.1016/j.ydbio.2010.05.497

Feng, L., Shi, Z. and Chen, X. (2017). Enhancer of polycomb coordinates multiplesignaling pathways to promote both cyst and germline stem cell differentiation inthe Drosophila adult testis.PLoSGenet. 13, e1006571. doi:10.1371/journal.pgen.1006571

Feng, L., Shi, Z., Xie, J., Ma, B. and Chen, X. (2018). Enhancer of polycombmaintains germline activity and genome integrity in Drosophila testis. Cell DeathDiffer. 25, 1486-1502. doi:10.1038/s41418-017-0056-5

Flaherty, M. S., Zavadil, J., Ekas, L. A. and Bach, E. A. (2009). Genome-wideexpression profiling in the Drosophila eye reveals unexpected repression of notchsignaling by the JAK/STAT pathway. Dev. Dyn. 238, 2235-2253. doi:10.1002/dvdy.21989

Flaherty, M. S., Salis, P., Evans, C. J., Ekas, L. A., Marouf, A., Zavadil, J.,Banerjee, U. and Bach, E. A. (2010). chinmo is a functional effector of the JAK/STAT pathway that regulates eye development, tumor formation, and stemcell self-renewal in Drosophila. Dev. Cell 18, 556-568. doi:10.1016/j.devcel.2010.02.006

Gao, J., Aksoy, B. A., Dogrusoz, U., Dresdner, G., Gross, B., Sumer, S. O., Sun,Y., Jacobsen, A., Sinha, R., Larsson, E. et al. (2013). Integrative analysis ofcomplex cancer genomics and clinical profiles using the cBioPortal. Sci. Signal. 6,pl1. doi:10.1126/scisignal.2004088

Gold, K. S. and Brückner, K. (2015). Macrophages and cellular immunity inDrosophila melanogaster. Semin. Immunol. 27, 357-368. doi:10.1016/j.smim.2016.03.010

Gorrini, C., Squatrito, M., Luise, C., Syed, N., Perna, D., Wark, L., Martinato, F.,Sardella, D., Verrecchia, A., Bennett, S. et al. (2007). Tip60 is a haplo-insufficient tumour suppressor required for an oncogene-induced DNA damageresponse. Nature 448, 1063-1067. doi:10.1038/nature06055

Grigorian, M., Mandal, L. and Hartenstein, V. (2011). Hematopoiesis at the onsetof metamorphosis: terminal differentiation and dissociation of the Drosophilalymph gland. Dev. Genes Evol. 221, 121-131. doi:10.1007/s00427-011-0364-6

Grmai, L., Hudry, B., Miguel-Aliaga, I. and Bach, E. A. (2018). Chinmo preventstransformer alternative splicing to maintain male sex identity. PLoS Genet. 14,e1007203. doi:10.1371/journal.pgen.1007203

Guerini, V., Barbui, V., Spinelli, O., Salvi, A., Dellacasa, C., Carobbio, A.,Introna, M., Barbui, T., Golay, J. and Rambaldi, A. (2008). The histonedeacetylase inhibitor ITF2357 selectively targets cells bearing mutatedJAK2(V617F). Leukemia 22, 740-747. doi:10.1038/sj.leu.2405049

Hanratty, W. P. and Ryerse, J. S. (1981). A genetic melanotic neoplasm ofDrosophila melanogaster. Dev. Biol. 83, 238-249. doi:10.1016/0012-1606(81)90470-X

Harrison, D. A., Binari, R., Nahreini, T. S., Gilman, M. and Perrimon, N. (1995).Activation of a Drosophila Janus kinase (JAK) causes hematopoietic neoplasiaand developmental defects. EMBO J. 14, 2857-2865. doi:10.1002/j.1460-2075.1995.tb07285.x

Herrera, S. C. and Bach, E. A. (2019). JAK/STAT signaling in stem cells andregeneration: from Drosophila to vertebrates. Development 146, dev167643.doi:10.1242/dev.167643

