The JAK/STAT Pathway Regulates Proximo- Distal Patterning in Drosophila

DEVELOPMENTAL DYNAMICS 236:2721 2730, 2007 RESEARCH ARTICLE The JAK/STAT Pathway Regulates Proximo- Distal Patterning in Drosophila Aidee Ayala-Camargo, 1 Laura A. Ekas, 1 Maria Sol Flaherty, 1 Gyeong-Hun Baeg, 2 and Erika A. Bach 1 * JAK/STAT signaling is thought to control growth and proliferation. However, here we show a novel role for this pathway in the patterning of Drosophila appendages. Loss of Stat92E function results mainly in ventralizations and multiplications of the proximo-distal axis in leg and antenna, primarily through the ectopic misexpression of wingless. We also show that the pathway ligand Unpaired is expressed in two domains in leg and antenna that abuts those of wingless and decapentaplegic. We report that JAK/STAT signaling represses both wingless and decapentaplegic, restricting them to their respective domains in leg and antenna. In a reciprocal manner, we show that wingless and decapentaplegic restrict unpaired to its two domains. Thus, a main function of the JAK/STAT pathway in leg and antennal development is to promote the formation of a single proximo-distal axis per disc by constraining the intersection of wingless and decapentaplegic to the center of the disc. Developmental Dynamics 236:2721 2730, 2007. 2007 Wiley-Liss, Inc. Key words: proximo-distal axis; Unpaired; JAK/STAT; leg; antenna; appendage; Drosophila; Wingless; Decapentaplegic Accepted 14 May 2007 INTRODUCTION In mammals and Drosophila, the Janus Kinase/Signal Transducer and Activator of Transcription (JAK/STAT) pathway is a critical regulator of proliferation and has been shown to promote the formation of tumors (Levy and Darnell, 2002; Arbouzova and Zeidler, 2006). For instance, an oncogenic allele of jak2 causes hematopoietic cancers and a dominant active stat3 allele causes tumors in nude mice (Bromberg et al., 1999; Bowman et al., 2000; Darnell, 2005). Similarly, hyper-activation of the Drosophila JAK/STAT pathway leads to dramatic tissue overgrowths that resemble human tumors (Harrison et al., 1995; Kiger et al., 2001; Tulina and Matunis, 2001; Bach et al., 2003). However, little is known about the role of this pathway in pattern formation. The presence of only one JAK and one STAT in Drosophila, compared to four JAKs and seven STATs in mammals, make the fruit fly an ideal system to study the role of this pathway during development. Three Unpaired ligands (Upd, Upd2, and Upd3) are thought to bind to the receptor Domeless (Dome), resulting in its activation and subsequent trans-phosphorylation and activation of the associated JAK kinase Hopscotch (Hop) (reviewed in Arbouzova and Zeidler, 2006). The activated Hop proteins phosphorylate and activate the transcription factor Stat92E, which then migrates to the nucleus and acts as a transcription factor. Development of the proximo-distal (P-D) axis is essential to the formation of mammalian limbs and arthropod appendages. Significant insights into P-D axis formation have been made in Drosophila imaginal discs, groups of larval epithelia that give rise to adult structures (Cohen, 1993). The Drosophila leg and antenna are homologous structures based on homeotic mutations, and both have been used successfully to study patterning of the P-D axis (Postlethwait and Schneider- The Supplementary Material referred to in this article can be viewed at www.interscience.wiley.com/jpages/1058-8388/suppmat 1 Pharmacology Department, New York University School of Medicine, New York, New York 2 Children s Cancer Research Laboratory, Pediatrics-Hematology/Oncology, New York Medical College, Valhalla, New York Grant Sponsor: Breast Cancer Alliance, Inc.; Grant Sponsor: Pharmacological Sciences Training Grant (N.I.H.); Grant number: GM066704-01. *Correspondence to: Erika A. Bach, Pharmacology Department, New York University School of Medicine, 550 First Avenue, MSB 497B, New York, NY 10016-6402. E-mail: erika.bach@med.nyu.edu DOI 10.1002/dvdy.21230 Published online 11 July 2007 in Wiley InterScience (www.interscience.wiley.com). 2007 Wiley-Liss, Inc.

2722 AYALA-CAMARGO ET AL. man, 1971; Lecuit and Cohen, 1997). The leg consists of five segments: the proximal coxa, followed by the trochanter, femur, tibia, and tarsus, and the distal-most structure, the claw. The adult antenna consists of six segments: from the proximal a1 to the distal a6, also called the arista. Like the vertebrate limb, the Drosophila leg and antenna arise from the differentiation of three different axes: antero-posterior (A-P), dorso-ventral Fig. 1. The JAK/STAT pathway is active in young leg and antennal discs. Unless noted otherwise, discs in all figures are oriented with anterior to the left and dorsal up. A D: upd-gal4, UAS-GFP (upd GFP) expression in second (A,C) and third instar (B,D) leg and antenna discs. The expression of upd GFP (green) abuts the Wg domain (red) in both leg (A,B) and antennal (C,D) discs. E H: Expression of 10XSTAT92E- GFP (abbreviated 10XSTAT-GFP ) (green) and Wg protein (red) in second (E), early third (G), and mid-third (F,H) instar leg and antennal discs, respectively. The 10XSTAT92E-GFP reporter is expressed throughout the leg, displaying lower levels on the ventral side where Wg is expressed (E,F). This reporter is also expressed in the distal antenna and exhibits lower levels in the dorsal, Wg-expression domain (G,H). Scale bar relative size. Fig. 2.

