GDNF and GFR -1 Are Components of the Axolotl Pronephric Duct Guidance System

Transcription

1 Developmental Biology 228, (2000) doi: /dbio , available online at on GDNF and GFR -1 Are Components of the Axolotl Pronephric Duct Guidance System Julie Drawbridge, 1 Christopher M. Meighan, and Elisa A. Mitchell Department of Biology, Rider University, Lawrenceville, New Jersey In mammals, secretion of GDNF by the metanephrogenic mesenchyme is essential for branching morphogenesis of the ureteric bud and, thus, metanephric development. However, the expression pattern of GDNF and its receptor complex the GPI-linked ligand-binding protein, GFR -1, and the Ret tyrosine kinase signaling protein indicates that it could operate at early steps in kidney development as well. Furthermore, the developing nephric systems of fish and amphibian embryos express components of the GDNF signaling system even though they do not make a metanephros. We provide evidence that GDNF signaling through GFR -1 is sufficient to direct pathfinding of migrating pronephric duct cells in axolotl embryos by: (1) demonstrating that application of soluble GFR -1 to an embryo lacking all GPI-linked proteins rescues PND migration in a dose-dependent fashion, (2) showing that application of excess soluble GFR -1 to a normal embryo inhibits migration and that inhibition is dependent upon GDNF-binding activity, and (3) showing that the PND will migrate toward a GDNF-soaked bead in vivo, but will fail to migrate when GDNF is applied uniformly to the flank. These data suggest that PND pathfinding is accomplished by migration up a gradient of GDNF Academic Press Key Words: axolotl; Ambystoma mexicanum; pronephric duct; nephric duct; Wolffian duct; GDNF; GFR -1; cell migration; c-ret; Ret. INTRODUCTION In the axolotl, the incipient pronephros and pronephric duct (PND; also known as the Wolffian, mesonephric, or nephric duct) first appear as a homogeneous bulge in the mesoderm directly ventral to somites 3 7, at approximately embryonic stage 20 (Bordzilovskaya et al., 1989). Cells in the caudal region of this bulge the presumptive PND then migrate posteriorly as a cohesive stream along the ventral border of the somites until they reach and fuse with the cloaca at approximately stages Tissue transplantation studies have revealed that PND migration is a tightly choreographed event controlled by strict directional, spatial, and temporal signals emanating from the overlying epidermis and possibly the underlying lateral mesoderm along the migratory pathway (Poole and Steinberg, 1981, 1982; Gillespie and Armstrong, 1986; Zackson and Steinberg, 1986; Drawbridge et al., 1995; Drawbridge and Steinberg, 1996). A recent resurgence in interest in amphibian pronephric development and mammalian mesonephric development has provided us with several candidate genes 1 To whom correspondence should be addressed. Fax: (609) drawbridge@rider.edu. expressed in the correct spatiotemporal pattern to be involved in PND motility and/or pathfinding (see Carroll et al., 1999, for a review). This study focuses on exploring the roles of glial cell line-derived neurotrophic factor (GDNF) and GDNF family receptor -1 (GFR -1), a ligand and coreceptor for the Ret transmembrane tyrosine kinase, in promoting PND migration in the axolotl. The Ret tyrosine kinase is activated by secreted signaling molecules that are distant members of the TGF family. However, these signaling proteins do not bind directly to Ret. Instead, Ret activation is mediated by a ligand-binding, glycosylphosphatidylinositol (GPI)-linked coreceptor (reviewed in Airakinsinen et al., 1999; Rosenthal, 1999). In vitro, Ret can be activated in cis, when the coreceptor is expressed on the same cell as Ret or in trans, by a soluble coreceptor or one expressed on an adjacent cell (Sanicola et al., 1997; Yu et al., 1998) (see Fig. 1). However, it has recently been shown that Ret signaling through GFR -1 in trans is less effective than in cis (Tansey et al., 2000), bringing into question the role of trans signaling in vivo. At present, we know of four GPI-linked coreceptor:ligand pairs that can trigger Ret signaling: GFR -1:GDNF, GFR - 2:neurturin (NTN), GFR -3:artemin, and GFR -4:persephin (see references in Airakinsinen et al., 1999; Rosenthal, /00 $35.00 Copyright 2000 by Academic Press All rights of reproduction in any form reserved.