Holz, A., Bossinger, B., Strasser, T., Janning, W. and Klapper, R. (2003). Thetwo origins of hemocytes in Drosophila. Development 130, 4955-4962. doi:10.1242/dev.00702

Honti, V., Csordás, G., Márkus, R., Kurucz, É., Jankovics, F. andAndó, I. (2010).Cell lineage tracing reveals the plasticity of the hemocyte lineages and of thehematopoietic compartments in Drosophila melanogaster. Mol. Immunol. 47,1997-2004. doi:10.1016/j.molimm.2010.04.017

Honti, V., Csordás, G., Kurucz, É., Márkus, R. and Andó, I. (2014). The cell-mediated immunity of Drosophila melanogaster: hemocyte lineages, immunecompartments, microanatomy and regulation. Dev. Comp. Immunol. 42, 47-56.doi:10.1016/j.dci.2013.06.005

Hou, X. S., Melnick, M. B. and Perrimon, N. (1996). Marelle acts downstream of theDrosophila HOP/JAK kinase and encodes a protein similar to the mammalianSTATs. Cell 84, 411-419. doi:10.1016/S0092-8674(00)81286-6

James, C., Ugo, V., Le Couédic, J.-P., Staerk, J., Delhommeau, F., Lacout, C.,Garçon, L., Raslova, H., Berger, R., Bennaceur-Griscelli, A. et al. (2005). Aunique clonal JAK2 mutation leading to constitutive signalling causespolycythaemia vera. Nature 434, 1144-1148. doi:10.1038/nature03546

Jones, A. V., Kreil, S., Zoi, K., Waghorn, K., Curtis, C., Zhang, L., Score, J.,Seear, R., Chase, A. J., Grand, F. H. et al. (2005). Widespread occurrence of theJAK2 V617F mutation in chronic myeloproliferative disorders. Blood 106,2162-2168. doi:10.1182/blood-2005-03-1320

Jung, S.-H., Evans, C. J., Uemura, C. and Banerjee, U. (2005). The Drosophilalymph gland as a developmental model of hematopoiesis. Development 132,2521-2533. doi:10.1242/dev.01837

Klampfer, L., Huang, J., Swaby, L.-A. and Augenlicht, L. (2004). Requirement ofhistone deacetylase activity for signaling by STAT1. J. Biol. Chem. 279,30358-30368. doi:10.1074/jbc.M401359200

Kralovics, R., Passamonti, F., Buser, A. S., Teo, S.-S., Tiedt, R., Passweg, J. R.,Tichelli, A., Cazzola, M. and Skoda, R. C. (2005). A gain-of-function mutation ofJAK2 in myeloproliferative disorders. N Engl. J. Med. 352, 1779-1790. doi:10.1056/NEJMoa051113

Kurucz, É., Márkus, R., Zsámboki, J., Folkl-Medzihradszky, K., Darula, Z.,Vilmos, P., Udvardy, A., Krausz, I., Lukacsovich, T., Gateff, E. et al. (2007).Nimrod, a putative phagocytosis receptor with EGF repeats in Drosophilaplasmatocytes. Curr. Biol. 17, 649-654. doi:10.1016/j.cub.2007.02.041

Lafave, L. M. and Levine, R. L. (2012). JAK2 the future: therapeutic strategies forJAK-dependent malignancies. Trends Pharmacol. Sci. 33, 574-582. doi:10.1016/j.tips.2012.08.005

Lanot, R., Zachary, D., Holder, F. and Meister, M. (2001). Postembryonichematopoiesis in Drosophila. Dev. Biol. 230, 243-257. doi:10.1006/dbio.2000.0123

Leatherman, J. L. and Dinardo, S. (2008). Zfh-1 controls somatic stem cell self-renewal in the Drosophila testis and nonautonomously influences germline stemcell self-renewal. Cell Stem Cell 3, 44-54. doi:10.1016/j.stem.2008.05.001