JAK/STAT PATHWAY IN DROSOPHILA 2723 (D-V), and P-D. The secreted factor Hedgehog (Hh) is produced by cells in the posterior compartment of the leg and induces the expression of two secreted factors, the Bone Morphogenetic Protein (BMP) Decapentaplegic (Dpp) and the Wnt protein Wingless (Wg), in anterior cells adjacent to the compartment boundary (Basler and Struhl, 1994; Tabata and Kornberg, 1994). The mutually antagonistic relationship between Dpp and Wg establishes the D-V axis by creating spatially restricted domains, in which each gene maintains its own expression pattern (Brook and Cohen, 1996; Jiang and Struhl, 1996; Johnston and Schubiger, 1996; Penton and Hoffmann, 1996; Theisen et al., 1996; Heslip et al., 1997). In addition to their roles in patterning the D-V axis, Dpp and Wg determine the P-D axis of the leg and antenna. The juxtaposition of dpp- and wg-expressing cells in the central region of both discs activates Distal-less (Dll), which encodes a homeodomain protein that is required for the formation of distal leg and antennal structures (Cohen et al., 1989; Diaz- Benjumea et al., 1994; Campbell and Tomlinson, 1995). Ectopic expression of Fig. 2. Loss of JAK/STAT signaling results in patterning defects of the P-D axis in adult leg and antenna. A F: Leg phenotypes, anterior is to the left. Legs and antennae from y hs-flp; FRT 82B CD2 y flies were labeled control. A control female mesothoracic leg (A). B,C: A control male prothoracic tibia exhibits rows of ventral transverse bristles (B). In y stat92e clones (labeled stat92e ), there are ectopic transverse rows of bristles (yellow arrowhead) immediately adjacent to the normal ones (white arrowhead) (C). D: y stat92e clones result in a supernumerary limb that arose from the femur of a female metathoracic leg and that extends proximo-distally to a claw (arrowhead). The supernumerary limb is shown at high magnification (D, inset). E,F: There are typically 10 sex combs in a control male prothoracic tarsus (E). In y stat92e clones, there are multiple additional y sex combs (F). G K: Antennal phenotypes. Control antenna (G). y stat92e clones lead to an antenna with two a3 a6 segments (arrowheads) (H). I, J: Unmarked dome 468 (I) or hop C111 (J) clones result in an antenna that has triplicated a6 segments (I, arrowheads) or duplicated a3 segments (J, arrowheads). K: y stat92e clones in a Minute background result in an antenna that is missing a3 a6 (red arrowhead) and that has a leg-like structure (black arrowhead). L: These defects are fully rescued by expression of a stat92e transgene (Stat92E FL ) (Ekas et al., 2006) in stat92e clones using the MARCM technique (Lee and Luo, 1999) (D). Trochanter (Tr). Bar indicates relative size (A,D). wg on the dorsal side, or dpp on the ventral side, of the leg disc results in ectopic expression of Dll and a secondary P-D axis (Campbell et al., 1993; Struhl and Basler, 1993; Diaz-Benjumea et al., 1994). Lower levels of Wg and Dpp initiate the transcription of dachshund (dac), which encodes a nuclear factor required for development of the femur and tibia (Mardon et al., 1994; Lecuit and Cohen, 1997). In addition to Dll and Dac, the TALE class homeodomain protein Homothorax (Hth) is essential for patterning of the proximal structures in the leg and antenna (Casares and Mann, 2001; Dong et al., 2001). In this study, we address the role of the JAK/STAT pathway in patterning by addressing its role in P-D axis formation. Specifically, we find that the JAK/ STAT pathway can repress both wg and dpp in an autonomous manner, although Stat92E represses wg more robustly than it does dpp. Furthermore, we show that both wg and dpp restrict upd to its anterior and posterior domains in leg and antennal discs. Through these reciprocal inhibitory interactions, the expression domains of upd, wg, and dpp become mutually exclusive. Thus, we show that JAK/STAT signaling regulates the development of one central P-D axis per leg or antennal disc by constraining the expression domains of dpp and wg. RESULTS Upd Expression and JAK/ STAT Pathway Activity Are Detected Throughout Leg and Antennal Imaginal Disc Development We recently reported that upd is expressed in two distinct domains in leg and antennal discs, one in the anterior and the other in the posterior compartment (Bach et al., 2007). In both discs, upd-expressing cells abut the Wg and Dpp domains (Fig. 1A D; see also Fig. 6G). An upd2 null mutant is viable and fertile and has no antennal or leg defects, while upd3 is not expressed in either disc (data not shown). These data suggest that these factors do not act in these discs or have redundant roles with Upd. To assess the range of Upd activity, we used a 10XSTAT92E-GFP reporter that specifically reflects JAK/STAT pathway activity in vivo (Bach et al., 2007). This reporter is activated throughout larval development in leg and antennal discs (Bach et al., 2007). Interestingly, its expression is lowest in the Wg-expressing domains (Fig. 1E H). These data indicate that in both discs the JAK/STAT pathway is activated by Upd, which acts over a long range. Additionally, the mutual spatial and temporal exclusion of cells with JAK/STAT activity from those that express Wg suggests that a negative regulatory interaction exists between these pathways. Loss of JAK/STAT Pathway Activity Leads to Defective Patterning of the P-D Axis To assess the role of stat92e in the developing leg and antenna, stat92e clones were randomly induced and marked by the absence of the yellow (y) gene product in the adult. In the leg, y stat92e clones are associated with supernumerary appendages and incomplete duplications (Fig. 2A,D, arrowhead and inset, and data not shown). Furthermore, stat92e clones can promote the formation of ventro-lateral