SEARCH

SEARCH BY CITATION

Keywords:

  • JAK/STAT;
  • pole cells;
  • Drosophila;
  • primordial germ cells;
  • gonad;
  • migration;
  • cell survival;
  • filopodia

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

The gonad is formed from two populations of cells originating at different locations: the primordial germ cells (PGCs), giving rise to either sperm or oocytes, and the somatic gonadal mesoderm precursors (SGPs), which support development of the gametes. Following the PGCs' migration during gastrulation, these two populations meet, forming the immature gonad. We present evidence that during embryonic development, the PGCs require the canonical JAK/STAT signalling cascade to migrate efficiently towards the SGPs. Loss of function for any element of the JAK/STAT pathway causes frequent germ cell mislocalisation. We have found that wild-type germ cells produce filopodia while they migrate through the mesoderm towards the gonad. Our observations suggest that PGCs use filopodia to migrate and to keep contact with each other. Interestingly, activation of the JAK/STAT pathway is required for these filopodia to form, and ectopic JAK/STAT activation enhances their formation. Developmental Dynamics 235:958–966, 2006. © 2006 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Primordial germ cells (PGCs) are essential for a sexually reproducing organism to produce gametes and future generations. The PGCs are the precursors of the adult gonadal stem cell population and so the identification of factors that control their migration and survival is important. The primitive gonad in the embryo is derived from two sub-populations of cells, the PGCs and the mesoderm-derived somatic gonadal precursors (SGPs). Recent work has highlighted a number of proteins that are required for PGCs migration during embryogenesis (reviewed in Howard,1998; Wylie,1999; Matova and Cooley,2001; Starz-Gaiano and Lehmann,2001; Kunwar and Lehmann,2003).

In Drosophila, the pole cells form the embryonic germline component of the gonads (PGCs). Pole cells can be first visualised extraembryonically at the posterior end of the blastoderm. With the start of gastrulation, pole cells are passively carried into the embryo by the amnioproctodeal invagination, where they remain until they pass through the midgut epithelium into the embryo and a repulsive signal, generated by the wunen (Zhang et al.,1997) and wunen-2 genes (Starz-Gaiano et al.,2001), repels them towards the mesoderm. Pole cells traversing the amnioproctodeal invagination have been shown to have an actin-rich cortex and actively growing cytoplasmic expansions, forming pseudopodia (Jaglarz and Howard,1995). These cells also have cytoplasmic ruffles as they move through the midgut epithelium (Callaini et al.,1995). Once inside the embryo cavity, they bifurcate and then migrate laterally over the mesodermal substrate. It is here that they attach to the specialised gonadal mesoderm (Brookman et al.,1992; Boyle and DiNardo,1995) that will ensheathe the pole cells and compact, forming the embryonic gonads at stage 14 (Moore et al.,1998a; Jenkins et al.,2003; Van Doren et al.,2003). In the wild type embryo, a fraction of the PGCs do not reach the gonadal mesoderm and are eliminated by a process of cell death (Coffman et al.,2002).

In higher animals, several extracellular attractant signals that control migration have been isolated (Doitsidou et al.,2002; Ara et al.,2003; Knaut et al.,2003). Hh plays an important role in chemotaxis of these mesodermally guided movements, but is thought to act redundantly with another signalling pathway (Deshpande et al.,2001). The Drosophila Hh pathway is thought to be required cell autonomously in the pole cells. In addition to this columbus, a gene encoding an enzyme that is involved in lipid metabolism is expressed in the SGP and is required for the emanation of an unknown extracellular attractant signal (Van Doren et al.,1998a).