2 Pronephric Duct Guidance System ). Although there is some overlap, each coreceptor: ligand pair is expressed in a specific subset of cell types in the developing nervous and excretory systems. Some promiscuity of the ligands for the different coreceptors has been demonstrated in vitro (Baloh et al., 1997; Buj-Bello et al., 1997; Creedon et al., 1997; Klein et al., 1997; Sanicola et al., 1997; Trupp et al., 1998). However, in vivo studies currently indicate, at least for GFR -1:GDNF and GFR -2: NTN, that each pair has physiologically distinct functions and ligand activation of Ret signaling is coreceptor specific (Cacalano et al., 1998; Enomoto et al., 1998; Rossi et al., 1999; Heuckeroth et al., 1999). The GFR -1:GDNF coreceptor:ligand pair is apparently the only one of the four pairs essential for development of the excretory system (Moore et al., 1996; Pichel et al., 1996; Sanchez et al., 1996; Cacalano et al., 1998; Enomoto et al., 1998; Rossi et al., 1999; Heuckeroth et al., 1999). GDNF is expressed in nephrogenic mesenchyme; GFR -1 and Ret are expressed in cells of the migrating tip of the PND primordium and the branching ureteric bud (Pachnis et al., 1993; Robertson and Mason, 1995; Schuchardt et al., 1995; Tsuzuki et al., 1995; Suvanto et al., 1996; Bisgrove et al., 1997; Marcos-Gutiérrez et al., 1997; Homma et al., 2000). When GFR -1, GDNF, or c-ret is eliminated via germline transformation in mice, the neonates die as a result of renal disgenesis and failed enteric innervation (Schuchardt et al., 1994; Durbec et al., 1996; Moore et al., 1996; Pichel et al., 1996; Sanchez et al., 1996; Moore et al., 1996; Cacalano et al., 1998). Further studies reveal that ureteric bud outgrowth and branching are defective in these mutants (Schuchardt et al., 1996; Sainio et al., 1997; Pepicelli et al., 1997; Tang et al., 1998; Srinivas et al., 1999). Thus, the GDNF/ GFR -1/Ret signaling complex is essential for development of the metanephros. Some evidence suggests that this signaling complex might also be required early in development of the excretory system, well before ureteric bud formation. First, c-ret, GFR -1, and GDNF expression is detected in PND epithelia and nephrogenic mesenchyme of amniote embryos well before the ureteric bud forms (Pachnis et al., 1993; Robertson and Mason, 1995; Schuchardt et al., 1995; Tsuzuki et al., 1995; Yu et al., 1998; Golden et al., 1999; Homma et al., 2000). Secondly, c-ret is expressed on the PND primordia of fish and amphibian embryos (Bisgrove et al., 1997; Marcos- Gutiérrez et al., 1997; Carroll et al., 1999), organisms that make neither a ureteric bud nor a metanephros; their terminal kidney is the mesonephros. Therefore, the function of the GDNF/GFR -1/Ret signaling complex in the excretory system must predate the amniotes invention of the metanephros. The observation that enteric innervation as well as kidney development fail in GFR -1, GDNF, or c-ret knockout mice provides an important clue about the potential function of these molecules in early excretory system development. Both the hindbrain (vagal) neural crest, which generates the enteric nervous system, and the PND must migrate to their final destinations in the embryo by following guidance information distributed along their respective pathways. Experiments in axolotl embryos have revealed that when neural crest of the head is transplanted to the flank of a donor embryo it follows the same trajectory a transplanted PND would take (Zackson and Steinberg, 1986; Thibaudeau et al., 1993). Thus, the two tissues which require functioning GDNF/GFR -1/Ret signaling complexes during development PND and hindbrain neural crest can also interpret the same guidance information on the flank. Furthermore, Zackson and Steinberg, (1989) and Thibaudeau et al. (1993) showed that cells of the migrating PND must express GPI-linked proteins on their cell surfaces in order to migrate. Here, we test the hypothesis that Ret signaling via the GDNF:GFR -1 ligand:coreceptor pair functions in either promoting or guiding PND cell migration in axolotl embryos. We show here that application of soluble GFR -1 to embryos denuded of GPI-linked proteins rescues PND migration in a dose-dependent fashion. Thus, GFR -1 is identified as the only GPI-linked protein necessary for PND migration. Further experiments reveal that in vivo application of soluble GFR -1 to the lateral flank of normal axolotl embryos inhibits PND migration and that inhibition is dependent on GDNF-binding activity. We also show that in vivo application of excess GDNF inhibits PND migration when applied uniformly to the flank. However, when a GDNF-impregnated agarose bead is implanted adjacent to the PND pathway, the PND will migrate toward the bead. Taken together, our data support a model in which GDNF signaling through GFR -1 mediates the pathfinding behavior of the migrating PND in axolotl embryos. Thus, we have identified the first molecular components of the axolotl PND guidance system and, in so doing, revealed a function for GDNF and GFR -1 in excretory system development in anamniote vertebrates. METHODS AND MATERIALS Embryo Handling Axolotl embryos were obtained from the Indiana University Axolotl Colony. Embryos were staged according to Bordzilovskaya et al. (1989). Control and experimental embryos for each experiment were same-stage siblings. Embryos were manually dejellied with watchmaker s forceps prior to microinjection or bead implantation. All procedures and subsequent rearing of embryos were carried out in full-strength Hepes-buffered Steinberg s solution (1 HBSt) supplemented with 100 U/ml penicillin and 100 g/ml streptomycin as previously described (Thibaudeau et al., 1993). At the termination of each experiment, embryos were fixed in 2.5% glutaraldehyde in 1 HBSt for 30 min to 1 h, rinsed three times with 1 HBSt, and denuded of epidermis. The epidermis was removed with electrolytically sharpened tungsten needles, hairloops, and watchmaker s forceps in petri dishes lined with nonhardening modeling clay. The pronephros and the PND were visualized directly subsequent to removal of the epidermis from the lateral flank. Albino embryos were stained for endogenous alkaline phosphatase with 5-bromo-4-chloro-3-indolyl phosphate