structures like ectopic transverse bristles of the prothoracic legs next to the normally positioned ventral ones (Fig. 2B,C yellow arrowhead). In addition, other ventro-lateral structures, such as the sex combs in the male prothoracic leg, are increased in number when they reside within y stat92e clones (Fig. 2E,F). The formation of supernumerary limbs and of both ectopic transverse bristles and sex combs has been observed when ectopic wg expression is present on the dorsal side of the leg disc (Campbell et al., 1993; Struhl and Basler, 1993). In the antenna, y stat92e clones exhibit supernumerary structures, including the duplication or triplication of segments a2-a6 (Fig. 2G,H arrowheads and data not shown). Other phenotypes, such as reduced or absent medial and distal segments (a3 a6), as well as antenna-to-leg transformations, are also observed (Fig. 2K, arrowheads). These defects are fully rescued by co-expression of a full-length stat92e transgene (Ekas et al., 2006), indicating that they were specifically due to the loss of JAK/ STAT signaling (Fig. 2L). Lastly, we de-

2724 AYALA-CAMARGO ET AL. termined that mutant clones of dome or hop, genes that act upstream of stat92e, result in antennal phenotypes similar to those observed in stat92e clones, including the formation of supernumerary segments (Fig. 2I,J, arrowheads). Taken together, these data indicate that the JAK/STAT pathway regulates P-D patterning in both the leg and antenna. Stat92E Is Required for the Proper Expression of Major P-D Patterning Genes To determine if the stat92e adult phenotypes arise during larval development, we examined the effect of loss of stat92e activity in leg and antennal imaginal discs. First, we examined the expression of major P-D patterning genes. In the leg, Dll and Dac are expressed in a central and medial domain, respectively, while Hth is expressed in the proximal domain (Fig. 3A; See Supplemental Fig. 1A, which can be viewed at www.interscience. wiley.com/jpages/1058-8388/suppmat). Third instar leg discs carrying stat92e clones often contain at least one secondary P-D axis (Fig. 3B, arrowhead). These ectopic P-D axes are only found on the dorsal side of the leg disc but can form in either the anterior or posterior compartment (Fig. 3B,D, arrowheads). Bric-a-brac (Bab), a BTB/POZ-domain protein required for the development of distal joints in the leg and antenna (Godt et al., 1993; Couderc et al., 2002), is also observed in these ectopic P-D axes (Fig. 3C,D, arrowhead). Consistent with a lack of defects in the coxa and trochanter in adult legs containing stat92e clones, we found that Hth expression is not altered in stat92e clones in the leg disc (Supplemental Fig. 1A,B). Similarly, additional P-D axes are observed in antennal discs containing stat92e clones. In the third instar antenna, Hth is expressed in a1, a2, and faintly in a3, while Dll is expressed in a2 a6 (Supplemental Fig. 1C) (Dong et al., 2000). We show one disc that contains three P-D axes, as assessed by Dll expression (Fig. 3E, asterisks). One of these ectopic axes occurs within the stat92e clone, while the bifurcation that generated the other two axes follows the stat92e clone boundary (Fig. 3E,E ). We presume that comparable antennal discs would give rise to the duplicated or triplicated antennal structures similar to those shown in Figure 2H J. In addition, we also observed absent or reduced Dll expression in large stat92e clones, and a concomitant expansion in Hth expression throughout the proximal and distal antenna (Supplemental Fig. 1C,D). These results are consistent with the adult phenotypes where a3 a6 segments are most frequently lost (Fig. 2K). To assess whether Dll and Hth can be directly regulated by JAK/STAT signaling, we used the flip-out technique to make random hop-expressing clones, which result in ligand-independent, autonomous activation of Stat92E (Struhl and Basler, 1993; Ekas et al., 2006). In hop-expressing clones, the expression patterns of Dll and Hth are not altered, arguing that this pathway does not directly affect the expression of these factors (Fig. 3F,F ; Supplemental Fig. 1E,E ). JAK/STAT Signaling Autonomously Represses wg Expression The defects observed in stat92e clones in leg and antennal discs and their corresponding adult structures suggested that wg was ectopically expressed. In wild type leg discs, Wg protein is expressed in a wedge in ventral anterior cells close to the A-P boundary, while dpp, as assessed by expression of the dpp-lacz reporter (Blackman et al., 1991), is synthesized by dorsal anterior cells adjacent to this boundary and at lower levels by ventral anterior cells (Fig. 4A). In wild type antennal discs, Wg and dpp domains are reversed relative to the leg (Fig. 4C). As expected, ectopic Wg is expressed within stat92e clones on the dorsal side of the leg disc. Moreover, we observe a secondary P-D axis adjacent to the ectopic Wg domain, as assessed by ectopic non-autonomous dpp expression and local outgrowths (Fig. 4B, arrowhead). Similarly, ectopic Wg is also seen in stat92e antennal clones, as is an additional P-D axis, and non-autonomous ectopic expression of dpp (Fig. 4D, arrowhead). Furthermore, the ectopic Wg protein observed in stat92e clones reflects autonomous ectopic transcription of the wg gene, as monitored by the enhancer trap wg P (Kassis et al., 1992) (Fig. 4E,E arrowhead). We note that the lateral posterior domain in the antennal disc appears to be the most sensitive to loss of JAK/STAT signaling, as ectopic wg is always observed in stat92e clones in this region (Fig. 4E). We next attempted to identify the region of the wg gene regulated by Stat92E. The 3 cis genomic region of the wg gene regulates wg expression in many imaginal discs (Pereira et al., 