In vertebrates, there is some evidence suggesting that the JAK/STAT pathway is required for PGC development. In mice, the Leukemia inhibitory factor (LIF) (Matsui et al.,1991; Cheng et al.,1994) and GP130 (IL-6 receptor signal transducing component), which both signal through JAK/STAT, are required for the in vitro proliferation of PGCs (Koshimizu et al.,1992). However, LIF−/− (Stewart et al.,1992; Escary et al.,1993) and LIFR−/− (Ware et al.,1995) mice produce fertile adults. As there is redundancy in most levels of the vertebrate JAK/STAT signalling cascade, it is difficult to ascertain a function for JAK/STAT signalling in gonadal development. The GP130 receptor subunit has been shown to be required for normal germ cell numbers in the gonad as GP130−/− male mice show a significantly reduced number of PGCs. GP130 is not, however, required for germ cell differentiation (Molyneaux et al.,2003). There is currently no evidence for the role of STAT during PGC migration or survival as STAT 1,2,4, and 6 knock-outs are fertile and STAT3, 5a, and 5b can not be assessed because they are embryonic lethal (Ihle,2001). Studying the requirement of the JAK/STAT pathway in germ cells in Drosophila should be helped by the absence of redundancy in the system. The Drosophila JAK/STAT signalling pathway has (Fig. 1D) fewer elements at each level of the signalling pathway: the ligands are encoded by the cytokine-like molecule of the Unpaired family (Harrison et al.,1998; Gilbert et al.,2005; Hombría et al.,2005; Agaisse et al.,2003), the receptor by domeless (dome) (Hou et al.,1996a; Brown et al.,2001), the Janus kinase (JAK) by hopscotch (hop) (Binari and Perrimon,1994), and the transcription factor by stat92E gene (Hou et al.,1996b; Yan et al.,1996). Once the receptor has been activated, a cascade of intracellular tyrosine phosphorylation occurs, which results in STAT translocation to the nucleus where it induces gene transcription of downstream targets. The JAK/STAT pathway has been shown to be essential for several gonadal functions including stem cell self-renewal in testes (Kiger et al.,2001; Tulina and Matunis,2001), border cell migration during oogenesis (Silver and Montell,2001; Beccari et al.,2002; Ghiglione et al.,2002), and patterning in the egg chamber in Drosophila (Baksa et al.,2002; McGregor et al.,2002; Xi et al.,2003).

thumbnail image

Figure 1. upd and upd3 are expressed in the embryonic gonad. A: Dorsal view of a stage-13 wild-type embryo stained with an upd3 RNA probe. Faint upd3 RNA expression can be detected in the gonad, arrowhead. B: Late wild-type embryo stained with the same probe as A revealing upd3 anterior expression in the gonad, arrowhead. C: Dorsal view of a stage-16 embryo stained with a upd RNA probe. The arrowhead indicates anterior gonadal expression of upd. B',C': Gonad outlined by dashed line, showing, respectively, anterior refinement of expression of upd3 and upd expression. D: Schematic representation of the Drosophila JAK/STAT pathway. The UPD ligands bind a homodimerised receptor (DOME), activating the JAK kinase (HOP), which in turn activates by phosphorylation the cytoplasmic STAT transcription factor (STAT92E).

Download figure to PowerPoint

In this study, we characterise the role of the JAK/STAT signalling cascade in the formation of the embryonic gonad. We find that pathway activity is required in the pole cells for their migration. We show that wild-type migratory pole cells make filopodial extensions contacting each other. These extensions are enhanced by pathway hyperactivation and lost when STAT is not activated. Although the pathway may have some secondary function as a chemo-attractant, we suggest that its major role is to enhance cell motility during migration.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

The Canonical JAK/STAT Pathway Plays a Role in Germ Cell Migration

As part of our work characterizing the functions of the JAK/STAT pathway in Drosophila, we studied the expression patterns of the genes with homology to upd. There are three genes of the upd family and they map closely to each other (Hombría and Brown,2002). One of them, CG15062, a gene also known as upd3, is expressed in the gonad at stage 13 when the gonad has just coalesced (Fig. 1A, arrowheads) and later refines anteriorly in the gonad (Fig. 1B). In situ RNA hybridizations showed that upd is also expressed in the gonad at late stages (Fig. 1C), in a similar position to upd3. However, upd2 (CG5988) is not expressed in the gonad or the germ cells (data not shown). Because stat92E is expressed at high levels in the germ cells during Drosophila gastrulation (Yan et al.,1996), and the canonical JAK/STAT pathway has been implicated in mouse embryonic germ cell development (Molyneaux et al.,2003), we decided to investigate if the pathway is required for PGC development. To identify the PGCs in different mutant backgrounds, we used the pole cell specific antibody Vasa (Hay et al.,1988).