3 118 Drawbridge, Meighan, and Mitchell (BCIP) for better visualization of the PND and other flank structures as described in Zackson and Steinberg, (1988). Data were recorded via photomicrography. Microinjection and Agarose Bead Implantation A General Valve Corp. Picospritzer II was used to microinject phosphatidylinositol-specific phospholipase C (PI-PLC; Glyko, Inc.), bovine serum albumin (BSA), soluble GFR -1(kindly provided by Dr. M. Sanicola, Biogen, Inc.), and recombinant GDNF (Peprotech, Inc.) into the subectodermal space on the lateral flank of stages embryos essentially as described (Thibaudeau et al., 1993). Injected proteins were concentrated and buffer exchanged into 1 HBSt by three rounds of dilution and concentration in Centricon 10 centrifugation concentrators (Amicon, Inc.). The concentration of PI-PLC used in the microinjection experiments was 600 U/ml; the final concentration of BSA was 1 mg/ml; GDNF (Peprotech, Inc.) and GFR -1 were at a concentration of 1.7 M (50 ng/ l GDNF; 85.6 ng/ l GFR -1) unless otherwise indicated. Approximately nl of protein solution was injected per embryo. Embryos were injected at stages (the beginning of PND migration), reared until uninjected sibling controls reached stages (the end of PND migration) then processed for observation of PND migration. Bead preparation was modified from the method of Tang et al. (1998). Briefly, heparin agarose beads (100 m diameter; Sigma, Inc.) were hydrated in 1 HBSt, rinsed two times in 1 HBSt, and then air-dried. Dried beads were then rehydrated in 10 ng/ l BSA, GDNF, or NTN (Peprotech, Inc.) in 1 HBSt for at least 1 h. After rehydration, individual beads were implanted into stages embryos through a small slit in the epidermis, within the mesodermal flank, posterior to the migrating tip of the PND and adjacent to the normal pathway. Beads were positioned laterally to the incision, rather than directly beneath it, to prevent extrusion prior to healing of the epidermis. Since it is known that the epidermis is important for normal PND migration and that the PND will migrate around an epidermal wound (Poole and Steinberg, 1984; Drawbridge et al., 1995), only embryos which healed completely were scored. In addition, neither embryos in which the bead was inserted directly onto the pathway nor embryos which exhibited abnormal PND migration on the unoperated side were scored. Embryos were reared to stages before fixation. Assessment of the Extent of PND Migration In the axolotl, the wave of somite segmentation and PND migration occur at the same time and move from rostral to caudal at the same rate (Gillespie et al, 1985; Gillespie and Armstrong, 1986). Gillespie and Armstrong (1986) showed that the posterior tip of the axolotl PND is located approximately two somite widths anterior to the posterior-most somite fissure throughout the entire period of PND migration. Thus, segmenting somites serve as anatomical landmarks to which the extent of PND migration can be compared. In agreement with Gillespie and Armstrong, we found that the posterior tips of PNDs in control uninjected and BSA-injected embryos were found two, and, sometimes, three somite widths anterior to the posterior-most somite fissure; these were scored as PND positions 2 and 3 respectively. A score of 4, 5, etc., indicates inhibition of PND migration. Such embryos were placed into the anterior to 3 category. RESULTS GFR -1 Is the PI-PLC-Sensitive Component of the PND Migration System Removal of GPI-linked proteins from stages axolotl embryos via subepidermal application of PI-PLC causes a specific syndrome of morphological defects which include failure of pharyngeal pouch morphogenesis, body elongation along the rostrocaudal axis, and PND extension (Zackson and Steinberg, 1989; Thibaudeau et al., 1993; Drawbridge and Steinberg, 2000). Thibaudeau et al. (1993) showed that the PND defect was due to removal of GPIlinked proteins from cells of the PND, rather than from cells along their migratory pathway. Since it has been shown that soluble, extracellular GFR -1 can mediate GDNF signaling in Ret-expressing cells (Fig. 1) (Sanicola et al., 1997; Tang et al., 1998), we performed an experiment to determine whether GFR -1 is the essential GPI-linked protein required for PND migration. Stages embryos were coinjected with PI-PLC and soluble GFR -1 ranging from 0.5 to 200 ng/ml. Table 1 shows that coinjection of GFR -1 with PI-PLC rescues PND migration in a dose-dependent fashion, the most effective rescue occurring at ng/ml GFR -1. At these concentrations of GFR -1 the PND extends to somite level 2 or 3 greater than 40% of the time. Figure 2 shows an example of GFR -1 rescue of PND migration in PI-PLCtreated embryos. Comparison of Figs. 2B and 2C reveals that anteroposterior elongation of the trunk is not rescued by GFR -1, even though PND migration is. Thus, we have a convenient internal control for the activity of PI-PLC; observation of normal PND migration in an embryo with foreshortened trunk morphology allows us to reliably conclude that rescue of PND migration is due to application of GFR -1 rather than to a reduction in PI-PLC activity. Exogenous GFR -1 Prevents PND Migration Although GFR -1 can rescue PND migration in embryos lacking GPI-linked proteins, we predicted that application of soluble GFR -1 to an untreated embryo should prevent PND migration. We surmised that excess soluble GFR -1 applied to the mesodermal flank would block access of endogenous PND cell-surface GFR -1 to endogenous GDNF. The data in Table 2 reveal that the posterior tips of PNDs in control uninjected and BSA-injected embryos are found two or three somite widths anterior (i.e., at positions 2 or 3 ) to the posterior-most somite fissure in greater than 95% of embryos evaluated. In contrast, injection of soluble GFR -1 inhibits PND migration in 73% of treated embryos. Coinjection of equimolar amounts of GDNF with GFR -1 relieves GFR -1 inhibition; approximately 90% of embryos so treated have PNDs at positions 2 or 3. Thus, GFR -1 applied in excess inhibits PND migration in vivo. Furthermore, GFR -1 s inhibition of PND migration depends upon its GDNF-binding activity.

4 Pronephric Duct Guidance System 119 FIG. 1. A diagram illustrating the interaction of the Ret tyrosine kinase with GFR -1 and GDNF. In vitro, Ret can be activated by GDNF in cis when GPI-linked GFR -1 is expressed on the Ret-expressing cell or in trans when GFR -1 is in soluble form or GPI linked to the surface of an adjacent cell. The significance of trans signaling in vivo has not been determined. GDNF Is a Chemoattractant to PND Cells in Vivo In vitro studies showing that GDNF-responsive neurons and ureteric bud epithelium from rodents will extend TABLE 1 The Effect of GFR -1 on PND Migration in PI-PLC-Treated Embryos ng/ml GFR -1 coinjected with 600 U/ml PI-PLC Position of the posterior PND tip relative to the posterior-most somite a 2 3 Anterior to /20 (0%) 0/20 (0%) 20/20 (100%) 2.5 0/11 (0%) 0/11 (0%) 11/11 (100%) 5 1/16 (6.3%) 0/16 (0%) 15/16 (93.8%) 15 3/16 (18.8%) 0/16 (0%) 13/16 (81.3%) 20 2/14 (14.3%) 0/14 (0%) 12/14 (85.7%) 25 8/18 (44.4%) 0/18 (0%) 10/18 (55.6%) 30 5/17 (29.4%) 2/17 (11.8%) 10/17 (58.8%) 35 4/14 (28.6%) 2/14 (14.3%) 8/14 (57.1%) 50 1/16 (6.0%) 0/16 (0%) 15/16 (94%) 100 1/20 (5.0%) 0/20 (0%) 19/20 (95%) 200 0/14 (0%) 0/14 (0%) 14/14 (100%) a 2 or 3 indicates that the PND tip is two or three somite widths anterior to the posterior-most somite fissure; anterior to 3 indicates that the PND tip is more anterior than position 3. toward an exogenous source of GDNF suggested to us that GDNF might help direct the migrating PND to the cloaca (Pichel et al., 1996; Sainio et al., 1997; Tang et al., 1998). Therefore, we developed an assay to determine whether the axolotl PND would migrate toward an ectopic source of GDNF and/or the GFR -2 ligand, NTN, in living embryos. Briefly, heparin agarose beads soaked in GDNF, NTN, or BSA were inserted into the lateral mesoderm of stages embryos, adjacent to the PND pathway, posterior to the migrating PND tip. When the PNDs of unoperated sibling control embryos reached the cloaca, the position of the PNDs relative to the implanted beads in experimental embryos was determined. Table 3 and Fig. 3 show that the PND migrated off its normal pathway toward a source of GDNF in 81.2% of embryos. In 3 of 13 cases in which the PND responded to the GDNF bead, the PND primordium bifurcated; part of the PND migrated to the GDNF bead while the other part migrated along the normal pathway (Fig. 3D). Neither NTN nor BSA beads were able to redirect the trajectory of the migrating PND. Exogenous GDNF Prevents PND Migration When Microinjected onto the Lateral Flank Possible functions of GDNF signaling in PND migration include promoting PND cell motility, providing directional cues to motile PND cells, or both. Our bead implantation results suggested strongly to us that PND pathfinding could occur via migration up a gradient of GDNF. This hypothesis

5 120 Drawbridge, Meighan, and Mitchell TABLE 2 The Effect of Subepidermal Microinjection of GFR -1 and GDNF on PND Migration Position of the posterior PND tip relative to the posterior-most somite Injected protein 2 3 Anterior to 3 None 45/59 (76.3%) 13/59 (22.0%) 1/59 (1.7%) BSA 39/53 (73.6%) 14/53 (26.4%) 0/53 (0%) PI-PLC 1/32 (3.1%) 3/32 (9.4%) 28/32 (87.5%) GFR -1 2/22 (9.1%) 4/22 (18.2%) 16/22 (72.7%) GDNF 3/15 (20%) 1/15 (6.7%) 11/15 (73.3%) GFR -1 GDNF a 6/10 (60%) 3/10 (30%) 1/10 (10%) a Equimolar concentrations of GDNF and GFR -1 were used. predicts that eliminating the gradient by providing a uniform concentration of GDNF to migrating PND cells should prevent migration. The data in Table 2 reveal that when GDNF is injected subepidermally onto the lateral flank, PND migration is inhibited relative to BSA-injected and uninjected controls. Thus, although the PND will migrate off its pathway toward a source of GDNF, it fails to migrate when presented with a uniform distribution of excess GDNF. DISCUSSION In 1989, Zackson and Steinberg showed that removal of GPI-linked proteins from the embryonic flank via subepidermal application of PI-PLC halted PND migration in axolotl embryos. Because of this and other evidence, they proposed that alkaline phosphatase a cell-surface, GPIlinked protein distributed in a gradient on the lateral flank mesoderm of tailbud stage axolotl embryos provides directional guidance cues to the migrating PND (Zackson and Steinberg, 1988, 1989). However, further studies showed that removal of GPI-linked proteins, including alkaline phosphatase, from the migration pathway did not prevent PND pathfinding, while removal of GPI-linked proteins from migrating PND cells did (Thibaudeau et al., 1993). FIG. 2. Soluble GFR -1 can rescue PND migration in PI-PLC treated embryos. Stages embryos were injected with either BSA (A), PI-PLC (B), or PI-PLC and 30 ng/ml soluble GFR -1 (C), cultured until controls reached stage 30 and denuded of epidermis. Posterior tips of the PNDs are indicated by arrows; the posteriormost somite is indicated with an asterisk. Note the foreshortened bodies of PI-PLC-treated embryos. FIG. 3. The PND migrates toward a source of GDNF in vivo. Stages embryos were implanted with GDNF-coated agarose beads (A and D), no bead (B), or a BSA-coated agarose bead (C) as described under Methods and Materials. At stages 30 34, the embryos were fixed and denuded of epidermis. In (A C) the embryos were stained with BCIP to visualize endogenous PND and lateral flank alkaline phosphatase activity. (D) The flank of a pigmented embryo in which a GDNF-coated bead was implanted dorsal to the normal PND pathway. Note that part of the PND in (D) migrated dorsally onto the bead; the other part migrated on the normal PND pathway. Arrowheads indicate the posterior tips of the PNDs. The white spots in (A), (C), and (D) are agarose beads.