2006). These authors found that a 2-kb enhancer called wg2.3z was the smallest enhancer from this 3 cis genomic region that could partially recapitulate wg expression in leg and other ventral discs (Supplemental Fig. 2A). We were unable to detect wg2.3z expression in leg discs by immuno-fluorescence using antibodies specific for -galactosidase (data not shown). However, we could detect its expression in wild type discs by X-Gal staining (Supplemental Fig. 2B,D). This enabled us to ask whether wg2.3z is ectopically expressed in discs containing stat92e clones. Indeed, this reporter is ectopically expressed in these discs and in regions where we observe re-patterning and overgrowth (Supplemental Fig. 2C,E, yellow arrowheads). However, we were unable to determine if Stat92E regulates this wg enhancer autonomously. The data presented to this point suggest that Stat92E represses wg. Therefore, we reasoned that ectopic activation of the JAK/STAT pathway should lead to wg repression. To address this issue, we assessed whether hop-expressing clones could autonomously repress expression of wg, as monitored by the wg P enhancer trap. As expected, in both leg and antennal discs, ectopic activation of JAK/STAT signaling represses wg in an autonomous manner (Fig. 5A,A and data not shown). These results indicate that JAK/STAT signaling represses wg. JAK/STAT Signaling Also Represses dpp Expression We also find that dpp can be autonomously expressed in stat92e clones (4/24 discs), although at a lower penetrance than ectopic expression of wg. We next specifically examined leg

JAK/STAT PATHWAY IN DROSOPHILA 2725 Fig. 3. Expression of proximal and distal markers in stat92e clones in leg and antennal discs. stat92e clones were generated in a Minute (B, D) or non-minute (E) background and lack GFP (green). All discs are third instar. In a wild type leg disc, Dac (blue) is expressed in the femur, tibia, and first tarsal segment, while Dll (red) is expressed in all tarsal segments and distal tibia (A). In a leg disc that contains large stat92e clones, a secondary P-D axis is observed in the posterior compartment (B, arrowhead). In a wild type leg disc, Bab (red) is expressed in 4 concentric rings within the Dll domain (C). A stat92e clone in the anterior compartment of the leg disc is associated with a secondary P-D axis, in which Bab is ectopically expressed (D, arrowhead). A stat92e clone that runs through the D-V axis in an antennal disc is associated with a triplicated P-D axis, as assessed by Dll (red) staining (E, asterisks). The expression of Hth follows the edge of the clone ventro-laterally and marks the proximal edge of the ectopic axes (E ). hop-expressing clones (abbreviated hop FO and marked by GFP) do not alter the expression of Dll (red) (F,F ). D, dorsal; V, ventral; A, anterior; P, posterior. Fig. 4. wg and dpp are ectopically expressed in stat92e clones. All discs are third instar. stat92e clones were generated in a Minute (B,F) or non- Minute (D,E) background and lack GFP (green). Wild type discs (A,C). A E: Wg is ectopically expressed in stat92e clones in leg and antennal discs. In a wild type leg disc, dpp (blue) is expressed in the dorsal domain, while Wg protein (red) is expressed ventrally (A). In a leg disc containing a dorsal anterior stat92e clone, Wg is ectopically expressed (B, arrowhead), which leads to a duplicated axis, as assessed by the local outgrowth and ectopic, non-autonomous dpp expression adjacent to the ectopic Wg (B). In a wild type antennal disc, dpp is expressed ventrally, while Wg is expressed dorsally (C). In an antennal disc containing stat92e clones, Wg expression expands into the ventral domain and is associated with a secondary P-D axis, in which ectopic non-autonomous expression of dpp is observed (D). In an antennal disc containing a stat92e clone in the lateral posterior domain, the wg gene (wg P in red), is ectopically expressed within the clone (E,E arrowhead). F: dpp is ectopically expressed within stat92e clones. A stat92e clone in the dorsal lateral domain of the posterior compartment of a leg disc leads to ectopic expression of dpp (blue) in an autonomous manner (F,F arrowhead). wg P (wg-lacz) (anti- -Gal) is red in E and dpplacz, (anti- -Gal) is blue in A D and F. Magenta lines mark one side of the clone boundary in F,F.

2726 AYALA-CAMARGO ET AL. various locations in the leg disc. We find that dpp can be repressed in dorsal hop flip-out clones located at the lateral edge of the dpp expression domain but not in those located close to the A-P boundary (Fig. 5B,B arrowhead, and data not shown). Previous work has shown that loss of wg expression or signaling in the leg disc leads to ectopic expression of dpp in the ventral domain (Brook and Cohen, 1996; Jiang and Struhl, 1996). However, we find that dpp is never expressed within ventral hop-expressing clones located in or adjacent to the wg expression domain (21 discs examined) (Fig. 5C,C arrowheads, and data not shown). Lastly, we find that the effects of Stat92E on wg and dpp are not due to alteration in Hh expression in the absence of Stat92E (Supplemental Fig. 1F,G). Taken together these results suggest that JAK/STAT signaling represses dpp. However, the fact that supernumerary P-D axes and local repatterning are always associated with stat92e clones on the dorsal side of the leg disc suggests that the repressive effects of Stat92E are greater on the wg gene than on the dpp gene. Fig. 5. wg and dpp are repressed by JAK/STAT signaling. All discs are third instar. A,A : A large hop-expressing clone (abbreviated hop FO and marked by GFP and an arrowhead in A ) in the ventral leg disc strongly represses the wg P enhancer trap (red) in an autonomous manner. B,B : hop-expressing clones in the dorsal domain of the