Research ArticleDevelopmental Biology

The Drosophila Female Germline Stem Cell Lineage Acts to Spatially Restrict DPP Function Within the Niche

See allHide authors and affiliations

Science Signaling  27 Jul 2010:
Vol. 3, Issue 132, pp. ra57
DOI: 10.1126/scisignal.2000740

Abstract

Maintenance of stem cells requires spatially restricted, niche-associated signals. In the Drosophila female germline stem cell (GSC) niche, Decapentaplegic (DPP) is the primary niche-associated factor and functions over a short range to promote GSC self-renewal rather than differentiation. Here, we show that the GSC lineage and, more specifically, the stem cells themselves participate in the spatial restriction of DPP function by activating epidermal growth factor receptor (EGFR)–mitogen-activated protein kinase (MAPK) signaling in the surrounding somatic cells. EGFR-MAPK signaling in somatic cells repressed the expression of dally, which encodes a glypican required for DPP movement and stability. Consequently, only GSCs close to the DPP source (the somatic cells in the niche) showed high signal activation and were maintained as stem cells, whereas cystoblasts outside the niche showed low signal activation and initiated differentiation. Thus, our data reveal that the reciprocal crosstalk between the GSCs and the somatic cells defines the spatial limits of DPP action and therefore the extent of the GSC niche.

Introduction

Homeostasis of adult organs relies on a small population of tissue-specific stem cells that can self-renew and at the same time generate tissue-specific differentiated cells to replenish outgoing cells. Stem cells reside in a specialized microenvironment, called a niche, which is provided by surrounding stromal cells. The niche provides not only a physical location but also the extrinsic signals to maintain stem cell self-renewal (15). In addition, increasing evidence implicates that dysfunction of the stem cell niche may be involved in diseases associated with excessive proliferation of stem cell–like populations and with tumorigenesis (68). The niche activities that promote stem cell identity must be spatially restricted to achieve this balance.

Drosophila female germline stem cells (GSCs) provide a well-established system for studying stem cells and their interactions with the niche (9). The niche, comprising three types of somatic cells—terminal filament (TF) cells, cap cells, and escort stem cells (ESCs) (10)—supports two to three self-renewing GSCs by providing niche-associated signals, including the Drosophila bone morphogenetic protein (BMP) homolog Decapentaplegic (DPP) (1115). DPP is produced by the niche cells, including cap cells and ESCs (16), and functions over a short range to repress GSC expression of bag-of-marbles (bam), a key factor that promotes differentiation (15, 17). Only GSCs within the niche show high DPP signal activation, repress bam expression, and maintain stem cell identity, whereas cystoblasts (CBs), the immediate daughters of GSC divisions, which are located outside the niche, show low DPP signal activation and consequently derepress bam expression and initiate differentiation. Although DPP can act as a long-range signal in other developmental contexts, its activity is spatially restricted in the germarium.

The epidermal growth factor receptor (EGFR) pathway is involved in cell fate specification through cell-cell interactions during animal development (18, 19). The EGFR pathway is activated upon binding of secreted, mature ligands to the receptor. The activated EGFR can potentially activate several downstream signaling cascades, including the phosphatidylinositol 3-kinase (PI3K), mitogen-activated protein kinase (MAPK), and phospholipase C–γ (PLC-γ) pathways (20). Stem cell tumor (stet) is a member of a family of Rhomboid intramembrane proteases in Drosophila whose function is required for the maturation of EGFR ligand; stet is required for proper interactions between early germ cells and somatic cells to maintain germline homeostasis in both males and females (21). Furthermore, EGFR signaling in somatic cells may organize a microenvironment required for proper male germline development (22).

Here, we show that in the Drosophila germarium, the GSC lineage—including the GSCs, which have traditionally been regarded as the passive signal-receiving entity within the niche—plays an active role in defining the spatial limit of the niche activity. We show that stet functions in the GSC lineage and is required for the maturation of multiple EGFR ligands, which initiate the RAS–RAF–MAPK kinase (MEK)–MAPK signaling cascade downstream of the EGFR in surrounding somatic cells. EGFR-MAPK signal activation in these somatic cells in turn suppresses expression of dally, which encodes the glypican required for DPP movement and stability. Consequently, only GSCs within the niche and next to the DPP source exhibit high downstream signal activation and are maintained as stem cells, whereas CBs located outside the niche show low signal activation and initiate differentiation. Thus, germline homeostasis is maintained through this reciprocal interaction.

Results

stet functions in the germ line to maintain germline homeostasis

Schulz et al. (21) previously showed that stet is required for proper interaction between early germ cells and somatic cells to maintain female germline homeostasis. In a candidate approach–based screen, we found that compromised EGFR signaling in ESCs and ECs (escort cells, the descendants of ESCs) resulted in the formation of ectopic GSC-like cells as observed in the stet mutant. GSCs are the anterior-most germ cells and can be identified by their direct contact with cap cells and the presence of an anterior spectrosome, a spherical form of fusome that is a germline-specific organelle and that functions in cyst formation (2325). CBs, the immediate daughters of GSCs, have a randomly positioned spectrosome, whereas differentiating germline cysts contain a branched fusome (9). The spectrosome and fusome are enriched with cytoskeletal proteins, including α-spectrin. Each wild-type germarium contains 5.3 ± 0.8 spectrosome-containing cells, including two to three GSCs situated next to two to three posteriorly located CBs, which in turn were situated next to more posteriorly located fusome-containing cysts (Fig. 1A). In contrast, germaria from animals expressing a stet null allele (stet1A3) (Fig. 1B and fig. S1) had more spectrosome-containing cells (14.1 ± 4.7, P < 0.01), consistent with a previous report (21). These phenotypes were completely reverted by restoring the germline expression of stet (Fig. 1C). These germaria had only 5.33 ± 1.26 (P < 0.01) spectrosome-containing cells. Virtually identical phenotypes were observed when stet function was specifically removed from the germ line (Fig. 1D). These data indicate that stet functions in the germ line to maintain germline homeostasis.