Lebestky, T., Chang, T., Hartenstein, V. and Banerjee, U. (2000). Specification ofDrosophila hematopoietic lineage by conserved transcription factors. Science288, 146-149. doi:10.1126/science.288.5463.146

Leitao, A. B. and Sucena, E. (2015). Drosophila sessile hemocyte clusters are truehematopoietic tissues that regulate larval blood cell differentiation. eLife 4,e06166. doi:10.7554/eLife.06166

Letourneau, M., Lapraz, F., Sharma, A., Vanzo, N., Waltzer, L. and Crozatier, M.(2016). Drosophila hematopoiesis under normal conditions and in response toimmune stress. FEBS Lett. 590, 4034-4051. doi:10.1002/1873-3468.12327

Levine, R. L., Wadleigh, M., Cools, J., Ebert, B. L., Wernig, G., Huntly, B. J. P.,Boggon, T. J., Wlodarska, I., Clark, J. J., Moore, S. et al. (2005). Activatingmutation in the tyrosine kinase JAK2 in polycythemia vera, essentialthrombocythemia, and myeloid metaplasia with myelofibrosis. Cancer Cell 7,387-397. doi:10.1016/j.ccr.2005.03.023

Levy, D. E. (1999). Physiological significance of STAT proteins: investigationsthrough gene disruption in vivo. Cell. Mol. Life Sci. 55, 1559-1567. doi:10.1007/s000180050395

Lleonart, M., Vidal, F., Gallardo, D., Diaz-Fuertes, M., Rojo, F., Cuatrecasas, M.,Lopez-Vicente, L., Kondoh, H., Blanco, C., Carnero, A. et al. (2006). New p53related genes in human tumors: significant downregulation in colon and lungcarcinomas. Oncol. Rep. 16, 603-608. doi:10.3892/or.16.3.603

Lorbeck, M., Pirooznia, K., Sarthi, J., Zhu, X. and Elefant, F. (2011). Microarrayanalysis uncovers a role for Tip60 in nervous system function and generalmetabolism. PLoS ONE 6, e18412. doi:10.1371/journal.pone.0018412

Luo, H., Hanratty,W. P. and Dearolf, C. R. (1995). An amino acid substitution in theDrosophila hopTum-l Jak kinase causes leukemia-like hematopoietic defects.EMBO J. 14, 1412-1420. doi:10.1002/j.1460-2075.1995.tb07127.x

Luo, H., Rose, P., Barber, D., Hanratty, W. P., Lee, S., Roberts, T. M., D’andrea,A. D. and Dearolf, C. R. (1997). Mutation in the Jak kinase JH2 domainhyperactivates Drosophila and mammalian Jak-Stat pathways.Mol. Cell. Biol. 17,1562-1571. doi:10.1128/MCB.17.3.1562

Makhijani, K., Alexander, B., Tanaka, T., Rulifson, E. and Brückner, K. (2011).The peripheral nervous system supports blood cell homing and survival in theDrosophila larva. Development 138, 5379-5391. doi:10.1242/dev.067322