Analysis of a null upd mutant did not reveal any defects on PGC migration (see Experimental Procedures section for alleles tested). Because there are no mutant alleles for upd3, we expressed an RNAi construct, UAS-iupd3 (Agaisse et al.,2003), in the germ cells and in the mesoderm using nos-GAL4VP16 and 24B-GAL4, but failed to find any PGC defects (data not shown). Given the homology shared by the UPD-ligands, we tested potential redundancy between upd and upd3 by simultaneously deleting all upd-like loci to assess if this would affect PGC development. Df(1)os1A is a deletion that removes among other genes upd and all upd-like ligands and is a complete null for all JAK/STAT pathway functions (Hombría et al.,2005). Compared to the characteristic wild-type embryo PGC migration (Fig. 2A–C), in Df(1)os1A embryos, we first detect a phenotype at stage 12 when the pole cells are more dispersed than in wild-type (Fig. 2D compared to A). At the time of gonadal coalescence, at stage 14, some pole cells do make it to the gonad but many are scattered throughout the posterior end of the embryo (Fig. 2E, arrow shows gonad). This is also the case in Df(1)osUE69, another chromosomal deletion mutant that also removes a similar region of the genome (Gilbert et al.,2005; Hombría et al.,2005). Although these results suggest that the germ cell migration phenotype is due to the loss of STAT activation caused by the absence of upd-like ligands, it could also be caused by another gene missing in the deletion.

thumbnail image

Figure 2. Requirement of the canonical JAK/STAT pathway for PGC migration. A–L: Whole mount antibody stains with anti-Vasa antibody. AC: Wild type. DF: Df(1)os1A. GI: stat92E null mutants. J: dome null mutants. K,L: Zygotic dome468. (A,D,G) stage 12, (B,E,H,K) stage 14, (C,F,I,J,L) late embryogenesis. All embryos are dorsal views and orientated anterior left.

Download figure to PowerPoint

To confirm that the JAK/STAT pathway is required for PGC migration, we stained stat92E (Fig. 2G–I) and dome (Fig. 2J) maternal zygotic null mutants (which will now be referred to as dome or stat92E null mutants, see Experimental Procedures section) with anti-Vasa. The phenotypes observed in dome null mutants are similar to those obtained with Df(1)os1A (Fig. 2D–F). However, stat92E null mutant embryos have a more extreme defect as in many cases pole cells remain extra-embryonically (Fig. 2G, arrow) and do not cross the gut. These pole cells do not die and can still be visualised later in embryogenesis on the surface of the embryo. However, the most striking phenotype is that most of the pole cells cross the midgut epithelium but fail to find the gonadal mesoderm. The observation that stat92E null has a more extreme phenotype than loss of the upd family or dome null mutants suggests that STAT92E may be activated by other signaling pathways. Li et al. (2003) have shown that, at the blastoderm stage, STAT92E is activated in the pole cells under the control of the TOR/RAS/MAPK pathway (Li et al.,2003). They proposed an early developmental function for STAT92E in germ cell proliferation and migration. Our observations indicate that the upd family also plays a role in the activation of STAT92E, once the pole cells exit the gut. A variable penetrance of the phenotypes can be observed with stat92E mutants as with the deletion of the upd family, as there are some embryos with no pole cells in the gonad while others are nearly normal. Ten percent of stat92E null embryos form 2 gonads while 60% of animals form only 1 gonad compared to 51% of Df(1)os1A embryos forming 2 gonads and 23% of animals forming only 1 gonad (n ≥ 30) (Fig. 2H, arrow).