6 Pronephric Duct Guidance System 121 TABLE 3 The Effect of GDNF and NTN Bead Implants on PND Pathfinding Treatment PND migrates to cloaca PND migrates to bead GDNF bead 3/16 (18.8%) 13 a /16 (81.2%) NTN bead 18/18 (100%) 0/18 (0%) BSA bead 21/21 (100%) 0/21 (0%) a In three cases the ND split ; only part of the primordium migrated toward the bead. Two discoveries (1) c-ret expression within the elongating PND primordia of fish and amphibians (Bisgrove et al., 1997; Marcos-Gutiérrez et al., 1997; Carroll et al., 1999) and (2) Ret s requirement for a GPI-linked coreceptor (Jing et al., 1996; Treanor et al., 1996) prompted us to ask whether GFR -1 might be the relevant target of PI-PLC on the migrating axolotl PND. The data shown in Fig. 2 and Table 1 demonstrate rescue of PND migration when soluble GFR -1 is coinjected with PI-PLC. If PND migration required more than one GPI-linked protein, then rescue could not have been achieved with this experiment. Thus, these data tell us that GFR -1 is the only essential GPI-linked component of the PND migration machinery. This experiment also provides indirect evidence that the manipulations of GDNF and GFR -1 reported here exert their effects on PND migration via activation of the Ret tyrosine kinase. First, Ret is expressed on the extending PND of amphibian embryos (Carroll et al., 1999), and it is the only transmembrane kinase known to be activated via GFR -1 (Airaksinen et al., 1999). Second, although it has been hypothesized that GPI-linked GFR -1 may signal independently of Ret in a subset of Ret-negative neurons (Airaksinen et al., 1999; Poteryaev et al., 1999), studies of c-ret and GFR -1 knockout mice suggest that GFR -1 cannot signal without Ret in the excretory system (Schuchardt et al., 1994; Durbec et al., 1996; Tomac et al., 2000). Finally, we rescued PND migration in PI-PLC-treated embryos with soluble GFR -1. Soluble GFR -1 cannot trigger a cellular response unless the responding cell expresses the Ret tyrosine kinase (Sanicola et al., 1997; Tang et al., 1998; Yu et al., 1998). PI-PLC application prevents anteroposterior elongation of the embryonic trunk as well as PND migration (Drawbridge and Steinberg, 2000). However, trunk elongation is not rescued by the concentrations of GFR -1 used in these experiments (Fig. 2). This is not surprising since GFR -1 expression has never been reported in the actively elongating tissues of the trunk dorsal mesoderm and ventral endoderm (Keller, 1987; Drawbridge and Steinberg, 2000). Furthermore, GFR -1 knockout mice exhibit no defects in elongation morphogenesis of these primary tissues. Thus, these data tell us that the syndrome of defects caused by in vivo application of PI-PLC (Zackson and Steinberg, 1989; Thibaudeau et al., 1993; Drawbridge and Steinberg, 2000) is caused by removal of different GPI-linked proteins from the different affected tissues. Although injection of soluble GFR -1 can rescue PND migration in PI-PLC-treated embryos, injection of soluble GFR -1 into untreated embryos inhibits PND migration. In the PI-PLC rescue experiment (Table 1), rescue of PND migration was shown to be dose dependent; either too much or too little GFR -1 fails to rescue migration. Thus, we conclude that the migrating PND can only respond to GDNF when the ratio of GFR -1 to GDNF is sufficiently high to promote signaling, but not so high that excess GFR -1 blocks access of GDNF to Ret. When GFR -1 is injected into a normal embryo (Table 2), the excess condition always exists; therefore, one would expect inhibition of PND migration. However, when an equimolar amount of GDNF is coinjected with GFR -1, PND migration is restored to nearly control levels. Therefore, GFR - 1 s ability to inhibit PND migration is dependent on its GDNF-binding activity. Furthermore, exogenous GDNF/ GFR -1 complexes cannot override the function of endogenous GFR -1 and GDNF. Attempts to determine the order in which Ret, GFR -1, and GDNF interact to trigger Ret kinase activity have yielded conflicting results. Jing et al. (1996) showed, via coprecipitation experiments, that GDNF binds GFR -1 in the absence of Ret and, thus, proposed that GDNF binding to GFR -1 is required to initiate interaction with Ret. However, Eketjall et al. (1999) showed that GDNF defective in GFR -1 binding activates Ret as efficiently as normal and, furthermore, retains the requirement for GFR -1 to activate Ret. These authors propose that, in vivo, GDNF interacts with preassociated GFR -1/Ret complexes. Examination of the data presented here indicates that the model of Eketjall et al. is more likely in axolotl PND migration. The data in Table 2 show that subepidermal application of GDNF alone inhibits PND migration relative to controls while coinjection of GFR -1 and GDNF does not affect normal PND migration. GDNF is also able to elicit a chemotactic response from the PND when applied in an agarose bead. Thus, free, exogenous GDNF can signal to the PND; GFR -1- associated GDNF cannot. If coinjected GFR -1/GDNF complexes had been able to activate Ret effectively, as proposed by Jing et al. (1996), then GDNF coinjected with GFR -1 should have had the same effect on PND migration as GDNF injected alone. In order for the axolotl PND to migrate to the cloaca, PND cells must be both motile and capable of responding to local, directional guidance cues (see Drawbridge and Steinberg, 1996, for a review). GFR -1 injection experiments showed that excess coreceptor prevents posterior translocation of the PND, but do not tell us whether we have interfered with cell motility, cell direction finding, or both. Our bead implantation and GDNF injection experiments help discriminate between these possibilities. Providing the migrating PND with a GDNF-soaked bead elicits the chemotactic response shown in Fig. 3, while subepidermal

7 122 Drawbridge, Meighan, and Mitchell application of a solution of GDNF inhibits PND migration (Table 2). If GDNF functioned solely to elicit cell motility in the PND, one would not expect inhibition of cell migration in the latter experiment. Therefore, we propose that the axolotl PND migrates up a gradient of GDNF to find its way to the cloaca. If the gradient is abolished through either blocking access of the PND to GDNF (via microinjection of soluble GFR -1), or by distributing GDNF homogeneously (via microinjection of GDNF), PND migration is inhibited due to lack of directional information. Thus, our experiments indicate that GDNF signaling promotes either PND pathfinding only or both PND pathfinding and cell motility; it does not function solely to elicit PND cell motility. There are three experimental manipulations that cause the axolotl PND to leave its normal pathway, in vivo: Rotation of the epidermis overlying the PND will cause it to turn in the direction of rotation; i.e., the PND recognizes anterior-to-posterior guidance cues presented to it by the epidermis (Drawbridge et al., 1995). Rotation of immobilized epidermal extracellular matrix (ECM) causes the same PND response, indicating that epidermal ECM is the source of those guidance cues (A. Morris, J. Drawbridge, and M. Steinberg, unpublished results). In this work we show that the PND will migrate off its pathway toward a point source of exogenous GDNF (Table 1). Are we observing two separate PND guidance systems, one intrinsic to epidermal ECM and one a gradient of GDNF? Or, are we altering the distribution of GDNF, and thus the PND trajectory, in all three experiments? At present, either is equally likely, although the fact that TGF s (as well as other growth factors) have known ECM-binding activity makes the latter explanation attractive to us (see Taipale and Keski-Oja, 1997; Sinha et al., 1998, for reviews). The fact that implanted NTN agarose beads did not affect the behavior of the migrating PND was surprising considering that NTN can trigger GFR -1-mediated Ret signaling in vitro (Buj-Bello et al., 1997; Creedon et al., 1997; Klein et al., 1997). However, this result does corroborate gene targeting experiments in mice showing that NTN is not necessary for excretory system development (Heuckeroth et al., 1999). Our bead implantations reveal that NTN does not signal efficiently to the migrating PND in vivo. In summary, the discovery of c-ret expression within the extending PND of fish and amphibian embryos (Bisgrove et al., 1997; Marcos-Gutiérrez et al., 1997; Carroll et al., 1999) implies that Ret signaling is involved in morphogenesis of the early PND in addition to its wellestablished role in branching morphogenesis of the ureteric bud a PND outgrowth formed only in amniote vertebrates. The experiments reported here reveal that the Ret ligand and coreceptor, GDNF and GFR -1, function in the pathfinding behavior of the axolotl PND primordium: the PND interprets the spatial distribution of GDNF via GFR -1 and, thus, extends posteriorly to the cloaca. ACKNOWLEDGMENTS We thank the Axolotl Colony at Indiana University for supplying us with axolotl embryos and Dr. Michele Sanicola for her kind gift of soluble GFR -1. Research in J.D. s laboratory is supported by Grants HD from the NIH, from the New Jersey Commission on Cancer Research, and IBN from the NSF. C.M.M. and E.A.M. participated in this project in partial fulfillment of their undergraduate curriculum requirements. The authors thank Ms. Mary E. Kite for excellent technical assistance. REFERENCES Airaksinen, M. S., Titievsky, A., and Saarma, M. (1999). GDNF family neurotrophic factor signaling: four masters, one servant? Mol. Cell. Neurosci. 5, Baloh, R. H., Tansey, M. G., Golden, J. P., Creedon, D. J., Heuckeroth, R. O., Keck, C. L., Zimonjic, D. B., Popescu, N. C., Johnson, E. M., Jr., and Milbrandt, J. (1997). Trn2, a novel receptor that mediates neurturin and GDNF signaling in through Ret. Neuron 18, Bisgrove, B. W., Raible, D. W., Walter, V., Eisen, J. S., and Grunwald, D. J. (1997). Expression of c-ret in the zebrafish embryo: Potential roles in motoneuronal development. J. Neurobiol. 33, Bordzilovskaya, N. P., Dettlaff, T. A., Duhon, S. T., and Malacinski, G. M. (1989). Developmental-stage series of axolotl embryos. In The Developmental Biology of the Axolotl (J. B. Armstrong and G. M. Malacinski, Eds.), pp Oxford Univ. Press, New York. Buj-Bello, A., Adu, J., Pinon, L. G., Horton, A., Thompson, J., Rosenthal, A., Chinchetru, M., Buchman, V. L., and Davies, A. M. (1997). Neurturin responsiveness requires a GPI-linked receptor and the Ret receptor tyrosine kinase. Nature 387, Carroll, T., Wallingford, J., Seufert, D., and Vize, P. D. (1999). Molecular regulation of pronephric development. Curr. Top. Dev. Biol. 44, Cacalano, G., Farinas, I., Wang, L-C., Hagler, K., Forgie, A., Moore, M., Armanini, M., Phillips, H., Ryan, A., Reichardt, L., Hynes, M., Davies, A., and Rosenthal, A. (1998). GFRalpha1 is an essential receptor component for GDNF in the developing nervous system and kidney. Neuron 21, Creedon, D. J., Tansey, M., Baloh, R., Osborne, P., Lampe, P., Fahrer, T., Heuckeroth, R., Millbrandt, J., and Johnson, E. M., Jr. (1997). Neurturin shares receptors and signal transduction pathways with glial cell line-derived neurotrophic factor in sympathetic neurons. PNAS 94, Drawbridge, J., and Steinberg, M. S. (1996). Morphogenesis of the axolotl pronephric duct: A model system for the study of cell migration in vivo. Int. J. Dev. Biol. 40, Drawbridge, J., and Steinberg, M. S. (2000). Elongation of axolotl tailbud embryos requires GPI-linked proteins and organizerinduced, active, ventral trunk endoderm cell rearrangements. Dev. Biol. 223, Drawbridge, J., Wolfe, A. E., Delgado, Y. L., and Steinberg, M. S. (1995). The epidermis is a source of directional information for the migrating pronephric duct in Ambystoma mexicanum embryos. Dev. Biol. 172, Durbec, P., Larsson-Blomberg, L. B., Schuchardt, A., Costantini, F., and Pachnis, V. (1996). Common origin and developmental

8 Pronephric Duct Guidance System 123 dependence on c-ret of subsets of enteric and sympathetic neuroblasts. Development 122, Eketjall, S., Fainzilber, M., Murray-Rust, J., and Ibanez, C. F. (1999). Distinct structural elements in GDNF mediate binding to GFRalpha1 and activation of the GFRalpha1 c-ret receptor complex. EMBO J. 18, Enomoto, H., Araki, T., Jackman, A., Heuckeroth, R. O, Snider, W. D., Johnson, E. M., and Milbrandt, J. (1998). GFR 1-deficient mice have deficits in the enteric nervous system and kidneys. Neuron 21, Gillespie, L. L., Armstrong, J. B., and Steinberg, M. S. (1985). Experimental evidence for a proteinaceous presegmental wave required for morphogenesis of axolotl mesoderm. Dev. Biol. 107, Gillespie, L. L., and Armstrong, J. B. (1986). Morphogenetic waves in the development of the lateral mesoderm in the Mexican axolotl (Ambystoma mexicanum) and their relationship to pronephric duct migration. J. Exp. Zool. 237, Golden, J., DeMaro, J., Osbourne, P., Millbrandt, J., and Johnson, E. M., Jr. (1999). Expression of Neurturin, GDNF and GDNF family-receptor mrna in the developing and mature mouse. Exp. Neurol. 158, Heuckeroth, R. O., Enomoto, H., Grider, J. R., Golden, J. P., Hanke, J. A., Jackman, A., Molliver, D., Bardgett, M. E., Snider, W. D., Johnson, E. M., and Millbrandt, J. (1999). Gene targeting reveals a critical role for neurturin in the development and maintenance of enteric, sensory and parasympathetic neurons. Neuron 22, Homma, S., Oppenheim, R. W., Yaginuma, H., and Kimura, S. (2000). Expression pattern of GDNF, c-ret and GFR s suggests novel roles for GDNF ligands during early organogenesis in the chick embryo. Dev. Biol. 217, Jing, S., Wen, D., Yu, Y., Holst, P. L., Luo, Y., Fang, M., Tamir, R., Antonio, L. Hu, Z., Cupples, R. Louis, J. C., Hu, S., Altrock, B. W., and Fox, G. M. (1996). GDNF-induced activation of the ret protein tyrosine kinase is mediated by GDNFR-alpha, a novel receptor for GDNF. Cell 85, Keller, R. (1987). Cell rearrangement in morphogenesis. Zool. Sci. 4, Klein, R. D., Sherman, D., Ho, W-H., Stone, D., Bennett, G. L., Moffat, B., Vandlen, R., Simmons, L., Gu, Q., Hongo, J., Devaux, B., Poulsen, K., Armanini, M., Nozaki, C., Asai, N., Goddard, A., Phillips, H., Henderson, C. E., Takahashi, M., and Rosenthal, A. (1997). A GPI-linked protein that interacts with Ret to form a candidate neurturin receptor. Nature 387, Marcos-Gutiérrez, C. V., Wilson, S. W., Holder, N., and Pachnis, V. (1997). The zebrafish homologue of the ret receptor and its pattern of expression during embryogenesis. Oncogene 14, Moore, M. W., Klein, R. D., Farinas, I., Sauer, H., Armanini, M., Philllips, H., Reichardt, L., Ryan, A., Carver-Moore, K., and Rosenthal, A. (1996). Renal and neuronal abnormalities in mice lacking GDNF. Nature 382, Pachnis, V., Mankoo, B., and Costantini, F. (1993). Expression of the c-ret proto-oncogene during mouse embryogenesis. Development 119, Pepicelli, C. V., Kispert, A., Rowitch, D. H., and McMahon, A. P. (1997). GDNF induces branching and increased cell proliferation in the ureter of the mouse. Dev. Biol. 192, Pichel, J. G., Shen, L., Sheng, H., Granholm, A., Drago., J., Grinberg, A., Lee, E., Huang, S., Saarma, M., Hoffer, B., Sariola, H., and Westphal, H. (1996). Defects in enteric innervation and kidney development in mice lacking GDNF. Nature 382, Poole, T. J., and Steinberg, M. S. (1981). Amphibian pronephric duct morphogenesis: Segregation, cell rearrangement and directed migration of the Ambystoma duct rudiment. J. Embryol. Exp. Morphol. 