leg disc can repress dpp (magenta) in an autonomous manner (arrowhead) on the lateral edge of the dpp expression domain. C,C : dpp (magenta) is not expressed in hop-expressing clones (arrowheads) in the ventral domain of the leg disc that reside both within and adjacent to the wg expression domain. wg P (wg-lacz) (anti- -Gal) is red in A; dpp-lacz, (anti- -Gal) is magenta in B,C. White boxes in A C indicate regions that are magnified at 2 in A C. discs in which dpp was ectopically expressed in an autonomous manner (n 13 discs) and we made the following observations: (1) in all cases, dpp is ectopically expressed only in the dorsal domain and never in the ventral domain (13/13 discs); and (2) small patches of ectopic dpp are observed at the dorsal lateral margin of the disc, close to the boundary with the proximal region that gives rise to the trocanter and coxa (10/13) (Fig. 4F,F arrowhead). Our results indicate that the loss of Stat92E function can result in ectopic expression of dpp within the stat92e clone boundary. To assess if JAK/ STAT signaling leads to autonomous repression of dpp, we examined dpp expression in hop-expressing clones at wg and dpp Restrict upd Expression It is not known how upd becomes restricted to two domains as in leg and antennal discs (Fig. 6A, arrowheads). We observed that the cells producing upd, dpp, and wg do not overlap in either leg or antennal discs (Fig. 6G and data not shown). These mutually exclusive expression domains, together with our results that Stat92E represses wg and dpp, raise the possibility that there are multiple negative regulatory interactions between these factors. To investigate this issue, we examined upd mrna in leg discs that had temperature-sensitive mutations in key P-D axis patterning genes. Notch (N) signaling is necessary and sufficient to induce transcription of the upd gene in the eye disc (Chao et al., 2004; Reynolds-Kenneally and Mlodzik, 2005). Consistent with these published reports, we observed loss of upd at the posterior margin of N ts eye discs at the restrictive temperature (data not shown). In contrast, upd remains in its two domains in leg and

JAK/STAT PATHWAY IN DROSOPHILA 2727 antennal discs when N signaling is reduced, thus ruling out the hypothesis that N regulates upd in these tissues (Fig. 6C, arrowheads, and data not shown). However, we observed that upd is negatively regulated by Hh signaling, as upd mrna is found throughout hh ts2 leg discs at the restrictive temperature (Fig. 6B). To investigate whether the repressive effects of Hh were due to Wg and/or Dpp, we examined upd expression in hetero-allelic combinations that behave as wg ts or dpp ts alleles (van den Heuvel et al., 1993; Kenyon et al., 2003). When wg signaling is reduced, we find that upd expression collapses into a single domain that is aberrantly localized to the ventral leg disc (Fig. 6D, arrowhead). Similarly, a single upd domain that localized to the dorsal leg disc is observed when dpp signaling is reduced (Fig. 6E, arrowhead). Consistently, we also find that Upd activity is significantly altered in the absence of dpp signaling. In a dpp ts leg disc, Wg is ectopically expressed in the dorsal domain, which is consistent with previous reports (Brook and Cohen, 1996; Jiang and Struhl, 1996; Penton and Hoffmann, 1996). Moreover, the 10XSTAT92E- GFP reporter is ectopically expressed at the dorsal-most region of these discs, consistent with the dorsal expression of upd in these discs, but is lost from the central region of the disc (compare Fig. 1E to 6F). Taken together, our results indicate that both wg and dpp repress upd and hence restrict its expression to two domains in the leg disc (Fig. 6H). DISCUSSION Numerous studies have documented the function of the JAK/STAT pathway in proliferation and growth control. However, we report here for the first time the crucial role of the JAK/STAT pathway in P-D axis formation during development. Specifically, we demonstrate that one of the main functions of the JAK/STAT pathway is to restrict wg gene expression to a small domain in Drosophila appendages. Moreover, we show that Stat92E represses dpp in leg discs, although less robustly than Stat92E represses wg. Furthermore, we show for the first time that upd is subject to negative regulation by both Wg Fig. 6. Wg and Dpp restrict the upd expression domains. In A E, upd mrna is detected by in situ hybridization. B E show leg discs from animals shifted to the restrictive temperature. A: InaWTleg disc, upd mrna has two expression domains, one in the anterior and the other in the posterior compartment (arrowheads). B: In a hh ts2 leg disc, upd is ectopically expressed throughout the disc. C: In a N ts leg disc, upd expression is not affected and two upd expression domains are still observed (arrowheads). D: In a wg IL114/CX3 mutant disc, only one upd expression domain, aberrantly localized to the ventral part of the disc, is observed (arrowhead). E: In a dpp hr4/hr56 mutant disc, only a single domain of upd expression, localized to the dorsal part of the disc, is observed (arrowhead). F: In a dpp hr4/hr56 mutant leg disc, Wg (marked by Wg protein staining in red) and the 10XSTAT-GFP reporter are ectopically expressed in the dorsal domain. The 10XSTAT-GFP reporter is no longer expressed in the central part of the disc (compared with Fig. 1E). G: Wg, dpp, and upd are expressed in mutually exclusive domains in the antennal disc. Wg protein (red); dpp-lacz (anti- galactosidase blue) and upd GFP green. H: Model of the role of the JAK/STAT pathway in P-D axis formation in a third instar leg disc. Thickness of the line is intended to reflect the strength of the inhibitory interaction. See text for details. Proximal (Pr) and Distal (Di). and Dpp. Our results are consistent with a model in which Upd is restricted by the repressive actions of Wg and Dpp to two reciprocal groups of cells in the anterior and posterior compartments of leg and antennal discs. Upd also acts at a distance to restrict wg and dpp expression to their appropriate domains in leg and antennal discs and ultimately promotes the formation of a single, central P-D axis per disc (Fig. 6H). It is not yet known whether the mutual antagonism of upd-wg and upd-dpp is direct or indirect. Additional experiments will be needed to address these issues. Nevertheless, this study shows for the first time that these three major patterning factors (Upd, Wg, and Dpp) cross-regulate one another and that this regulation is crucial for proper P-D patterning of leg and antennal discs. Although we show that Stat92E autonomously represses both wg and dpp, it must be stressed that Stat92E appears to more strongly repress wg than it does dpp. This conclusion is supported by the following observations. First, ectopic wg is observed more frequently than ectopic dpp in stat92e clones. Second, stat92e clones only give rise to P-D axis duplication and overgrowth on the dorsal side of

2728 AYALA-CAMARGO ET AL. the leg disc. We presume that Stat92E repression of dpp may be cell-type specific, as we have previously shown that Stat92E represses wg but not dpp in the eye disc (Ekas et al., 2006). In contrast, our work demonstrates that the repression of wg by activated Stat92E is a broad mechanism utilized during pattern formation in imaginal discs and one that has recently been highlighted (Tolwinski, 2007). We have demonstrated the repression of wg by Stat92E in the eye and notal region of the wing disc (Ekas et al., 2006), as well as antennal and leg discs (this study). In the eye disc, we have identified a small (263 bp) enhancer that is negatively regulated by Stat92E in an autonomous manner (wg 2.11Z) (Ekas et al., 2006). The inability to look at the wg2.3z enhancer within stat92e clones precludes us from making conclusions about autonomous regulation of wg by Stat92E in leg and antennal discs through this enhancer (this study). Nevertheless, the wg2.11z enhancer is contained within wg2.3z, which leaves open the possibility that Stat92E regulates these two enhancers by the same molecular mechanism (Ekas et al., 2006; Pereira et al., 2006). Canonical Stat92E binding sites are not present in either wg enhancer, which suggests that Stat92E does not directly repress wg but rather acts through another factor. In the simplest scenario, activated Stat92E induces the expression of a repressor that acts directly on wg in eye, wing (notal region), antennal, and leg discs. However, we cannot rule out the possibility that Stat92E represses wg by distinct mechanisms in different discs. Moreover, there may be cryptic Stat92E binding sites in these wg enhancers, through which Stat92E may directly repress wg. Future work will be needed to address these issues. Do Mammalian STATs Repress BMP and Wnt Genes in Limb Formation? Our study raises the possibility that the JAK/STAT pathway also plays an important role during mammalian development by negatively regulating the expression of vertebrate Wnts and BMPs. Furthermore, our study also highlights the question of whether Wnts and BMPs regulate expression of cytokines/growth factors that activate JAK/STAT signaling during development. While it is not currently known whether the JAK/STAT pathway functions during mammalian limb development, BMP and Wnt genes do play fundamental roles in this process (Capdevila and Izpisua Belmonte, 2001). The elucidation of the roles of the JAK/STAT pathway during development have been hampered by the early embryonic lethality of stat3-deficient mice, and likely genetic redundancy between Stat3 and Stat5, which are most similar to Stat92E (Takeda et al., 1997; Levy and Darnell, 2002). The effects of a stat3/stat5 double knock-out mouse have not been reported. The generation of mice with conditional single stat3 or with double stat3 and stat5 deficiency only in the limb bud will be required to address this issue. Nevertheless, given the homology between molecules involved in limb development in Drosophila and mammals and the critical role of Stat92E in Drosophila appendage development presented here, our observations are likely to provide new insights into the role of STAT proteins in limb development in higher organisms. EXPERIMENTAL PROCEDURES D. melanogaster Stocks These stocks are described in FlyBase: y; stat92e 85C9 ; stat92e 397 ; dome 468 ; hop C111 ; dpp-lacz; wg-lacz (wg P ); UAS-hop; UAS-gfp; P{AyGAL4}25 P{UAS-GFP.S65T}T2; hs-flp MKRS/ TM6B; hs-flp; ey-flp; FRT 82B ubi- GFP(S65T) nls /TM6B, Tb 1 ; FRT 82B M(3)96C, Ubi-GFP; N ts ;hh ts2 /TM6B, Tb; awg temperature sensitive hetero-allelic combination wg IL114/CX3, referred to as wg ts in the text; a dpp temperature-sensitive hetero-allelic combination dpp hr4/hr56, referred to as dpp ts in the text. We also used y, hs- FLP; FRT 82B hscd2, y /TM2 and y, hs-flp; FRT 82B hscd2, y Minute/ TM2 (Casares and Mann, 2001), upd2 null allele (Hombria et al., 2005); upd3-gal4, UAS-GFP (Agaisse et al., 2003); 10XSTAT92E-GFP (Bach et al., 2007), UAS-3HA-stat92E, referred to as Stat92E FL in text and figures (Ekas et al., 2006), and wg2.3z (Pereira et al., 2006). stat92e Clones stat92e 85C9 and stat92e 397 were used interchangeably as they produce identical phenotypes (Silver and Montell, 2001; Ekas et al., 2006). We induced stat92e clones randomly using hs-flp or in the eye-antennal disc using eyflp. Collections of larvae of the genotype y, hs-flp; FRT 82B stat92e/ FRT 82B hscd2, y were heat shocked for 2 hr at 37 C during first or second instar. The adult legs and antennae were dissected and mounted in a 2:l mixture of Canada balsam and methyl salicylate. To provide stat92e cells with a growth advantage, we employed the Minute technique, which allows mutant cells to grow faster than surrounding heterozygous cells and to encompass the majority of the compartment in which they arise (Morata and Ripoll, 1975). To rescue stat92e antennal phenotypes, we used the MARCM technique (Lee and Luo, 1999) to drive expression of the UAS-3HA-stat92E transgene only in stat92e clones. hop and dome Clones We used dome 468, a strong hypomorphic allele, and hop C111, an amorphic allele, and generated clones in the eye-antennal disc using the ey-gal4 UAS-flp, GMR-hid technique (Stowers and Schwarz, 1999). This system relies upon the presence of a recessive cell lethal mutation on the chromosome arm of interest to generate large clones in the antenna. Temperature-Sensitive Experiments Flies of the appropriate genotype were allowed to lay eggs at the permissive temperature 18 C for 2 days. Their offspring were then shifted to the restrictive temperature 29 C for 2 days, and then they were returned to 18 C. For the N ts experiments, we isolated male larvae for the dissection of discs. For hh ts2 experiments, Tb larvae were dissected. The wg and dpp alleles were maintained over a compound chromosome SM6 TM6B and Tb larvae were isolated for dissec-

JAK/STAT PATHWAY IN DROSOPHILA 2729 tion. Discs were fixed and sense and anti-sense upd ribo-probes were generated according to the protocol in Bach et al. (2003). Antibody Staining and Microscopy Antibody stainings were performed as described (Bach et al., 2003). We used mouse anti-wg (1:15), mouse anti- galactosidase (1:20), mouse anti-dac (1:200) (all from the Developmental Studies Hybridoma Bank); rabbit anti-hth (1:1,000) (Pai et al., 1998); mouse anti-dll (1:500) (Duncan et al., 1998); rabbit anti- -galactosidase (1: 2,000) (Cappel); rabbit anti-hh (1:1,000) (Rogers et al., 2005); rat anti- Bab II (1:1,000) (Couderc et al., 2002). We used fluorescent secondary antibodies at 1:250 (Jackson Laboratories). We collected fluorescent images (at 25 ) using a Zeiss LSM 510 confocal microscope, and bright field pictures of adult legs or antennae (at 10, 20, or 40 ) using a Zeiss Axioplan microscope with a Spot camera. ACKNOWLEDGMENTS We thank R. Mann, J. Treisman, F. Casares, J. Hombria, H. Agaisse, H. Sun, I. Duncan, F. Laski, E. Rogers, and the Bloomington Stock center and the Developmental Studies Hybridoma Bank (University of Iowa, Department of Biological Sciences, Iowa City, IA 52242) for fly stocks and antibodies. We are indebted to J. Treisman, R. Dasgupta, and members of the Bach lab for reading the manuscript and for insightful comments. L.A.E. was supported in part from an NIH institutional training grant. E.A.B was supported in part by a Young Investigator Award from the Breast Cancer Alliance, Inc. and Whitehead Fellowship for Junior Faculty. REFERENCES Agaisse H, Petersen UM, Boutros M, Mathey-Prevot B, Perrimon N. 2003. Signaling role of hemocytes in Drosophila JAK/STAT-dependent response to septic injury. Dev Cell 5:441 450. Arbouzova NI, Zeidler MP. 2006. JAK/ STAT signalling in Drosophila: insights into conserved regulatory and cellular functions. Development 133:2605 2616. Bach EA, Vincent S, Zeidler MP, Perrimon N. 2003. A sensitized genetic screen to identify novel regulators and components of the Drosophila janus kinase/signal transducer and activator of transcription pathway. Genetics 165:1149 1166. Bach EA, Ekas LA, Ayala-Camargo A, Flaherty MS, Lee H, Perrimon N, Baeg GH. 2007. GFP reporters detect the activation of the Drosophila JAK/STAT pathway in vivo. Gene Expr Patterns 7:323 331. Basler K, Struhl G. 1994. Compartment boundaries and the control of Drosophila limb pattern by hedgehog protein. Nature 368:208 214. Blackman RK, Sanicola M, Raftery LA, Gillevet T, Gelbart WM. 1991. An extensive 3 cis-regulatory region directs the imaginal disk expression of decapentaplegic, a member of the TGF-beta family in Drosophila. Development 111:657 666. Bowman T, Garcia R, Turkson J, Jove R. 2000. STATs in oncogenesis. Oncogene 19:2474 2488. Bromberg JF, Wrzeszczynska MH, Devgan G, Zhao Y, Pestell RG, Albanese C, Darnell JE Jr. 1999. Stat3 as an oncogene. Cell 98:295 303. Brook WJ, Cohen SM. 1996. Antagonistic interactions between wingless and decapentaplegic responsible for dorsal-ventral pattern in the Drosophila Leg. Science 273:1373 1377. Campbell G, Tomlinson A. 1995. Initiation of the proximodistal axis in insect legs. Development 121:619 628. Campbell G, Weaver T, Tomlinson A. 1993. Axis specification in the developing Drosophila appendage: the role of wingless, decapentaplegic, and the homeobox gene aristaless. Cell 74:1113 1123. Capdevila J, Izpisua Belmonte JC. 2001. Patterning mechanisms controlling vertebrate limb development. Annu Rev Cell Dev Biol 17:87 132. Casares F, Mann RS. 2001. The ground state of the ventral appendage in Drosophila. Science 293:1477 1480. Chao JL, Tsai YC, Chiu SJ, Sun YH. 2004. Localized Notch signal acts through eyg and upd to promote global growth in Drosophila eye. Development 131:3839 3847. Cohen SM. 1993. Imaginal disc development. Cold Spring Harbor, NY: Cold Spring Harbor Press. 747 841 p. Cohen SM, Bronner G, Kuttner F, Juergens G, Jackle H. 1989. Distal-less encodes a homoeodomain protein required for limb development in Drosophila. Nature 338:432 434. Couderc JL, Godt D, Zollman S, Chen J, Li M, Tiong S, Cramton SE, Sahut-Barnola I, Laski FA. 2002. The bric a brac locus consists of two paralogous genes encoding BTB/POZ domain proteins and acts as a homeotic and morphogenetic regulator of imaginal development in Drosophila. Development 129:2419 2433. Darnell JE. 2005. Validating Stat3 in cancer therapy. Nat Med 11:595 596. Diaz-Benjumea FJ, Cohen B, Cohen SM. 1994. Cell interaction between compartments establishes the proximal-distal axis of Drosophila legs. Nature 372:175 179. Dong PD, Chu J, Panganiban G. 2000. Coexpression of the homeobox genes Distalless and homothorax determines Drosophila antennal identity. Development 127:209 216. Dong PD, Chu J, Panganiban G. 2001. Proximodistal domain specification and interactions in developing Drosophila appendages. Development 128:2365 2372. Duncan DM, Burgess EA, Duncan I. 1998. Control of distal antennal identity and tarsal development in Drosophila by spineless-aristapedia, a homolog of the mammalian dioxin receptor. Genes Dev 12:1290 1303. Ekas LA, Baeg GH, Flaherty MS, Ayala- Camargo A, Bach EA. 2006. JAK/STAT signaling promotes regional specification by negatively regulating wingless expression in Drosophila. Development 133: 4721 4729. Godt D, Couderc JL, Cramton SE, Laski FA. 1993. Pattern formation in the limbs of Drosophila: bric a brac is expressed in both a gradient and a wave-like pattern and is required for specification and proper segmentation of the tarsus. Development 119:799 812. Harrison DA, Binari R, Nahreini TS, Gilman M, Perrimon N. 1995. Activation of a Drosophila Janus kinase (JAK) causes hematopoietic neoplasia and developmental defects. Embo J 14:2857 2865. Heslip TR, Theisen H, Walker H, Marsh JL. 1997. Shaggy and dishevelled exert opposite effects on Wingless and Decapentaplegic expression and on positional identity in imaginal discs. Development 124:1069 1078. Hombria JC, Brown S, Hader S, Zeidler MP. 2005. Characterisation of Upd2, a Drosophila JAK/STAT pathway ligand. Dev Biol. Jiang J, Struhl G. 1996. Complementary and mutually exclusive activities of decapentaplegic and wingless organize axial patterning during Drosophila leg development. Cell 86:401 409. Johnston LA, Schubiger G. 1996. Ectopic expression of wingless in imaginal discs interferes with decapentaplegic expression and alters cell determination. Development 122:3519 3529. Kassis JA, Noll E, VanSickle EP, Odenwald WF, Perrimon N. 1992. Altering the insertional specificity of a Drosophila transposable element. Proc Natl Acad Sci USA 89:1919 1923. Kenyon KL, Ranade SS, Curtiss J, Mlodzik M, Pignoni F. 2003. Coordinating proliferation and tissue specification to promote regional identity in the Drosophila head. Dev Cell 5:403 414. Kiger AA, Jones DL, Schulz C, Rogers MB, Fuller MT. 2001. Stem cell self-renewal specified by JAK-STAT activation in response to a support cell cue. Science 294: 2542 2545.

2730 AYALA-CAMARGO ET AL. Lecuit T, Cohen SM. 1997. Proximal-distal axis formation in the Drosophila leg. Nature 388:139 145. Lee T, Luo L. 1999. Mosaic analysis with a repressible cell marker for studies of gene function in neuronal morphogenesis. Neuron 22:451 461. Levy DE, Darnell JE Jr. 2002. Stats: transcriptional control and biological impact. Nat Rev Mol Cell Biol 3:651 662. Mardon G, Solomon NM, Rubin GM. 1994. dachshund encodes a nuclear protein required for normal eye and leg development in Drosophila. Development 120: 3473 3486. Morata G, Ripoll P. 1975. Minutes: mutants of drosophila autonomously affecting cell division rate. Dev Biol 42:211 221. Pai CY, Kuo TS, Jaw TJ, Kurant E, Chen CT, Bessarab DA, Salzberg A, Sun YH. 1998. The Homothorax homeoprotein activates the nuclear localization of another homeoprotein, extradenticle, and suppresses eye development in Drosophila. Genes Dev 12:435 446. Penton A, Hoffmann FM. 1996. Decapentaplegic restricts the domain of wingless during Drosophila limb patterning. Nature 382:162 164. Pereira PS, Pinho S, Johnson K, Couso JP, Casares F. 2006. A 3 cis-regulatory region controls wingless expression in the Drosophila eye and leg primordia. Dev Dyn 235:225 234. Postlethwait JH, Schneiderman HA. 1971. Pattern formation and determination in the antenna of the homoeotic mutant Antennapedia of Drosophila melanogaster. Dev Biol 25:606 640. Reynolds-Kenneally J, Mlodzik M. 2005. Notch signaling controls proliferation through cell-autonomous and non-autonomous mechanisms in the Drosophila eye. Dev Biol 285:38 48. Rogers EM, Brennan CA, Mortimer NT, Cook S, Morris AR, Moses K. 2005. Pointed regulates an eye-specific transcriptional enhancer in the Drosophila hedgehog gene, which is required for the movement of the morphogenetic furrow. Development 132:4833 4843. Silver DL, Montell DJ. 2001. Paracrine signaling through the JAK/STAT pathway activates invasive behavior of ovarian epithelial cells in Drosophila. Cell 107: 831 841. Stowers RS, Schwarz TL. 1999. A genetic method for generating Drosophila eyes composed exclusively of mitotic clones of a single genotype. Genetics 152:1631 1639. Struhl G, Basler K. 1993. Organizing activity of wingless protein in Drosophila. Cell 72:527 540. Tabata T, Kornberg TB. 1994. Hedgehog is a signaling protein with a key role in patterning Drosophila imaginal discs. Cell 76:89 102. Takeda K, Noguchi K, Shi W, Tanaka T, Matsumoto M, Yoshida N, Kishimoto T, Akira S. 1997. Targeted disruption of the mouse Stat3 gene leads to early embryonic lethality. Proc Natl Acad Sci USA 94:3801 3804. Theisen H, Haerry TE, O Connor MB, Marsh JL. 1996. Developmental territories created by mutual antagonism between Wingless and Decapentaplegic. Development 122:3939 3948. Tolwinski NS. 2007. Meeting report: new advances in signaling and pattern formation. Fly 1(2):116 117. Tulina N, Matunis E. 2001. Control of stem cell self-renewal in Drosophila spermatogenesis by JAK-STAT signaling. Science 294:2546 2549. van den Heuvel M, Harryman-Samos C, Klingensmith J, Perrimon N, Nusse R. 1993. Mutations in the segment polarity genes wingless and porcupine impair secretion of the wingless protein. Embo J 12:5293 5302.