Although all the above results suggest that the canonical JAK/STAT pathway is required for normal germ cell migration, the source of the ligand is unlikely to come from the gonad, as migration defects are observed from stage 12 before upd and upd3 are expressed in the gonads.

Pole Cell Migration Defects Are Not Caused by the Abnormal Segmentation or Somatic Gonadal Precursor Defects

The gonad requires the specification and development of the SGPs (Zhang et al.,1997) as well as the germ cells. Many mutations have been discovered that affect the SGPs and in turn affect the correct migration of pole cells (Boyle and DiNardo,1995; Greig and Akam,1995; Boyle et al.,1997; Moore et al.,1998b). The gonadal mesoderm is comprised of cells derived from parasegments 10–12, which coalesce and gather in the 5th abdominal segment. As JAK/STAT pathway mutants have abdominal segmental defects consisting of the deletion of denticle belts A4, A5, and the fusion of A6 with A7, the abnormal pole cell migration we observed could be indirectly caused by a primary mesoderm defect. To establish if this is the case, we investigated PGC behaviour in dome zygotic mutant embryos, which do not have segmental defects. The receptor, dome, is expressed maternally and zygotically. Maternal dome contribution is sufficient to allow normal segmentation in dome zygotic mutants, and thus allows us to detect PGC defects late in embryogenesis in an almost normal embryo (Brown et al.,2001). The analysis of dome zygotic mutants reveals that although most pole cells do migrate into the gonad due to rescue by maternally supplied dome, some pole cells mislocalise and remain spread through the embryo (Fig. 2K,L).

To test whether loss of JAK/STAT signalling affects the mesodermal component, we used two SGP markers, 412 and eyes absent (eya) (Brookman et al.,1992; Boyle et al.,1997; Bonini et al.,1998). In stage 11 wild-type embryos, 412 labels the SGPs among other cells (Broihier et al.,1998) (Fig. 3A, arrows) and at stage 14, when the gonads have coalesced, 412 is expressed exclusively in the SGP (Fig. 3B).

thumbnail image

Figure 3. JAK/STAT signalling is not required for gonadal mesoderm formation. A–D: RNA in situ with the 412 probe revealing the SGP. A,B: Wild-type embryos. C,D: Df(1)os1A embryos. A–C: Embryos at stage 11 with arrowheads pointing at the gonadal mesoderm precursors; B,D: stage 14 embryos.

Download figure to PowerPoint

Df(1)os1A embryos have normal gonadal expression of 412 (Fig. 3C,D), and a similar analysis using the antibody against EYA shows that in Df(1)os1A embryos, EYA staining and gonadal mesoderm cell numbers are normal (27–36 cells). We also found that ectopic upd expression in the mesoderm using the mesodermal line 24B-Gal4 does not cause an increase in the number of SGP cells as marked by either 412 or EYA (not shown). This shows that JAK/STAT signalling does not affect the specification of gonadal mesoderm and that pole cell migration defects are not caused by the loss of SGPs.

The JAK/STAT Pathway Is Activated in Migrating Pole Cells

A lack of obvious defects in the gonadal mesoderm and the expression of stat92E RNA in the pole cells (Yan et al.,1996), suggests that JAK/STAT signalling may be needed autonomously in the pole cells. To test if this is the case, we investigated for signs of pathway activation in the pole cells. A pre-requisite for JAK/STAT activation is the presence of a dimerised receptor (Brown et al.,2003). We have shown that the βlue-βlau technique allows the detection of DOME receptor oligomerisation by β-galactosidase complementation. This technique uses two different mutant βGAL proteins (Δα and αω) attached to the C terminal end of the DOME receptor. If receptor dimerisation occurs, the βGAL mutants come into proximity and are capable of reconstituting βGAL activity (Brown et al.,2003). Using the germ cell–specific driver, nos-GAL4VP16, we expressed both UAS-domeless-Δα and UAS-domeless-Δω in the pole cells and observed β-galactosidase activity at stages 11–16, indicating that DOME is dimerised at these stages and thus is competent to activate the pathway in the pole cells (Fig. 4A,B). We did not observe dimerisation at earlier stages due probably to late expression of the GAL4 driver, as the expression of other UAS lines driven by nos-GAL4VP16 was also found to be very weak at earlier stages.