14


Disea

seModels&Mechan

isms

https://doi.org/10.1016/j.modgep.2006.08.003https://doi.org/10.1016/j.modgep.2006.08.003https://doi.org/10.1016/j.modgep.2006.08.003https://doi.org/10.1016/j.modgep.2006.08.003https://doi.org/10.1074/jbc.C500186200https://doi.org/10.1074/jbc.C500186200https://doi.org/10.1074/jbc.C500186200https://doi.org/10.1074/jbc.C500186200https://doi.org/10.1074/jbc.C500186200https://doi.org/10.1534/genetics.118.300223https://doi.org/10.1534/genetics.118.300223https://doi.org/10.1534/genetics.118.300223https://doi.org/10.7554/eLife.34890https://doi.org/10.7554/eLife.34890https://doi.org/10.7554/eLife.34890https://doi.org/10.7554/eLife.34890https://doi.org/10.1073/pnas.95.24.14337https://doi.org/10.1073/pnas.95.24.14337https://doi.org/10.1073/pnas.95.24.14337https://doi.org/10.1016/S0960-9822(01)00524-3https://doi.org/10.1016/S0960-9822(01)00524-3https://doi.org/10.1016/S0960-9822(01)00524-3https://doi.org/10.1158/2159-8290.CD-12-0095https://doi.org/10.1158/2159-8290.CD-12-0095https://doi.org/10.1158/2159-8290.CD-12-0095https://doi.org/10.1158/2159-8290.CD-12-0095https://doi.org/10.1007/BF00341734https://doi.org/10.1007/BF00341734https://doi.org/10.1007/BF00341734https://doi.org/10.1016/j.ydbio.2010.05.497https://doi.org/10.1016/j.ydbio.2010.05.497https://doi.org/10.1016/j.ydbio.2010.05.497https://doi.org/10.1016/j.ydbio.2010.05.497https://doi.org/10.1371/journal.pgen.1006571https://doi.org/10.1371/journal.pgen.1006571https://doi.org/10.1371/journal.pgen.1006571https://doi.org/10.1371/journal.pgen.1006571https://doi.org/10.1038/s41418-017-0056-5https://doi.org/10.1038/s41418-017-0056-5https://doi.org/10.1038/s41418-017-0056-5https://doi.org/10.1002/dvdy.21989https://doi.org/10.1002/dvdy.21989https://doi.org/10.1002/dvdy.21989https://doi.org/10.1002/dvdy.21989https://doi.org/10.1016/j.devcel.2010.02.006https://doi.org/10.1016/j.devcel.2010.02.006https://doi.org/10.1016/j.devcel.2010.02.006https://doi.org/10.1016/j.devcel.2010.02.006https://doi.org/10.1016/j.devcel.2010.02.006https://doi.org/10.1126/scisignal.2004088https://doi.org/10.1126/scisignal.2004088https://doi.org/10.1126/scisignal.2004088https://doi.org/10.1126/scisignal.2004088https://doi.org/10.1016/j.smim.2016.03.010https://doi.org/10.1016/j.smim.2016.03.010https://doi.org/10.1016/j.smim.2016.03.010https://doi.org/10.1038/nature06055https://doi.org/10.1038/nature06055https://doi.org/10.1038/nature06055https://doi.org/10.1038/nature06055https://doi.org/10.1007/s00427-011-0364-6https://doi.org/10.1007/s00427-011-0364-6https://doi.org/10.1007/s00427-011-0364-6https://doi.org/10.1371/journal.pgen.1007203https://doi.org/10.1371/journal.pgen.1007203https://doi.org/10.1371/journal.pgen.1007203https://doi.org/10.1038/sj.leu.2405049https://doi.org/10.1038/sj.leu.2405049https://doi.org/10.1038/sj.leu.2405049https://doi.org/10.1038/sj.leu.2405049https://doi.org/10.1016/0012-1606(81)90470-Xhttps://doi.org/10.1016/0012-1606(81)90470-Xhttps://doi.org/10.1016/0012-1606(81)90470-Xhttps://doi.org/10.1002/j.1460-2075.1995.tb07285.xhttps://doi.org/10.1002/j.1460-2075.1995.tb07285.xhttps://doi.org/10.1002/j.1460-2075.1995.tb07285.xhttps://doi.org/10.1002/j.1460-2075.1995.tb07285.xhttps://