Fig. 1

EGFR-MAPK signaling functions in somatic cells to maintain germline homeostasis. Red, α-spectrin; blue, nuclei. (A) A wild-type (WT) germarium stained for α-spectrin, Vasa (green; to label germ cells), and nuclei. Spectrosome-containing cells (GSCs and CBs) were confined to the anterior region (n = 49 germaria). (B) A stet1A3 germarium with ectopic spectrosome-containing cells. Cells located in the posterior region contained a spectrosome (n = 76 germaria). (C) Germline expression of stet rescued the stet1A3 mutant phenotype. Spectrosome-containing cells were restricted to the anterior region (n = 100 germaria). (D) A stet1A3 germline mutant germarium (marked by loss of GFP signal) had ectopic spectrosome-containing cells (n = 100 germaria). (E) A germarium with a germ line triply mutant (marked by loss of GFP signal) for grk, spi, and krn had ectopic spectrosome-containing cells (n = 20 germaria). (F) Germline expression of mature Krn in the stet1A3 mutant rescued the ectopic spectrosome-containing cell phenotype (n = 105 germaria). (G and H) Germaria with compromised EGFR (G; n = 47 germaria) or pointed (H; n = 25 germaria) function specifically in ESCs and ECs had ectopic spectrosome-containing cells. (I) Restoration of MAPK signaling in stet1A3 germaria suppressed formation of ectopic spectrosome-containing cells (n = 20 germaria). Left, anterior. Scale bars, 10 μm.

stet has been implicated in the cleavage and, thus, maturation of EGFR ligands (2628). Of the four known ligands in Drosophila, Spitz, Keren, and Gurken are produced as membrane-bound precursors and require cleavage processing to generate secreted mature ligands, whereas Vein is produced as a secreted protein (18). Germaria with germ lines singly mutant for either keren or gurken or doubly mutant for gurken and spitz did not have ectopic spectrosome-containing cells (fig. S2, A to D). However, germaria with germ lines triply mutant for spitz, keren, and gurken had ectopic spectrosome-containing cells, reminiscent of the stet1A3 germaria (Fig. 1E). Consistent with the role of STET in the generation of mature ligands, overexpression of the mature forms of Spitz, Keren, or Gurken (thus bypassing cleavage processing) in the stet1A3 germ line largely reverted these phenotypes (Fig. 1F and fig. S2, E and F). Although stet1A3 germaria had 14.1 ± 4.7 spectrosome-containing cells, stet1A3 germaria overexpressing Spitz had 3.89 ± 1.13 (P < 0.01) spectrosome-containing cells, and stet1A3 germaria overexpressing Keren or Gurken had 4.47 ± 0.97 (P < 0.01) and 5.07 ± 0.98 (P < 0.01) spectrosome-containing cells, respectively. Additionally, germaria with germ lines mutant for star, the chaperone required for transport of proligands from endoplasmic reticulum to Golgi before cleavage, also had ectopic spectrosome-containing cells (fig. S2G). These results suggest that stet is involved in the maturation of multiple EGFR ligands in the germ line.

Somatic EGFR-RAS-RAF-MEK-MAPK signaling prevents the formation of ectopic spectrosome-containing cells

We then asked whether the somatic cells (ESCs and ECs) enclosing the early germ cells are the signal-receiving cells. Indeed, germaria with compromised EGFR function in ESCs and ECs had ectopic spectrosome-containing cells (9.47 ± 2.84, P < 0.01) resembling those observed in stet1A3 germaria (Fig. 1G). Upon ligand binding, activated EGFR can initiate three downstream signaling cascades, including the PI3K, MAPK, and PLC-γ pathways (20). Germaria with defects in the PI3K or PLC-γ pathway in ESCs and ECs did not exhibit ectopic spectrosome-containing cells (5.07 ± 0.92 for plc-γ1, 5.75 ± 1.22 for plc-γ2, and 5.35 ± 1.3 for PI3KRNAi) (fig. S3, A to D). In contrast, germaria with compromised RAS-RAF-MEK-MAPK pathway activity specifically in ESCs and ECs (referred to as EGFR-MAPK germaria) had ectopic spectrosome-containing cells (13.3 ± 5.6, P < 0.01 for ras85DRNAi, and 10.8 ± 3.14, P < 0.01 for mekRNAi), reminiscent of germaria with compromised EGFR activity in ESCs and ECs or lacking stet function in the germ line (Fig. 1H and fig. S3, E to G). Furthermore, restoration of MAPK signaling in ESCs and ECs in stet1A3 germaria partially suppressed the formation of ectopic spectrosome-containing cells (Fig. 1I). Together, these data show that EGFR functions in ESCs and ECs through canonical RAS-RAF-MEK-MAPK signaling and downstream of germline stet activity to control tissue homeostasis.

We examined the signaling activation in germaria by monitoring phosphorylated extracellular signal–regulated kinase (pERK) (29). In wild-type germaria, as reported previously (30), pERK was detected in somatic cells including cap cells, ESCs, and ECs, but not in germ cells (Fig. 2A). However, in stet1A3 germaria, pERK signaling was not detected in ESCs or ECs (Fig. 2B). In wild-type germaria, ESCs directly enclosing GSCs often showed high signal activation, as indicated by pERK staining (Fig. 2A and fig. S4), in line with the notion that STET-mediated generation of EGFR ligands in GSCs is involved in the activation of EGFR signaling in ESCs to suppress formation of ectopic spectrosome-containing cells. Supporting this notion, restoration of stet function in GSCs, but not in CBs and germline cysts, fully reverted the ectopic spectrosome-containing cell phenotype in a stet1A3 background. These results are consistent with the notion that removal of stet function from GSCs leads to the formation of ectopic spectrosome-containing cells in stet1A3 germaria.

Fig. 2

stet functions in GSCs to maintain germline homeostasis. Red, α-spectrin. (A and B) pERK (green) was detected in somatic cells in WT germaria (A; n = 120 germaria) but not in stet1A3 germaria (B; n = 120 germaria). High amounts of pERK were often detected in ESCs (white arrows). (C) A germarium with one GSC mutant for stet (marked by the lack of GFP signal) exhibited non–cell autonomous effects. The GFP (stet) GSC was located out of presented focal plane. WT GFP+ cells located in the anterior region and outside the niche (indicated by white arrows) contained a spectrosome, whereas GFP+ cells located in the posterior region (arrowheads) had a fusome (n = 30 germaria). (D) A stet1A3 germarium with stet expression restored in CBs and germline cysts had more spectrosome-containing cells (compare with Fig. 1A; n = 115 germaria). Blue, nuclei. (E) Restoration of stet in stet1A3 germ line in GSCs, CBs, and germline cysts prevented the formation of ectopic spectrosome-containing cells (n = 100 germaria). Left, anterior. Scale bars, 10 μm.

Surprisingly, germaria with one mutant GSC clone showed non–cell autonomous defects. The green fluorescent protein–positive (GFP+) (stet+) cells located outside the niche and in the midway of the germarium also contain a spectrosome, indicating a GSC- or a CB-like stage (Fig. 2C and fig. S5), which suggests that loss of stet function in GFP (stet) cells resulted in ectopic (resulting in ectopic DPP production outside the niche) or expanded (resulting in DPP functioning over a longer range) niche activity. To confirm the role of stet function in GSCs, we conducted rescue experiments in stet1A3 germaria by restoring stet function in GSCs or CBs. To restore stet function in CBs and germline cysts, we used a Pbam-stet transgene that expresses the stet gene under the control of the bam promoter and recapitulates bam expression (15, 17) in CBs and cysts. This transgene largely reverted the phenotype of ectopic spectrosome-containing cells and restored the female sterile phenotype associated with the stet1A3 mutant [7.34 ± 2.0 spectrosome-containing cells; P < 0.01] (Fig. 2D and fig. S6). However, many germaria still had more spectrosome-containing cells compared with their wild-type counterparts (5.3 ± 0.8) (fig. S6, C, D, and G). Similar results were obtained when stet function was restored in CBs and germline cysts by a Pbam-Gal4 driver with the UAS/GAL system (fig. S6, E to G). To restore stet function in GSCs, CBs, and germline cysts, we used a nos-Gal4 driver, which is expressed in all germ cells, to drive expression of UASp-stet in a stet background. Consistent with a role for stet in GSCs, restoration of stet expression in GSCs, CBs, and germline cysts prevented the formation of ectopic spectrosome-containing cells in stet1A3 germaria. These germaria had 5.33 ± 1.26 (P < 0.01) spectrosome-containing cells, whereas stet mutant germaria had 14.1 ± 4.7 (Fig. 2E and fig. S6). Together, these results demonstrate that although stet also functions in the germ line downstream of GSCs, its function in GSCs is critical in preventing the formation of ectopic spectrosome-containing cells.

stet mutant germaria exhibit ectopic DPP activity

What is the role of EGFR-MAPK signal activation in ESCs and ECs? We examined the fate of ectopic spectrosome-containing cells in stet1A3 germaria. DPP, the primary niche-associated signal, is produced by cap cells and ESCs and functions as a short-range gradient (15, 16). As a result of restricted DPP function, each wild-type germarium, which has one niche, supports only two to three GSCs that receive high DPP signal and suppress the expression of bam (31). CBs outside the niche receive low DPP signal, derepress bam expression, and initiate differentiation. In wild-type germaria, Pbam-gfp [the gfp gene under the control of the bam promoter, which serves as a bam transcription reporter (32)] was not detected in GSCs within the niche but expressed in CBs and differentiating cysts (Fig. 3A). In contrast, in stet1A3 germaria, many ectopic spectrosome-containing cells, especially those located in the anterior region, expressed undetectable or low amounts of Pbam-gfp, whereas those spectrosome-containing cells located more posteriorly had a GFP signal (Fig. 3B). BAM-GFP (32) (which is under the control of the bam promoter) is expressed in wild-type CBs and early cysts but was not detected in ectopic spectrosome-containing cells (Fig. 3, C and D). We then examined DPP signal activation, which is responsible for bam silencing in GSCs in ectopic spectrosome-containing cells. In wild-type germaria, the expression of Dad-lacZ (33), a reporter of DPP signal activation, was largely confined to the anterior tip of the germarium including GSCs within the niche (Fig. 3E). However, many ectopic spectrosome-containing cells in stet1A3 germaria exhibited Dad-lacZ expression, indicative of DPP signal activation (Fig. 3F). We next examined phosphorylated Mad (pMad) (34), which requires strong signal activation and serves as a direct readout of DPP signal activation. In wild-type germaria, pMad was high in GSCs and undetectable in CBs (Fig. 3G and fig. S7A). Variable pMad amounts were also observed in posterior regions of germaria (Fig. 3G). In stet1A3 germaria, high amounts of pMad were detected in GSCs within the niche; however, lower amounts of pMad were also detected in many spectrosome-containing cells outside the niche position, suggesting the existence of a pMad gradient and possibly expanded DPP activity (Fig. 3H and fig. S7, B and C). We also noted that in some germaria, the germ cells located at the posterior regions exhibited high amounts of pMad (fig. S7C). Together, these results suggest that many ectopic spectrosome-containing cells were GSC-like, hence further indicating that ectopic DPP signal activation occurs in cells outside the normal niche in stet1A3 germaria. Supporting the role of the DPP pathway in formation of ectopic GSC-like cells in this background, removal of one copy of dpp or thickvein (tkv, the type I receptor of DPP) suppressed the formation of ectopic GSC-like cells in a stet1A3 background (Fig. 3, I and J, and fig. S8, A and B). Similar effects were observed when one copy of dpp was removed from animals in which EGFR-MAPK activity was specifically compromised in ESCs and ECs (fig. S8C).

Fig. 3

stet1A3 germaria exhibit expanded DPP signaling activity. Red, α-spectrin. (A and B) Pbam-gfp (green) was expressed in CBs and germline cysts in WT germaria (A; n = 100 germaria) but not in most ectopic spectrosome-containing cells in stet1A3 germaria (B; n = 40 germaria). (C and D) BAM-GFP (green) was expressed in CBs and early cysts in WT germaria (C; n > 100 germaria) but not in ectopic spectrosome-containing cells in stet1A3 germaria (D; n > 50 germaria). (E and F) Dad-lacZ (green) was largely restricted to GSCs in WT germaria (E; n > 100 germaria) but was expressed in many ectopic spectrosome-containing cells in stet1A3 germaria (F; n = 50 germaria). (G) In WT germaria, pMad (green) was detected in high amounts in GSCs but not in CBs (n = 150 germaria). (H) In stet1A3 germaria, pMad was found not only in GSCs within the niche but also in many spectrosome-containing cells outside the niche in lower amounts (n = 70 germaria). (I) A stet1A3 germarium showing ectopic spectrosome-containing cells (n ≥ 100 germaria). (J) Removal of one copy of dpp (dpphr56) suppressed the formation of ectopic spectrosome-containing cells in stet1A3 germaria (n = 55 germaria). Green, Vasa. (K) In a WT germarium, dpp transcripts were present in cap cells (n > 100 germaria) as well as in ESCs (white arrow in K′). (L) In a stet1A3 germarium, dpp transcripts were present only in cap cells (n = 110 germaria). Blue, LamC to label cap cells; green, dpp mRNA. Left, anterior. Scale bars, 10 μm.

Ectopic DPP expression outside the niche (14), expansion of the niche (caused by ectopic cap cell formation), or expansion of niche activity (in which DPP functions over a longer range) can all lead to formation of ectopic GSC-like cells in stet1A3 and EGFR-MAPK germaria. To distinguish between these possibilities, we examined dpp expression in these germaria. In wild-type germaria, as we previously reported (16), dpp transcripts were consistently detected in cap cells and often in ESCs (Fig. 3K and fig. S9). Consistent with this expression pattern, compromising dpp function specifically in TF and cap cells [with a combination of a tissue-specific driver and RNA interference (RNAi)–mediated knockdown] depleted GSCs, with more than 80% of the germaria showing precocious GSC loss (fig. S10, A and B). In contrast, removal of dpp function from TF cells, cap cells, ESCs, and ECs led to a complete loss of GSCs in all germaria examined (fig. S10, C and D). In stet1A3 germaria, dpp transcripts were also detected in cap cells (Fig. 3L and fig. S9D). These data show that dpp expression was not ectopically induced outside the niche in stet1A3 germaria. Supporting this notion, quantitative polymerase chain reaction (QPCR) analyses of dpp transcripts of samples from these somatic cells (including ESCs and ECs but not cap cells) isolated from wild-type and EGFR-MAPK germaria showed that dpp transcription was not increased in EGFR-MAPK germaria (Fig. 4). These results suggest that formation of ectopic GSC-like cells in EGFR-MAPK germaria is not due to ectopic dpp expression outside the niche. Furthermore, stet1A3 germaria had similar numbers of cap cells compared to wild-type germaria, ruling out the possibility that ectopic cap cell formation caused formation of ectopic GSC-like cells (Fig. 3, K and L) (35, 36).

Fig. 4

The abundance of dally but not dlp or dpp is increased in somatic cells from germaria with compromised EGFR function in ESCs and ECs. (A) A germarium showing the restriction of c587-Gal4 expression to ESCs and ECs. Green, GFP; red, Fas III to label prefollicular somatic cells. Left, anterior. Scale bar, 10 μm. n > 100 germaria. (B) Control cell-sorting profile for WT ovaries (GFP-negative). (C) Cell-sorting profile for c587-Gal4/UAS-gfp ovaries showing GFP-positive and PI-negative cell population (green). (D) QPCR analyses of dpp, dally, and dlp transcripts from samples of sorted cells, done in triplicate on each of three independent RNA samples. Data are reported as fold change between WT samples and EGFR knockdown samples. In EGFR-MAPK germaria, the abundance of dally transcript was increased, whereas that of dpp and dlp transcripts was unchanged.

The alternative possibility is that niche activity is expanded so that DPP, produced by cap cells albeit at a similar amount, now functions over a longer than normal range, possibly through transport or stability (or both) outside the niche in both stet1A3 and EGFR-MAPK germaria. This is consistent with the non–cell autonomous defects observed in stet1A3 germaria with one mutant GSC clone (Fig. 2C and fig. S5) and the greater number of pMad-positive cells detected outside the niche in stet1A3 mutant germaria (Fig. 3H and fig. S7, B and C). Together, our data suggest that in wild-type germaria, somatic EGFR-MAPK signaling acts to spatially restrict DPP function within the niche, possibly by restricting its transport or stability outside the niche. Consequently, only GSCs within the niche exhibit high downstream signal activation and are maintained as stem cells, whereas CBs outside the niche show low signal reception and initiate differentiation.

EGFR-MAPK signaling in ESCs and ECs suppresses dally transcription

To explore how DPP transport or stability may be affected by loss of stet in the germ line or loss of EGFR-MAPK signaling in somatic cells, we focused our attention on Drosophila glypicans, which have an established role in DPP transport and stability (3739). Two Drosophila glypicans, Dally and Dally-like (Dlp), are required for transport of DPP and can compensate for the loss of the other during wing development (37). Removal of dally, but not dlp, from ESCs and ECs partially suppressed the formation of ectopic spectrosome-containing cells in stet1A3 germaria (Fig. 5, A to D). Most stet1A3 germaria (95%) had ectopic spectrosome-containing cells outside the niche (Fig. 5B). When dally function was removed from ESCs and ECs in a stet1A3 background, 19% of these germaria had ectopic spectrosome-containing cells and 81% of these germaria had fusome-containing cysts outside the niche (Fig. 5C). When dlp function was further compromised in ESCs and ECs in a stet1A3 background, most germaria (98%) also had ectopic spectrosome-containing cells, as observed in stet1A3 germaria (Fig. 5D). We then examined DPP signal activation, which serves as a readout for niche activity. The expression domain of Dad-lacZ, which is largely restricted to the niche position in wild-type germaria but was expanded in stet1A3 germaria, was largely restricted within the niche position when dally but not dlp function was removed in ESCs and ECs of stet1A3 germaria (Fig. 5, E to H). Together with the well-established role of Dally on DPP transport and stability, these results suggest that increased Dally activity (in ESCs and ECs) in stet1A3 germaria is responsible for the expanded DPP (niche) activity by promoting the transport or stability of DPP outside the niche, and indicate that in wild-type germaria, EGFR-MAPK signaling in ESCs and ECs spatially restricts DPP function by antagonizing Dally activity.

Fig. 5

Compromising dally but not dlp function in ESCs and ECs suppresses stet1A3 mutant phenotypes. Red, α-spectrin; blue, nuclei. (A) A WT germarium showing spectrosome-containing cells restricted to the anterior region (n > 100 germaria). (B) A stet1A3 germarium exhibiting ectopic spectrosome-containing cells (n = 76 germaria). (C and D) Removal of dally (C; n = 96 germaria) but not dlp (D; n = 86 germaria) from ESCs and ECs suppressed the stet1A3 phenotype. (E and F) Dad-lacZ expression (green) was restricted to the anterior region in a WT germarium (E; n > 100 germaria) and expanded in a stet1A3 germarium (F; n = 130 germaria). (G and H) Dad-lacZ expression was largely confined in the anterior region in a stet background when dally (G; n = 50 germaria) but not dlp (H; n = 45 germaria) was removed from ESCs and ECs. Left, anterior. Scale bars, 10 μm.

How does EGFR signaling antagonize Dally activity? Because functional removal of pointed (which encodes the downstream transcriptional effector of the EGFR-MAPK pathway) from ESCs and ECs also led to expanded DPP activity (Fig. 1H and fig. S8, D to F), we favor the possibility that EGFR-MAPK signaling antagonizes Dally at the transcription level. We examined dally expression in wild-type and stet1A3 germaria. dally was expressed at variable amounts in wild-type germaria. Although dally transcripts were evident in prefollicle cells in every germarium examined, its expression in the niche was detectable in about 50% of germaria examined. Among germaria showing dally expression, dally messenger RNA (mRNA) was detected in only one to two cap cells per germarium, but not in ESCs or ECs surrounding early germ cells, which may be due to low abundance (Fig. 6A and fig. S11). However, in stet1A3 and EGFR-MAPK germaria, dally transcripts were detected in ESCs and ECs (Fig. 6, B to E, and fig. S11D), supporting the notion that stet-mediated EGFR signal activation in ESCs and ECs functions to suppress dally transcription to limit DPP transport or stability (or both) outside the niche. To confirm this notion, we carried out QPCR analyses for dally transcripts of samples from ESCs and ECs isolated from wild-type and EGFR-MAPK germaria. Indeed, although dlp transcript abundance was similar in wild-type and EGFR-MAPK germaria, dally transcript abundance was increased in ESCs and ECs of EGFR-MAPK germaria (Fig. 4D). Together, these results show that somatic EGFR-MAPK signal activation spatially restricts DPP function by transcriptional regulation of dally, the molecule required for DPP transport and stability.

Fig. 6

Somatic EGFR-MAPK signaling inhibits dally expression in ESCs and ECs. (A to E) Green, α-spectrin; red, dally mRNA. (A) In a WT germarium, dally expression was detected in a cap cell [white arrow, (A′)] and prefollicular cells (white arrows), but not in ESCs or ECs (n = 300 germaria). (B to E) Four different sections of a stet1A3 germarium showing the presence of ectopic dally transcripts in ESCs and ECs (white arrowheads). White arrows mark the anterior position of prefollicular cells. (F to J) Red, α-spectrin; green, Dad-lacZ. (F) A germarium with ectopic dally expression in ESCs and ECs had ectopic spectrosome-containing cells. Dad-lacZ was largely restricted to the anterior region in a WT germarium (G; n = 26 germaria) but was expressed in ectopic spectrosome-containing cells in a germarium with ectopic dally expression in ESCs and ECs (H; n = 82 germaria). (I) In addition to GSCs within the niche, pMad (green) was also detected in some spectrosome-containing cells outside the niche in a germarium with ectopic dally expression in ESCs and ECs (n = 80 germaria). (J) A germarium with compromised dally function in ESCs and ECs showed restricted Dad-lacZ (green) expression (n = 30 germaria). Left, anterior. Scale bars, 10 μm.

Discussion

Our data reveal that crosstalk between the GSC lineage and somatic cells defines the spatial limits of DPP action and therefore the extent of the GSC niche in the Drosophila ovary (fig. S12). In wild-type germaria, stet functions in the GSC lineage (including in GSCs) to generate multiple EGFR ligands, which could bind to receptors on ESCs and ECs to activate EGFR-MAPK signaling that represses dally expression. As a result, the expression of dally in ESCs and ECs is low, and DPP is not stabilized outside the niche. Hence, only GSCs within the niche and next to the DPP source, produced by cap cells and ESCs, can activate downstream signaling, suppress bam expression, and maintain a stem cell identity; in contrast, CBs outside the niche receive low DPP signal, derepress bam expression, and initiate differentiation. In stet mutant germaria, somatic EGFR-MAPK signaling is abolished as a result of loss of stet function in the germ line. Consequently, dally is ectopically expressed in ESCs and ECs and can now transport or stabilize DPP outside the niche. As a result, DPP can act over a longer than normal range across the germarium (thus expanding niche activity), which can now support the formation of more GSC-like cells outside the normal niche.

Consistent with our model, when dally was ectopically expressed in ESCs and ECs, many germaria (76%) had ectopic spectrosome-containing cells with Dad-lacZ expression outside the niche (Fig. 6, F to H). Although high amounts of pMad were detected in GSCs, low amounts of pMad were also detected in some spectrosome-containing cells outside the niche in 30% of germaria examined (Fig. 6I). Together, these data suggest that ectopic dally expression causes expanded DPP function (and consequently, expanded niche activity) in germaria, which is reminiscent of stet1A3 or EGFR-MAPK germaria. Conversely, compromising dally function in ESCs and ECs resulted in fewer spectrosome-containing cells in germaria and a more restricted expression domain of Dad-lacZ (Fig. 6J).

Several lines of evidence indicate that GSCs are involved in the spatial restriction of DPP function within the niche. First, ESCs directly enclosing GSCs often showed strong EGFR-MAPK signal activation (Fig. 2A and fig. S4), suggesting that GSCs may be involved in the generation of EGFR ligands. Second, only when stet function was removed from GSCs, CBs, and germline cysts did these germaria exhibit ectopic spectrosome-containing cells outside the niche (Fig. 1D), indicating a role for stet function in GSCs. Third, although restoration of stet function from CBs and germline cysts with a Pbam-stet transgene in stet1A3 germaria did not fully suppress the ectopic spectrosome-containing cell phenotype, putting back stet function in GSCs, CBs, and germline cysts in a stet mutant background fully reverted the stet1A3 mutant phenotype (Fig. 2, D and E, and fig. S6). The bam promoter used in our study appeared to recapitulate the endogenous bam expression pattern; however, because of the lack of proper methods to visualize STET distribution in vivo, we cannot formally exclude the possibility that the Pbam-stet transgene may not generate the same functional amount of STET in CBs as endogenous stet at the appropriate rate, which led to the incomplete rescue in a stet background. Nevertheless, these data support the notion that stet function in GSCs is important in maintaining germline homeostasis.

As previously reported (21), ESCs and ECs in stet1A3 germaria do not enclose early germ cells, including GSCs, with cellular extensions (fig. S13, A and B). It is possible that in wild-type germaria, these cellular extensions of ESCs and ECs function as a gate to prevent the movement (diffusion) of signaling molecules including DPP outside the niche. Disruption of these cellular extensions may result in longer than normal signaling capacity of signaling molecules other than DPP that lead to these phenotypes. Alternatively, as shown in the Drosophila male germ line (22), lack of these cellular extensions may disrupt the necessary microenvironment required for the differentiation of CBs leading to these phenotypes. However, several lines of evidence indicate that these phenotypes observed in stet1A3 and EGFR-MAPK germaria are associated with expanded DPP function caused by defective EGFR-MAPK signaling in ESCs and ECs.

First, in wild-type germaria, pMad is detected in GSCs (including dividing GSCs) (Fig. 3G and fig. S7A), but not in CBs just outside the niche. However, in stet1A3 germarium, high amounts of pMad were detected in GSCs; low amounts of pMad were also detected in many spectrosome-containing cells outside the niche, suggesting expanded DPP activity (Fig. 3H and fig. S7, B and C). Second, Dally is also involved in the signaling activity of other signaling molecules, including hedgehog (Hh) and wingless (Wg), both of which are also expressed in cap cells (40, 41). High dally expression in stet1A3 germaria may result in ectopic Hh and Wg signaling. However, ectopic Hh (fig. S14A) or Wg (fig. S14B) expression in germaria did not result in similar phenotypes (41). Although we do not exclude the possibility that ectopic signaling (through Hh, Wg, or another pathway) in stet1A3 or EGFR-MAPK germaria may contribute to these phenotypes, these observations suggest that increased Hh or Wg signaling caused by ectopic dally expression in these backgrounds may not be the major cause of these observed phenotypes. Third, removing one copy of dpp or tkv suppressed these phenotypes in stet1A3 or EGFR-MAPK germaria, suggesting that these phenotypes are specifically associated with ectopic DPP signaling (Fig. 3, I and J, and fig. S8, A to C). Furthermore, although most stet1A3 germaria (90.5%) had ectopic spectrosome-containing cells outside the niche, many germaria (70%) contained fusome-containing cysts outside the niche when DPP function was also removed from the cap cells in stet1A3 germaria after the formation of ectopic spectrosome-containing cells (fig. S15). These data suggest that formation of ectopic spectrosome-containing cells is associated with increased DPP signaling and that stet1A3 mutant germ cells can undergo differentiation upon withdrawal of DPP signaling (thus implying that their arrested differentiation was due to increased DPP signal reception and not to defects in the differentiation process). Fourth, removing dally function in stet1A3 germaria partially suppressed the formation of ectopic spectrosome-containing cells, in line with Dally’s known role in the transport and stability of DPP (Fig. 5, A to C). Lastly, other germline mutants that cause tumors, such as bam, which also exhibit defective cellular extensions of ESCs and ECs toward early germ cells as observed in stet1A3 germaria, did not show expanded DPP activity (fig. S13, C to I) (42), suggesting that defective cellular extensions per se is not sufficient for the expanded DPP (niche) activity observed in stet1A3 or EGFR-MAPK germaria. However, we cannot rule out the possibility that defective cellular extensions also contribute to observed phenotypes in a stet1A3 background.

Our results show that expanded DPP function through increased dally activity is associated with the formation of ectopic spectrosome-containing cells in a stet1A3 background. We also noted that stet1A3 germaria exhibited stronger phenotypes than dally-overexpressing germaria (compare Fig. 1B and Fig. 6F) and compromising dally function in stet1A3 germarium only partially reverted stet1A3 mutant phenotypes (Fig. 5G), suggesting that additional mechanisms downstream of EGFR-MAPK signaling other than ectopic dally expression are likely involved. In the Drosophila male germ line, downstream effectors of EGFR signaling in somatic cells function through cellular extensions of somatic cells toward early germ cells to organize a microenvironment necessary for proper germ cell development (22). Although our data suggest that the defective cellular extension of ESCs and ECs may not be a primary cause of observed phenotypes in stet1A3 germaria, we cannot rule out the possibility that the defective cellular extensions also contributed to the observed phenotypes. It will be interesting to see whether a similar mechanism also functions in the female germ line and whether this mechanism is disrupted in stet1A3 mutant germaria. Furthermore, although germaria ectopically overexpressing Hh or Wg did not phenocopy stet1A3 germaria, our results do not rule out the possibility that, in addition to expanded DPP activity, increased signal activity through Hh, Wg, or another pathway may also contribute to the observed phenotypes in stet1A3 germaria.

The spatial extent of the GSC niche must be tightly controlled to balance stem cell self-renewal with differentiation to maintain tissue homeostasis. Reduced niche activity often associates with tissue dysfunction, whereas expanded niche activity can result in excessive stem cell production, which is also deleterious. Our results demonstrate that the Drosophila female GSC lineage plays an active role in the spatial restriction of the Drosophila stem cell niche to maintain germline homeostasis. Given the conserved mode of dynamic interaction between stem cells and their niche across species, it will be interesting to see whether other stem cell systems also function similarly to restrict their niche.

Materials and Methods

Drosophila stocks

Information about strains used in this study was described in the text or in FlyBase. y1w1118, bamΔ86, dpphr56, dlp2, egfrf24, grkHF, grk2B6, krn27-7-B, plc-γ1, plc-γ2, spi2A14, starIIN, tkv906, BAM-GFP, dad-lacZ1883, Fax-GFP, Pbam-gfp, PZ(2)0078, PZ(2)1444, tubP-Gal80TS, UASp-gfp, UAS-gfp, UASp-s-grk, UASp-s-spi, UASp-stet, UAS-wg, UAS-hh, UAS-dally, UAS-phlgof, c587-Gal4, Pbam-Gal4, nos-Gal4, 0078GalER, egfrRNAi [National Institute of Genetics (NIG), Japan], dallyRNAi [Vienna Drosophila RNAi Center (VDRC)], dlpRNAi (VDRC), ras85DRNAi (VDRC), rafRNAi (VDRC), pntRNAi (VDRC), mekRNAi (VDRC), dppRNAi (Bloomington), UASp-skrn-c-myc, and Pbam-stet transgenes were generated by standard germline transformation.

78Gal4 was converted from PZ(2)0078, which is expressed in TF and cap cells (43) (fig. S9A). 1444Gal4 was converted from PZ(2)1444, which is expressed in some TF cells, cap cells, ESCs, and ECs (44) (fig. S9C).

All crosses were maintained at 25°C, and progenies with proper genotypes were aged at 28°C (unless stated elsewhere) for 2 to 6 days before analyses. Mutant clones were generated as previously described (16).

RNAi-mediated knockdown in ESCs and ECs

For c587-Gal4–mediated knockdown, c587-Gal4 is expressed in ESCs and ECs, but not in cap cells (15). The crosses were set up and maintained at 25°C. Progenies with proper genotypes were aged at 28°C before analyses.

For EGFR knockdown in ESCs and ECs, RNAi-mediated knockdown of EGFR was most effective in animals lacking one copy of egfrf24 function. This combination was used throughout this study.

Overexpression in ESCs and ECs

For UAS-dally overexpression, the crosses were set up at 25°C, and progenies with proper genotypes were aged at 31°C for 4 days before analyses.

For UAS-hh and UAS-wg overexpression, tubP-Gal80TS was used to suppress c587-Gal4 activity during early development. Crosses were set up and maintained at 18°C, and progenies with proper genotypes were collected and aged at 30°C for 7 days before analyses.

stet mutant generation and rescue experiment

stet1A3 deletion allele was generated through P-element–mediated imprecise excision. This allele deletes a common region of two annotated polypeptides of the stet locus (STET-PA and STET-PB), which encodes seven transmembrane domains necessary for enzymatic activity, and rob162A, the intragenic gene with unknown function. Thus, this allele is believed to be a functional null allele of stet. stet1A3 homozygotes are viable but sterile. The phenotype of female sterility was reverted by the introduction of stet in the germ line with nos-Gal4 driver, suggesting that it is associated with stet loss of function.

For the rescue experiment, nos-Gal4 driver (which is expressed in all germ cells) was used to drive UASp-stet expression in the germ line in GSCs, CBs, and germline cysts. We noticed that UASp-gfp driven by Pbam-Gal4 showed a slight delay in expression (which is likely due to the nature of the UAS/Gal4 system itself) compared with endogenous bam expression, whereas Pbam-gfp transgene faithfully recapitulates the endogenous bam expression pattern as reported previously (32). Accordingly, the introduction of UASp-stet by Pbam-Gal4 in the stet1A3 mutant did not effectively revert the formation of ectopic GSC-like cells (fig. S6). To introduce stet in the germ line from CBs and their descendants, we generated a Pbam-stet transgene. For this transgene, the 1-kb bam promoter region, which faithfully recapitulates endogenous bam expression pattern (32), was amplified from genomic DNA and verified by DNA sequencing. This promoter was placed 5′ of the stet coding sequencing and subcloned into CaSpeR-4 vector for germline transformation. Consistent with its expression without a delay (as endogenous bam gene), the introduction of Pbam-stet transgene in the stet1A3 mutant showed better rescue ability compared to Pbam-Gal4 (fig. S6).

0078GalER-mediated dppRNAi knockdown

Females with the genotypes dppRNAi-stet1A3/stet1A3 and 0078GalER/+;dppRNAi-stet1A3/stet1A3 were collected and aged for 4 days at room temperature (22°C) to develop the stet1A3 mutant phenotype and subsequently subjected to 17-β-estradiol treatment [17-β-estradiol (260 mg/ml) in 100% dimethyl sulfoxide was added to wet yeast paste] to activate the 0078GalER driver (45) for an additional 4 days before analyses.

Fluorescence-activated cell sorting

Females cultured at room temperature with the genotypes c587-Gal4/+;UAS-gfp/+ and c587-Gal4/+;UAS-gfp/egfrf24;egfrRNAi/+ were dissected in Schneider’s medium supplemented with 10% fetal bovine serum. The transparent part of the ovaries (including germaria and early-stage follicles) was retained, rinsed with phosphate-buffered saline (PBS), and dissociated with 0.5% trypsin plus collagenase (2.5 mg/ml; Invitrogen) in PBS (46). The cell suspensions were filtered twice through a 40-μm nylon mesh. Cells were collected and resuspended in 2 ml of PBS (~107 cells/ml) with propidium iodide (PI) (10 μg/ml), incubated for 15 min at room temperature, and sorted with a FACSVantage machine (Becton-Dickinson) by gating GFP-positive and PI-negative cells with the exclusion mode.

Synthesis of complementary DNA and real-time QPCR analysis

RNAs were extracted from sorted GFP+ cells (PicoPure RNA Extraction kit) and subjected to complementary DNA (cDNA) synthesis (SuperScript III First-Strand Synthesis System, Invitrogen) according to the manufacturer’s instructions. Oligo(dT)20 primer was used for cDNA synthesis. QPCR was carried out by means of a 7900HT Fast Real-Time PCR system (Applied Biosystems) according to the manufacturer’s manual with primers for Actin5C as an internal control.

Immunohistochemistry

Collection, fixation, antibody staining (except pMad staining), and fluorescent RNA in situ hybridization (for dpp and dally) of ovaries were carried out as previously described (47). For pMad staining, ovaries were fixed in 3.7% formaldehyde in PBT (PBS plus 0.1% Triton X-100) for 20 min, washed with PBT for 60 min, and blocked with 5% normal goat serum (NGS) in PBT before the addition of antibody against pMad (diluted in 5% NGS in PBT). TSA Cyanine 3 system (Perkin Elmer) was used to develop the dally in situ signal for Fig. 6, and TSA Fluorescein system was used to develop the dally in situ signal for fig. S10.

The following primary antibodies were used: mouse antibodies against α-spectrin [3A9, 1:30, Developmental Studies Hybridoma Bank (DSHB)], LamC (LC28.26, 1:25, DSHB), and Fas III (DSHB); rabbit antibodies against pMad (1:500, a gift from P. ten Dijke), GFP (1:2000, Molecular Probes), α-spectrin (48) (1:2000), phosphorylated ERK1/2 (1:200, Cell Signaling), and β-galactosidase (1:3000, Cappel); and guinea pig antibody against Vasa (1:3000, a gift from T. Kai). Alexa Fluor 488–, Alexa Fluor 555–, or Alexa Fluor 633–conjugated goat secondary antibodies against mouse, rabbit, and guinea pig (1:500, Molecular Probes) were used to detect the primary antibodies. TO-PRO-3 (Invitrogen) was used for DNA staining. Single-section images were taken with a Zeiss LSM510 META confocal microscope and processed with Adobe Photoshop. The following primers were used to amplify the DNA template for dally RNA in situ: 5′-CGTGTGTGCCAGTGTGTGT-3′ (forward) and 5′-TAATACGACTCACTATAGGGCGCCGCCTGTGTATCTATG-3′ (reverse); primers for the sense probe were 5′-ATTTAGGTGACACTATAGAACGTGTGTGCCAGTGTGTGT-3′ (forward) and 5′-CGCCGCCTGTGTATCTATG-3′ (reverse).

Statistical analyses of spectrosome-containing cells

For statistical analyses of spectrosome-containing cells in the study, numbers of spectrosomes were counted from randomly selected germaria under a fluorescence microscope, and P values were determined by Kruskal-Wallis test with a post-test (Dunns: compare all pairs of columns), representing probability of difference between sample populations, where a P < 0.01 threshold is considered a significant difference. The data analysis was done using Microsoft Office Excel 2007 and GraphPad Prism 5.0.

Acknowledgments

Acknowledgments: We are grateful to P. ten Dijke, C. Ghiglione, T. Kai, X. Lin, J. A. McDonald, D. McKearin, D. Montell, N. Perrimon, P. Rorth, T. Tabata, T. Xie, Y. M. Yamashita, DSHB, Bloomington Stock Center, NIG-FLY Stock Center, Szeged Stock Center, and VDRC for reagents. We thank the TRiP at Harvard Medical School (NIH National Institute of General Medical Sciences R01-GM084947) for providing transgenic RNAi fly stocks used in this study. We acknowledge Z. Liu for his technical support in pMad staining. We also thank B. Chia, S. Cohen, and A. Spradling for comments and suggestions and B. Chia for critical reading of the manuscript. Funding: This work was supported by Temasek Life Sciences Laboratory and Singapore Millennium Foundation. Author contributions: Y.C. designed the experiments, analyzed the data together with T.M.L. and M.L., and wrote the manuscript. T.M.L. contributed to the design of the experiments, was involved in the discussion of results, and helped in writing the manuscript. M.L. was involved in experimental design and execution, analyzed the results, and helped in manuscript preparation. Competing interests: The authors declare that they have no competing interests. The following fly strains require a materials transfer agreement (MTA) from VDRC: dallyRNAi, dlpRNAi, ras85DRNAi, rafRNAi, pntRNAi, and mekRNAi. The egfrRNAi fly strain requires an MTA from NIG, Japan.

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/3/132/ra57/DC1

Fig. S1. The generation of the stet1A3 mutant.

Fig. S2. stet functions in the germ line and is involved in the generation of multiple EGFR ligands.

Fig. S3. EGFR functions through the MAPK but not the PI3K or PLC-γ pathway.

Fig. S4. pERK is expressed in ESCs and ECs.

Fig. S5. stet1A3 mutant germarium exhibits non–cell autonomous defects.

Fig. S6. stet is required in GSCs for germline homeostasis.

Fig. S7. pMad expression domain is expanded in stet1A3 germaria.

Fig. S8. Dad-lacZ expression domain is expanded in EGFR-MAPK germaria.

Fig. S9. dpp transcripts are detected in wild-type and stet1A3 germaria.

Fig. S10. DPP functions in the niche.

Fig. S11. dally mRNA is detected in wild-type and EGFR-MAPK germaria.

Fig. S12. A model.

Fig. S13. Defective cellular extension is not sufficient for expanded DPP function outside the niche.

Fig. S14. Germaria with ectopic Hh or Wg expression do not exhibit the formation of ectopic spectrosome-containing cells.

Fig. S15. Conditional removal of DPP from the niche suppresses the stet1A3 mutant phenotype.

References and Notes

View Abstract

Navigate This Article