thumbnail image

Figure 4. Signs of JAK/STAT pathway activation in pole cells. A,B: Wild type embryos expressing βlue-βlau constructs in the pole cells revealing DOME homo-dimerisation in the pole cells by the use of the X-GAL substrate. A: Stage 12; B: stage 14. C,D: Progeny from nos-GAL4VP16 females mated with UAS-stat92E-GFP males and pole cells visualised on a confocal microscope as a projection. (C) Stage-12 wild-type and (D) Df(1)os1A stage-14 pole cells.

Download figure to PowerPoint

To test STAT activation in the pole cells more directly, we monitored STAT92E sub-cellular localization using a STAT-GFP protein. STAT translocation to the nucleus and transcriptional activity is required for all JAK/STAT-dependent signalling and can be used as an indicator of pathway activity (Lillemeier et al.,2001). Expression of UAS-stat92E-GFP with nos-GAL4VP16 in wild-type pole cells allowed us to analyse STAT localisation from stage 11. Initially, STAT-GFP can be seen residing in all cellular compartments, but from late stage 12 (Fig. 4C) until stage 16, STAT-GFP shows a bias towards its nuclear distribution. To show that the nuclear accumulation was related to ligand activity and not just to the normal shuttling of STAT between the nucleus and cytoplasm, we repeated the experiment in a Df(1)os1A background. STAT-GFP does not accumulate in the nucleus of stages 11–16 in flies that are missing all JAK/STAT ligands (Fig. 4D). These results show that the UAS-stat92E-GFP transgene is a useful tool for the analysis of activated STAT signalling. The observation that stat92E accumulates in the nucleus of the migrating pole cells suggests that the migration defects we observe are due, at least in part, to STAT activation in the pole cells. It also shows that the removal of the upd family has an impact on STAT cellular localization, as well as pole cell migration.

The observation that STAT92E mutants have an effect on the PGCs, contrasts with a recent report claiming that STAT92E and phosphorylated STAT92E are not detected in the pole cells prior to gonadal coalescence (Wawersic et al.,2005). Our results indicate that low levels of STAT92E undetectable by the antibody must be present in the cells before coalescing. In fact, Wawersic et al. can induce high levels of STAT in the female germ cells by expressing the activated form of JAK: HOPTum-l (Wawersic et al.,2005). This activation effect could only be possible if, as we propose, STAT92E is already expressed in the germ cells. We believe that JAK/STAT activation after gonad coalescence may stabilize STAT or activate a positive feed-back loop boosting stat92E levels of expression in the germ cells (compare levels of STAT in Fig. 4C and D). It is probably this high level that Wawersik et al. are detecting in their antibody stainings.

JAK/STAT Ligands Do Not Seem to Act as Major Chemo-Attractant Cues

Reports studying hedgehog (hh) function during pole cell migration show that it acts as a chemo-attractant (Deshpande et al.,2001). To determine if the UPD protein family could have a similar effect during the mesodermal migratory phase, we expressed upd in the mesoderm using either of the pan-mesoderm lines 24B-Gal4 or twi-Gal4. Figure 5A shows a stage-14 24B-Gal4 UAS-upd embryo stained with Vasa antibody (for 24B-Gal4 driver pattern, see Deshpande et al.,2001). Although some cells migrate abnormally as previously reported (Li et al.,2003), this phenotype is very variable, with most cells reaching the gonad, arguing against a major chemo-attractant role. Seventy-eight percent of embryos have fully formed gonads with a few wandering pole cells; 11% form 1 gonad; 11% are missing the gonads; and there are no completely wild-type embryos (n = 19).

thumbnail image

Figure 5. Pole cell migration in different activated JAK/STAT backgrounds. A–D: Embryos stained with Anti-Vasa antibody. A: Embryo with 24B-GAL4 driving UAS-upd in the mesoderm, revealing some mislocalisation of pole cells. B: nos-GAL4VP16 driving UAS-upd. C: Close-up of a nos-GAL4VP16 driving UAS-hoptum-l. D: Wild-type embryo. (A) Stage-14 and (B–D) stage-13 embryos.