63, Poole, T. J., and Steinberg, M. S. (1982). Evidence for the guidance of pronephric duct migration by a craniocaudally traveling adhesion gradient. Dev. Biol. 92, Poole, T. J., and Steinberg, M. S. (1984). Different modes of pronephric duct origin among vertebrates. Scan. Electron Microsc./1984 1, Poteryaev, D., Titievsky, A., Sun, Y., Thomas-Crusells, J., Lindahl, M., Billaud, M., Saarma, M., and Arumäe, U. (1999). GDNF triggers a novel Ret-independent Src kinase family-coupled signaling via a GPI-linked GDNF receptor 1. FEBS Lett. 463, Robertson, K., and Mason, I. (1995). Expression of ret in the chicken embryo suggests roles in regionalisation of the vagal neural tube and somites and in development of multiple neural crest and placodal lineages. Mech. Dev. 53, Rosenthal, A. (1999). The GDNF protein family: Gene ablation studies reveal what they really do and how. Neuron 22, Rossi, J., Luukko, K., Poteryaev, D., Laurikäinen, A., Sun, Y. F., Laakso, T., Eerikainen, S., Tuominen, L., Laakso, M., Rauvala, H., Arumäe, U., Pasternack, M., Saarma, M., and Airaksinen, M. S. (1999). Retarded growth and deficits in the enteric and parasympathetic nervous system in mice lacking GFR 2, a functional neurturin receptor. Neuron 22, Sainio, K., Suvanto, P., Davies, J., Wartiovaara, J., Wartiovaara, K., Saarma, M., Arumäe, U., Meng, X., Lindahl, M., Pachnis, V., and Sariola, H. (1997). Glial-cell-line-derived neurotrophic factor is required for bud initiation from ureteric bud epithelium. Development 124, Sanchez, M. P., Silos-Santiago, I., Frisen, J., He, B., Lira, S. A., and Barbacid, M. (1996). Renal agenesis and the absence of enteric neurons in mice lacking GDNF. Nature 382, Sanicola, M., Hession, C., Worley, D., Carmillo, P., Ehrenfels, C., Walus, L., Robinson, S., Jaworski, G., Wei, H., Tizard, R., Whitty, A., Pepinsky, R., and Cate, R. (1997). Glial cell line-derived neurotrophic factor-dependent RET activation can be mediated by two different cell-surface accessory proteins. PNAS 94, Schuchardt, A., D Agati, V., Larsson-Blomberg, L., Costantini, F., and Pachnis, V. (1994). Defects in the kidney and enteric nervous system of mice lacking the tyrosine kinase receptor Ret. Nature 367, Schuchardt, A., Srinivas, S., Pachnis, V., and Costantini, F. (1995). Isolation and characterization of a chicken homologue of the c-ret proto-oncogene. Oncogene 10, Schuchardt, A., D Agatti, V., Pachnis, V., and Costantini, F. (1996). Renal agenesis and hypodysplasia in ret k-mutant mice result from defects in ureteric bud development. Development 122, Sinha, S., Nevett, C., Shuttleworth C. A., and Kielty, C. M. (1998). Cellular and extracellular biology of the latent transforming growth factor-beta binding proteins. Matrix Biol. 17, Srinivas, S., Wu, Z., Chen, C-M., D Agati, V., and Costantini, F. (1999). Dominant effects of RET receptor misexpression and ligand-independent RET signaling on ureteric bud development. Development 126, Suvanto, P., Hiltunen, J. O., Arumäe, U., Moshnyakov, M. Sariola, H. Sainio, K., and Saarma, M. (1996). Localization of glial cell

9 124 Drawbridge, Meighan, and Mitchell line-derived neurotrophic factor (GDNF) mrna in rat by in situ hybridization. Eur. J. Neurosci. 8, Taipale, J., and Keski-Oja, J. (1997). Growth factors in the extracellular matrix.. FASEB J. 11, Tang, M., Worley, D., Sanicola, M., and Dressler, D. R. (1998). The RET glial cell-derived neurotrophic factor (GDNF) pathway stimulates migration and chemoattraction of epithelial cells. J. Cell Biol. 142, Tansey, M. G., Baloh, R. H., Milbrandt, J., and Johnson, E. M., Jr. (2000). GFRalpha-mediated localization of RET to lipid rafts is required for effective downstream signaling, differentiation, and neuronal survival. Neuron 25, Thibaudeau, G., Drawbridge, J., Dollarhide, A. W., Haque, T., and Steinberg, M. S. (1993). Three populations of migrating amphibian embryonic cells utilize different guidance cues. Dev. Biol. 159, Tomac, A. C., Grinberg, A., Huang, S. P., Nosrat, C., Wang, Y., Borlongan, C., Lin, S. Z., Chiang, Y. H., Olson, L., Westphal, H., and Hoffer, B. J. (2000). Glial cell line-derived neurotrophic factor receptor alpha1 availability regulates glial cell line-derived neurotrophic factor signaling: Evidence from mice carrying one or two mutated alleles. Neuroscience 95, Treanor, J. J., Goodman, L., de Sauvage, F., Stone, D., Poulsen, K., Beck, C., Gray, C., Armanini, M., Pollack, R., Hefti, F., Phillips, H., Goddard, A., Moore, M., Buj-Bello, A., Davies, A., Asai, N., Takahashi, M., Vandlen, R., Henderson, C., and Rosenthal, A. (1996). Characterization of a multicomponent receptor for GDNF. Nature 382, Trupp, M., Raymoschek, C., Belluardo, N., and Ibanez, C. (1998). Multiple GPI-anchored receptors control GDNF-dependent and independent activation of the c-ret receptor tyrosine kinase. Mol. Cell. Neurosci. 11, Tsuzuki, T., Takahashi, M., Asai, N., Iwashita, T., Matsuyama, M., and Asai, J. (1995). Spatial and temporal expression of the ret proto-oncogene product in embryonic, infant and adult rat tissues. Oncogene 10, Yu, T., Scully, S., Yu, Y., Fox, G. M., Jing, S., and Zhou, R. (1998). Expression of GDNF family receptor components during development: Implications in the mechanisms of interaction. J. Neurosci. 18, Zackson, S. L., and Steinberg, M. S. (1986). Cranial neural crest cells exhibit directed migration on the pronephric duct pathway: Further evidence for an in vivo adhesion gradient. Dev. Biol. 117, Zackson, S. L., and Steinberg, M. S. (1988). A molecular marker for cell guidance information in the axolotl embryo. Dev. Biol. 127, Zackson, S. L., and Steinberg, M. S. (1989). Axolotl pronephric duct cell migration is sensitive to phosphatidylinositol-specific phospholipase C. Development 105, 1 7. Received for publication August 7, 2000 Revised September 8, 2000 Accepted September 11, 2000 Published online October 25, 2000