doi.org/10.1242/dev.167643https://doi.org/10.1242/dev.167643https://doi.org/10.1242/dev.167643https://doi.org/10.1242/dev.00702https://doi.org/10.1242/dev.00702https://doi.org/10.1242/dev.00702https://doi.org/10.1016/j.molimm.2010.04.017https://doi.org/10.1016/j.molimm.2010.04.017https://doi.org/10.1016/j.molimm.2010.04.017https://doi.org/10.1016/j.molimm.2010.04.017https://doi.org/10.1016/j.dci.2013.06.005https://doi.org/10.1016/j.dci.2013.06.005https://doi.org/10.1016/j.dci.2013.06.005https://doi.org/10.1016/j.dci.2013.06.005https://doi.org/10.1016/S0092-8674(00)81286-6https://doi.org/10.1016/S0092-8674(00)81286-6https://doi.org/10.1016/S0092-8674(00)81286-6https://doi.org/10.1038/nature03546https://doi.org/10.1038/nature03546https://doi.org/10.1038/nature03546https://doi.org/10.1038/nature03546https://doi.org/10.1182/blood-2005-03-1320https://doi.org/10.1182/blood-2005-03-1320https://doi.org/10.1182/blood-2005-03-1320https://doi.org/10.1182/blood-2005-03-1320https://doi.org/10.1242/dev.01837https://doi.org/10.1242/dev.01837https://doi.org/10.1242/dev.01837https://doi.org/10.1074/jbc.M401359200https://doi.org/10.1074/jbc.M401359200https://doi.org/10.1074/jbc.M401359200https://doi.org/10.1056/NEJMoa051113https://doi.org/10.1056/NEJMoa051113https://doi.org/10.1056/NEJMoa051113https://doi.org/10.1056/NEJMoa051113https://doi.org/10.1016/j.cub.2007.02.041https://doi.org/10.1016/j.cub.2007.02.041https://doi.org/10.1016/j.cub.2007.02.041https://doi.org/10.1016/j.cub.2007.02.041https://doi.org/10.1016/j.tips.2012.08.005https://doi.org/10.1016/j.tips.2012.08.005https://doi.org/10.1016/j.tips.2012.08.005https://doi.org/10.1006/dbio.2000.0123https://doi.org/10.1006/dbio.2000.0123https://doi.org/10.1006/dbio.2000.0123https://doi.org/10.1016/j.stem.2008.05.001https://doi.org/10.1016/j.stem.2008.05.001https://doi.org/10.1016/j.stem.2008.05.001https://doi.org/10.1126/science.288.5463.146https://doi.org/10.1126/science.288.5463.146https://doi.org/10.1126/science.288.5463.146https://doi.org/10.7554/eLife.06166https://doi.org/10.7554/eLife.06166https://doi.org/10.7554/eLife.06166https://doi.org/10.1002/1873-3468.12327https://doi.org/10.1002/1873-3468.12327https://doi.org/10.1002/1873-3468.12327https://doi.org/10.1016/j.ccr.2005.03.023https://doi.org/10.1016/j.ccr.2005.03.023https://doi.org/10.1016/j.ccr.2005.03.023https://doi.org/10.1016/j.ccr.2005.03.023https://doi.org/10.1016/j.ccr.2005.03.023https://doi.org/10.1007/s000180050395https://doi.org/10.1007/s000180050395https://doi.org/10.1007/s000180050395https://doi.org/10.3892/or.16.3.603https://doi.org/10.3892/or.16.3.603https://doi.org/10.3892/or.16.3.603https://doi.org/10.3892/or.16.3.603https://doi.org/10.1371/journal.pone.0018412https://doi.org/10.1371/journal.pone.0018412https://doi.org/10.1371/journal.pone.0018412https://doi.org/10.1002/j.1460-2075.1995.tb07127.xhttps://doi.org/10.1002/j.1460-2075.1995.tb07127.xhttps://doi.org/10.1002/j.1460-2075.1995.tb07127.xhttps://doi.org/10.1128/MCB.17.3.1562https://doi.org/10.1128/MCB.17.3.1562https://doi.org/10.1128/MCB.17.3.1562https://doi.org/10.1128/MCB.17.3.1562https://doi.org/10.1242/dev.067322https://doi.org/10.1242/dev.067322https://doi.org/10.1242/dev.067322

Mandal, L., Banerjee, U. and Hartenstein, V. (2004). Evidence for a fruit flyhemangioblast and similarities between lymph-gland hematopoiesis in fruit fly andmammal aorta-gonadal-mesonephros mesoderm. Nat. Genet. 36, 1019-1023.doi:10.1038/ng1404

Markus, R., Laurinyecz, B., Kurucz, E., Honti, V., Bajusz, I., Sipos, B., Somogyi,K., Kronhamn, J., Hultmark, D. and Ando, I. (2009). Sessile hemocytes as ahematopoietic compartment in Drosophila melanogaster. Proc. Natl. Acad. Sci.USA 106, 4805-4809. doi:10.1073/pnas.0801766106

Marubayashi, S., Koppikar, P., Taldone, T., Abdel-Wahab, O., West, N., Bhagwat,N., Caldas-Lopes, E., Ross, K. N., Gönen, M., Gozman, A. et al. (2010). HSP90is a therapeutic target in JAK2-dependentmyeloproliferative neoplasms inmice andhumans. J. Clin. Invest. 120, 3578-3593. doi:10.1172/JCI42442

Mccarthy, A., Deiulio, A., Martin, E. T., Upadhyay, M. and Rangan, P. (2018).Tip60 complex promotes expression of a differentiation factor to regulate germlinedifferentiation in female Drosophila. Mol. Biol. Cell 29, 2933-2945. doi:10.1091/mbc.E18-06-0385

Minakhina, S. and Steward, R. (2006). Melanotic mutants in Drosophila: pathwaysand phenotypes. Genetics 174, 253-263. doi:10.1534/genetics.106.061978

Mondal, B. C., Shim, J., Evans, C. J. and Banerjee, U. (2014). Pvr expressionregulators in equilibrium signal control and maintenance of Drosophila bloodprogenitors. eLife 3, e03626. doi:10.7554/eLife.03626

O’shea, J. J., Gadina, M. and Schreiber, R. D. (2002). Cytokine signaling in 2002:new surprises in the Jak/Stat pathway. Cell 109 Suppl., S121-S131. doi:10.1016/S0092-8674(02)00701-8

Owusu-Ansah, E. and Banerjee, U. (2009). Reactive oxygen species primeDrosophila haematopoietic progenitors for differentiation. Nature 461, 537-541.doi:10.1038/nature08313

Petraki, S., Alexander, B. and Brückner, K. (2015). Assaying blood cell populationsof the Drosophila melanogaster Larva. J. Vis. Exp., e52733. doi:10.3791/52733

Rizki, T. M. (1978). The circulatory system and associated cells and tissues. InThe Genetics and Biology of Drosophila (ed. M. Ashburner and T. R. F. Wright).pp. 397-452. New York, London: Academic Press.

Rust, K., Tiwari, M. D., Mishra, V. K., Grawe, F. and Wodarz, A. (2018). Myc andthe Tip60 chromatin remodeling complex control neuroblast maintenance andpolarity in Drosophila. EMBO J. 37, e98659. doi:10.15252/embj.201798659

Sachs, Z., Been, R. A., Decoursin, K. J., Nguyen, H. T., Mohd Hassan, N. A.,Noble-Orcutt, K. E., Eckfeldt, C. E., Pomeroy, E. J., Diaz-Flores, E., Geurts,J. L. et al. (2016). Stat5 is critical for the development and maintenance ofmyeloproliferative neoplasm initiated by Nf1 deficiency. Haematologica 101,1190-1199. doi:10.3324/haematol.2015.136002

Sarov, M., Barz, C., Jambor, H., Hein, M. Y., Schmied, C., Suchold, D., Stender,B., Janosch, S., KJ, V. V., Krishnan, R. T. et al. (2016). A genome-wide resourcefor the analysis of protein localisation in Drosophila. eLife 5, e12068. doi:10.7554/eLife.12068

Searle, N. E., Torres-Machorro, A. L. and Pillus, L. (2017). Chromatin regulationby the NuA4 acetyltransferase complex is mediated by essential interactionsbetween Enhancer of Polycomb (Epl1) and Esa1. Genetics 205, 1125-1137.doi:10.1534/genetics.116.197830

Sheikh, B. N. andAkhtar, A. (2019). Themany lives of KATs - detectors, integratorsand modulators of the cellular environment. Nat. Rev. Genet. 20, 7-23. doi:10.1038/s41576-018-0072-4

Silvers, M. and Hanratty, W. P. (1984). Alterations in the production of hemocytesdue to a neoplastic mutation of Drosophila melanogaster. J. Invertebr. Pathol. 44,324-328. doi:10.1016/0022-2011(84)90030-2

Sorrentino, R. P., Tokusumi, T. and Schulz, R. A. (2007). The Friend of GATAprotein U-shaped functions as a hematopoietic tumor suppressor in Drosophila.Dev. Biol. 311, 311-323. doi:10.1016/j.ydbio.2007.08.011

Stofanko, M., Kwon, S. Y. and Badenhorst, P. (2010). Lineage tracing oflamellocytes demonstrates Drosophila macrophage plasticity. PLoS ONE 5,e14051. doi:10.1371/journal.pone.0014051

Tefferi, A. (2016). Myeloproliferative neoplasms: a decade of discoveries andtreatment advances. Am. J. Hematol. 91, 50-58. doi:10.1002/ajh.24221

Tepass, U., Fessler, L. I., Aziz, A. and Hartenstein, V. (1994). Embryonic origin ofhemocytes and their relationship to cell death in Drosophila. Development 120,1829-1837.

Terriente-Félix, A., Pérez, L., Bray, S. J., Nebreda, A. R. and Milán, M. (2017). ADrosophila model of myeloproliferative neoplasm reveals a feed-forward loop inthe JAK pathway mediated by p38 MAPK signalling. Dis. Model. Mech. 10,399-407. doi:10.1242/dmm.028118

Walz, C., Ahmed, W., Lazarides, K., Betancur, M., Patel, N., Hennighausen, L.,Zaleskas, V. M. and van Etten, R. A. (2012). Essential role for Stat5a/b inmyeloproliferative neoplasms induced by BCR-ABL1 and JAK2(V617F) in mice.Blood 119, 3550-3560. doi:10.1182/blood-2011-12-397554

Wang, Y., Fiskus, W., Chong, D. G., Buckley, K. M., Natarajan, K., Rao, R.,Joshi, A., Balusu, R., Koul, S., Chen, J. et al. (2009). Cotreatment withpanobinostat and JAK2 inhibitor TG101209 attenuates JAK2V617F levels andsignaling and exerts synergistic cytotoxic effects against humanmyeloproliferative neoplastic cells. Blood 114, 5024-5033. doi:10.1182/blood-2009-05-222133

Wood, W. and Jacinto, A. (2007). Drosophila melanogaster embryonichaemocytes: masters of multitasking. Nat. Rev. Mol. Cell Biol. 8, 542-551.doi:10.1038/nrm2202

Xu, P., Li, C., Chen, Z., Jiang, S., Fan, S., Wang, J., Dai, J., Zhu, P. and Chen, Z.(2016). The NuA4 core complex acetylates nucleosomal histone H4 through adouble recognition mechanism.Mol. Cell 63, 965-975. doi:10.1016/j.molcel.2016.07.024

Yan, R., Small, S., Desplan, C., Dearolf, C. R. and Darnell, J. E. Jr. (1996).Identification of a Stat gene that functions in Drosophila development. Cell 84,421-430. doi:10.1016/S0092-8674(00)81287-8

Yan, D., Hutchison, R. E. and Mohi, G. (2012). Critical requirement for Stat5 in amouse model of polycythemia vera. Blood 119, 3539-3549. doi:10.1182/blood-2011-03-345215

Yang, H. and Hultmark, D. (2016). Tissue communication

enhancer of polycomb and the tip60 complex repress hematological tumor initiation … · research...

Documents