Download figure to PowerPoint

To further investigate if the pole cells are responding to UPD as a chemo-attractant, we expressed UPD only in the pole cells using nos-Gal4VP16. Here the pole cells would receive a more intense, localized autocrine signal. We hypothesized that there would be little or reduced pole cell migration if UPD was acting as an attractant, “fooling” the cells into perceiving they have reached the attractant source. In these conditions, no migration defects were observed (Fig. 5B). Similarly, expression of hyperactivated JAK, UAS-hopTum-l, in the pole cells with nos-Gal4VP16 (Harrison et al.,1995) does not result in migration defects. In both cases, we observed higher numbers of pole cells in the gonad. This effect may be caused by extra proliferation in the germ cells induced by STAT activation (Li et al.,2003).

Pole Cell Behaviour Before Gonadal Coalescence Is Affected by the JAK/STAT Pathway

To determine the cause for the migration phenotypes in terms of cellular behaviour, we investigated potential mechanisms that wild type migrating pole cells use. Migrating pole cells form cellular extensions consisting of pseudopodia and membrane ruffling as they pass through the posterior midgut (Callaini et al.,1995; Jaglarz and Howard,1995). As an initial step, we decided to analyse if the JAK/STAT pathway affects the formation of extracellular protrusions in pole cells. To outline the pole cells and to visualize cortical actin, we expressed the actin-binding region of Moesin fused to GFP (Dutta et al.,2002), known as UAS-GMA. Using the driver line nos-GAL4VP16 enabled us to observe any extensions that the pole cells made during normal migration. In the wild type stage 11, GMA is normally found at the cortical region of each pole cell and concentrates at the plasma membrane where pole cells contact each other (Fig. 6A and C, white arrowhead). Filopodia can be visualised in wild type embryos at stage 11 and up to late stage 12. Two kinds of filopodia can be observed. Long filopodial extensions up to 12 μm in length (Fig. 6A, white arrow) that only make contact between distant pole cells; and short filopodia ranging from 1–4 μm in length that extend into the surrounding mesodermal substrate (Fig. 6A and C, yellow arrows). These smaller projections occur from stage 11 to 14 and their formation does not appear to be affected by pole cell clustering. About 30% of all cells make more than one small projection but never more than four. Pseudopodia occur in about 15% of all cells (Callaini et al.,1995; Jaglarz and Howard,1995) (Fig. 6B, blue arrow). When the pole cells coalesce and become ensheathed by the SGP, pseudopodia and long filopodia cease to form but the smaller filopodia are maintained until stage 14. When the pole cells are surrounded by the gonad mesoderm, GMA becomes more punctate, but still does accumulate at membranes where pole cells touch each other (Fig. 6C, white arrowhead).

thumbnail image

Figure 6. Filopodia formation in pole cells during migration along the mesoderm. Close up of pole cells marked using nos-VP16 GAL4 UAS-GMA. Wild type pole cells migrating in the posterior end of the embryo (A,B), or at the coalescence stage (C). GFP was visualised using confocal microscopy and all projections are represented here. Wild type pole cells form different types of extensions. Long filopodia connecting separate groups of cells (A, white arrow), shorter filopodia exploring the substrate (A yellow arrows), and moesin-rich pseudopodia (B, blue arrow). A white arrowhead indicates concentration of moesin in abutting cells. In Df(1)os1A embryos (D,E), filopodia are mostly absent but abutting membranes still accumulate moesin. Pole cells with hyper-activated JAK (F,G; expressing UAS-hopTum-l) have increased spots of GMA and form cellular projections at late stages (G). (A) Stage-11, (B, D, F) stage-12, and (C, E, G) late embryogenesis after gonadal coalescence. Scale bar in A = 5 μm.

Download figure to PowerPoint

We next studied how these motile behaviours are affected in hyperactivation or loss-of-function JAK/STAT mutants. Pole cells in Df(1)os1A embryos have plasma membranes with less ruffles and appear to be smoother as visualized with GMA, compared to wild type (compare Fig. 6B with D). There are two types of mislocalising pole cells that can be observed: ones that are singular and do not show any intense plasma membrane localisation of GMA. The other type are pole cells that form small groups clumping together. The latter pole cells show that JAK/STAT signaling is not required for pole cell aggregation. However, as they do not migrate to the gonad, they are not ensheathed by gonadal mesoderm and, subsequently, they become directly attached to each other. In the Df(1)os1A background, we found that GMA did localise to the plasma membranes of abutting pole cells, but was much reduced in the cortical areas (Fig. 6D,E). We found that pole cells in Df(1)os1A embryos extended virtually no filopodia (only one out of 1,500 Df(1)os1A pole cells had filopodia).

Expression of a hyperactivated JAK protein in the pole cells using nos-Gal4VP16 UAS-hopTum-l induced premature clumping of the pole cells from stage 12 (Fig. 6F). The pole cells also maintain high numbers of the smaller extracellular projections (Fig. 6F, yellow arrows) even into the late stages of embryogenesis (Fig. 6G, yellow arrows). However, the most striking observation is the high accumulation of spots of GMA at stage 12. This is normally seen at stage14 in wild type, but is premature and more pronounced in the hopTum-l expressing animals suggesting a higher number of stable cell-cell contacts.

These results suggest that the role of the JAK/STAT pathway during pole cell migration is to enhance extracellular protrusions, therefore facilitating migration.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

The Canonical JAK/STAT Pathway and Pole Cell Migration

A recurrent theme of JAK/STAT signalling controlling cellular movements is becoming apparent (reviewed by Hirano et al.,2000; Hou et al.,2002). In Dictyostelium, Dd-STATa is required for proper chemotaxis during development (Mohanty et al.,1999), zebrafish requires STAT3 for convergent extension movements during gastrulation (Yamashita et al.,2002), and STAT3 is essential for skin wound healing (Sano et al.,1999) and also the migration of hematopoietic progenitor cells through the activation of JAKs and STATs by the chemokine SDF-1 (Zhang et al.,2001). In Drosophila, the pathway is also required for convergent extension movements of the hindgut (Johansen et al.,2003) and border cell migration (Silver and Montell,2001; Ghiglione et al.,2002).

The work presented here and recent work in Drosophila shows that the JAK/STAT pathway is required in the pole cells at three different stages during embryogenesis. The first stage of requirement is at the cellularisation stage, when pole cells require STAT and the Ras pathway for proliferation and invasive behaviour (Li et al.,2003). For these functions, STAT activation is achieved via the receptor tyrosine kinase TORSO. Our results indicate that there is a second stage after the initial TOR activation of STAT, when there is a role for canonical Drosophila UPD-family/DOME/STAT signalling in the migration of the pole cells through the mesoderm. We have found that upd3 and later upd are expressed in the embryonic gonad. However, this expression is too late to be responsible for pole cell migration towards the gonad, indicating that non-gonadal ligand reaching the pole cells from the surrounding epithelial tissues is enough to enhance the pole cell motility. Although, upd mutants do not affect germ cell migration, deletions removing multiple members of the UPD family cause a reduction in STAT nuclear localisation and pole cell migration defects similar to those observed in DOME mutants. These results suggest that a combination of UPD ligands control STAT activation after gastrulation when TOR is inactive. While STAT and TOR mutants cause pole cells to fail to migrate through the midgut, null dome, JAK (hop) mutants (Li et al.,2003), or the deletion of the UPD family members do not. This indicates the requirement for the canonical JAK/STAT pathway must be later in the migratory phase through the mesoderm. At this stage, the pathway is required in the pole cells to increase cell movement by enhancing the formation of filopodia. Finally, in a third stage after pole cell ensheathment, male-specific gonad expression of UPD is responsible for germline sexual development (Wawersic et al.,2005).

Our results suggest that the JAK/STAT pathway is not required for cell adhesion but for cell motility during the mesodermal migratory phase. In fact, mislocalised pole cells clump, suggesting that JAK/STAT signalling does not affect homotypic cell adhesion affinities (Van Doren et al.,2003). In the absence of efficient protrusion formation, cells can still find the guiding cues, but they do it more inefficiently and as a result many more get mislocalised. The absence of upd expression in the gonadal mesoderm at the second stage when the pole cells migrate through the mesoderm, and the partially correct migration to the gonads in UPD null mutants, suggest that the UPD family members are not acting as major chemo-attractant cues.

In summary, we show that the germ cells form extracellular projections during their mesodermal migratory phase and that the canonical JAK/STAT pathway can induce this autonomously in the PGCs. Hyperactivation of the pathway induces premature clumping of the germ cells, facilitates the movement of cells to the gonad, and sustains the production of filopodia even when coalesced with the SGP.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Molecular Biology

Full-length stat92E cDNA, including the EPEPLVL motif, was PCR amplified using the following primers, STAT-F-B: 5′ GAGGTACCGAGCATGAGCTTGTTGAAGCGC 3′ and STAT-R-B: 5′ ACGGATCCGAAAAGTTCTCAAAGTTTGTAATC 3′. The product was then trimmed and cloned to make a fusion at the C terminal end to EGFP and inserted into the Kpn1 and Xba1 sites of pUAST. The pEGFP-C1 plasmid was used as the source of EGFP (Clontech, Palo Alto, CA). The ends of the construct were sequenced and the core of the construct replaced with a fragment from the stat92E cDNA (in case of PCR errors). The construct does not contain any UTR's (other than the SV40 poly A trailer in UAST) and was injected into w1118 flies using helper plasmid.

Flies and Genetics

Flies were reared at 25°C under standard conditions. nos-GAL4VP16 (Van Doren et al.,1998b), 24B-GAL4 (Brand and Perrimon,1993), twi-GAL4, UAS-upd, UAS-hoptum-l, UAS-moeGFP, UAS-GMA, UAS-actGFP, Df(1)os1A, Df(1)osUE69, dome217, dome468, stat92E6346, upd1, upd4 were used in this study. dome217 and stat92E6346 germ line clones were generated as described previously [33] (Brown et al.,2001).

Immunohistochemistry

To view pole cell cellular extensions, nos-GAL4VP16 females were crossed to UAS-GMA males. Live embryos were dechorinated and mounted in vectashield (VectaLabs) and viewed on the confocal microscope for up to 1 hr.

Antibodies: EYA (Bonini et al.,1993) (previously known as clift) was obtained from Developmental Studies Hybridoma Bank. Vasa (Hay et al.,1988) was a kind gift from C. Extavour.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

We thank Rob White for the 412 probe. UAS-iupd3 is a gift from Herve Agaisse. We are grateful to Norbert Perrimon in whose laboratory some of the experiments were carried out. We also thank Nan Hu for assistance in some experiments, and Barry Denholm, Nan Hu, Bridget Lovegrove, and Sol Sotillos for constructive criticisms of the manuscript. This work was supported by the Wellcome Trust (S.B. and J.C.-G.H.), the Royal Society and Ministerio de Educación y Ciencia; BFU2004/1069BMC (J.C.-G.H.), and Emmy Noether Award of the Deutsche Forschungsgemeinschaft. (M.Z.)

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES