Bidirectional inhibitory interactions between the embryonic chicken metanephros and lumbosacral nerves in vitro

Authors


Abstract

During chicken embryonic development the metanephros forms from the uretic duct at embryonic day (E) 7. As the metanephric tissue develops between E7 and E10, it comes into close apposition with lumbosacral nerves. Coculturing of metanephric and nerve explants demonstrated that the Schwann cells of the sciatic nerve inhibit the migration of metanephric cells in a contact-dependent manner. Conversely, metanephric cells inhibit dorsal root ganglion axon extension in a contact-dependent manner. However, metanephric cells are not inhibited by contact with growth cones or axons. Dorsal root ganglion growth cones become sensitive to the inhibitory signals on the surfaces of metanephric cells around E8, a time when the metanephros is expanding into the territory occupied by nerves in vivo. These observations demonstrate inhibitory bidirectional tissue–tissue interactions in vitro and provide a novel model system for the study of contact-based guidance of both neuronal and non-neuronal cell migration. Developmental Dynamics 231:190–198, 2004. © 2004 Wiley-Liss, Inc.

INTRODUCTION

Embryonic development is dependent on numerous cell–cell interactions. Both diffusible and cell surface signals between disparate cell populations determine the organization and differentiation of organs. A major role of cell–cell interactions during development is the transmission of signals leading to the establishment of boundaries between distinct cell types (Cowan and Henkemeyer, 2002). For example, during development of the hindbrain, rhombomeres form through inhibitory interactions between adjacent segments (Cooke and Moens, 2002). Similarly, the interactions of neuronal growth cones with inhibitory guidance cues is of fundamental importance to the development of appropriate axon projection patterns (Holder and Klein, 1999).

The chicken embryo has been a valuable model system in the clarification of embryological mechanisms in vivo and in vitro. Dissection of embryonic day 10 (E10) dorsal root ganglia (DRG) for use in culturing experiments requires the removal of surrounding tissues. A major impediment to removing the DRGs from the embryo is the late embryonic kidney, the metanephros. The metanephric tissue is in direct apposition with both the DRGs and the associated nerves and care must be taken to remove the metanephric tissue while preparing DRGs for culturing. The metanephros forms at E7 as an epithelial budding from the uretic duct and expands during the next 3 developmental days into the lumbosacral region of the embryo (Rienhoff, 1922). Between E7 and E10, neuronal differentiation and the extension of new axons into nerves continues in the DRG (Hamburger and Levi-Montalcini, 1949; Levi-Montalcini and Levi, 1943). In addition, during the period that the metanephros expands into the lumbosacral region, the nerves do not exhibit a fully developed perineurium (Du Plessis et al., 1996). The chick sciatic nerve perineurium does not mature until E10, a time point when the metanephros has fully surrounded the lumbosacral nerves and the initial segment of the sciatic nerve. Thus, these observations indicate that cells of the developing metanephros have the potential to come into contact with Schwann cells within nerves and possibly even the growth cones of extending axons, suggesting the hypothesis that mechanisms exist to prevent intermingling between metanephric and nerve cells.

In this study, we describe the in vitro inhibition of DRG axon extension by metanephric cells in a contact-dependent manner and the in vitro contact-mediated inhibition of metanephric cell motility by nerve-derived Schwann cells. These observations identify bidirectional inhibitory cell–cell interactions between DRGs and metanephric tissue in vitro. We suggest these inhibitory interactions serve to keep DRG, and the associated nerves, and the expanding metanephros in separate organ-specific domains.

RESULTS

Metanephric Cells Inhibit DRG Axon Extension in a Contact-Dependent Manner

To determine potential interactions between metanephric cells and neurons, we cocultured explants of E10 DRG and metanephros, reflective of a time during development when both tissues have made extensive contact in vivo. After overnight culturing, we observed that DRG axons failed to extend on territory occupied by metanephric cells migrating out from the explants (Fig. 1A). Axons could be observed as far as the edge of the front of the sheet of metanephric cells, indicating that the inhibition of axon outgrowth was not due to diffusible factors. To directly investigate whether the inhibition of axon extension by metanephric cells was due to a contact-mediated mechanism, we analyzed time-lapse videos of individual DRG growth cones interacting with metanephric cells under the same culturing conditions. These studies revealed that, after contact with metanephric cells, growth cones underwent collapse and retraction, stalled upon contact with metanephric cells, or, in rare cases, extended onto metanephric cells (Fig. 1B). Stalling growth cones often exhibited branching behavior (e.g., see right panel of Fig. 1B). Branching during stalling in response to contact with a cell surface inhibitory guidance cue has been described also in other systems (Oakley and Tosney, 1993; Thies and Davenport, 2003). No effects were observed on growth cones before contact with metanephric cells, further arguing against an inhibitory diffusible factor being released by metanephric cells. The observation that growth cones collapse and retract after contact with metanephric cells indicates the presence of an active inhibitory signal on metanephric cells and not the absence of axon extension supporting factors on the metanephric cell surface. The responses of growth cones to contact with inhibitory metanephric cells are not evident during contact with chicken primary fibroblasts or Schwann cells that growth cones can readily extend over after contact (Wessells et al., 1980; our unpublished observations). Thus, the inhibition of growth cones through contact is not a generalized response to non-neuronal cells.

Figure 1.

Metanephric cells inhibit dorsal root ganglia (DRG) axon extension in a contact-dependent manner. A: Overnight coculture of embryonic day (E) 10 DRG and metanephric (M) explants. Note the sheet of cells migrating from the metanephric explant. DRG axons are found throughout the substratum not occupied by metanephric cells. However, DRG axons are excluded from the territory occupied by metanephric cells. The inset shows a high magnification of the border of the metanephric cell sheet denoted by a white box. Note the presence of axons bundles at the edge of the metanephric sheet (arrows). B: Examples of the responses of E10 growth cones to contact with cells at the edge of the sheet of cells migrating from E10 metanephric explants. The time-lapse sequence on the left shows an example of a growth cone undergoing retraction after contact. The sequence on the right shows an example of a growth cone stalling at the edge. Note that the growth cone neither advances nor retracts for a 40-min time period. Numbers in panels reflect elapsed minutes from the first panel in the sequence. C: Histogram of the percentage of growth cones retracting (black bars), stalling (gray bars), or extending (white bars) onto E10 metanephric cells after contact. The ages (En) below the bars reflect the age of the DRG growth cones exhibiting that response. Note that all E6 DRG growth cones (n = 14) extend and none retract or stall. Conversely, the majority of E10 DRG growth cones (n = 23) retract and stall, whereas only approximately 10% extend onto the cells. The E8 DRG growth cones (n = 18) show an intermediate distribution of responses. D: Both E7 and E10 metanephric cells are inhibitory to E10 DRG axon extension and growth cones exhibit similar response profiles regardless of metanephric age (n = 23 growth cones for both E10 and E7 metanephros). E: Example of an E6 growth cone extending over/under a metanephric cell. Note that the growth cone does not collapse and exhibits a broad fan-shaped morphology (black arrows) even when in complete contact with the metanephric cell (edge of cell shown with white arrows). Numbers in top right corner of each panel of E represent elapsed minutes. Scale bars = 200 μm in A, 5 μm in B,E.

Given that possible interactions between metanephros and DRG axons in vivo would occur between E7, when the metanephros is forming, and E10, we investigated the time course of the development of the inhibitory interactions using heterochronic cocultures analyzed by time-lapse microscopy. These studies revealed that the sensitivity of DRG growth cones to E10 metanephric cells was minimal at E6, a time before development of possible contact-mediated interaction in vivo, and increased between E6 and E10 (Fig. 1C). The 0%, 39%, and 91% of E6, E8, and E10 axons, respectively, either retracted or stalled upon contact with E10 metanephric cells. Conversely, 100%, 61%, and 9% of E6, E8, and E10 axons extended onto the metanephric cells. E10 DRG growth cones were inhibited by metanephric tissue of all developmental stages tested (E7–E10; Fig. 1D). Thus, these data show that DRG neurons become sensitive to the metanephric inhibitory signal during the period when potential tissue-to-tissue interactions develop in vivo. Conversely, metanephric tissue of all stages tested was inhibitory to responsive E10 DRG growth cones.

To determine whether additional neuronal types, predicted to not interact with metanephric tissue due to anatomical localization, are also capable of responding to metanephric inhibitory signals, we cocultured E7 retinal explants and E10 ciliary ganglia with E10 metanephric tissue. Similar to E10 DRG, the extension of axons from both retinal and ciliary explants was inhibited by contact with cells migrating from metanephric explants (not shown). Kidneys receive sympathetic afferent innervation (reviewed in DiBona, 2002), suggesting that sympathetic axons may not be responsive to metanephric inhibitory signals. As determined by anti–neuron-specific tubulin immunocytochemistry, the axons of E10 sympathetic chain ganglia neurons extended into territory occupied by metanephric cells (Fig. 2A). These observations indicate that the metanephric inhibitory signal may be conserved in additional guidance systems as diverse neuronal types respond to metanephric tissue. On the other hand, the observation that DRG neurons develop sensitivity to the metanephric cells and that sympathetic chain ganglia are insensitive, demonstrate that the inhibition can be specific, even within the same neuronal cell type.

Figure 2.

Immunocytochemical studies of interactions between metanephric cells and nerve cells. A: Embryonic day (E) 10 sympathetic axons labeled with neuron-specific tubulin antibodies (green) extending deep into territory occupied by E10 metanephric cells (red background represents F-actin staining of metanephric cells migrating from two adjacent explants, M) after 1 day in vitro. Sympathetic explants were positioned to the right of the metanephric explants in this picture. B: After 6 days of coculturing, metanephric cells have taken over most available territory and the dorsal root ganglia axons have consolidated into a few large fascicles (explants were E10 at time of culturing). The fascicle is denoted in the phase image and the inset below by arrows). The lower panel shows neuron specific tubulin staining (green) and F-actin staining (red) for a fascicle similar to that shown in the upper panel. Note the fascicle contains axons, which are not present on the metanephric cell (red) -covered substratum. C: Similar staining as in B showing metanephric cells (red) extending beneath axons (green), the metanephric explant is to the right of this picture, and cells were migrating to the left. Note the axons appear to be “swept” to the left by the metanephric cells and exhibit fasciculation. D: Coculture of unlabeled E10 sciatic nerve explant (SN) and dissociated CellTracker-labeled E10 metanephric cells. The metanephric cells appear red (actin filaments) with a green cytoplasm (CellTracker). In the right panel, a blue dotted line approximating the border between the metanephric cells and nerve-derived cells has been drawn. The right panel shows a higher magnification view of the white-boxed area in the left panel, the border between cell types is shown as a pale yellow line. Note that the cells outside the border, to the left, are metanephric cells (labeled green), whereas the cells inside the border, to the right, have unlabeled cytoplasm. In addition the metanephric cells exhibit much brighter F-actin staining than the nerve-derived cells. E: P0 (green) and F-actin (red) stained dissociated E10 nerve culture. Note the irregular morphology of the P0-labeled Schwann cells relative to the P0 negative, but rich in F-actin and stress fibers, fibroblasts. Scale bars = 50 μm in A,C, D, 100 μm in B, 20 μm in E.

Growth cone collapse and retraction in response to inhibitory signals often requires activity of the RhoA-ROCK pathway (reviewed in Gallo and Letourneau, 2004). We therefore tested whether blockade of ROCK using the specific inhibitor y-27632 altered the response of E10 DRG growth cones to E10 metanephric cells. In the presence of 20 μM y-27632, the response of DRG growth cones to contact with metanephric cells was altered. Although y-27632 did not allow the majority axons to extend on the metanephric cells, it largely blocked metanephric cell induced axon retraction (Fig. 1C). A total of 52% and 11% of axons underwent retraction after contact with metanephric cells in the absence and presence of y-27632, respectively. In the presence of y-27632, the majority of DRG growth cones stalled after contact, neither extending nor retracting. y-27632 increased the percentage of E10 axons that extended onto metanephric cells from 9% to 28%. Thus, ROCK is required for retraction after contact with metanephric cells, and inhibition of ROCK only modestly allows axons to extend into metanephric territory.

Long-Term Coculture of Metanephric Tissue and DRG Results in Axonal Fasciculation

To determine the effects of continued interactions between DRG axons and metanephric cells, we cultured explants for up to 6 days in vitro (n = 8 cocultures). These studies revealed that, with increasing time in culture, DRG axons progressively lost territory to the expanding metanephric cells. DRGs cultured alone exhibit a radial pattern of axon extension. However, by 3 days in the presence of metanephric explants, axons congregated into fascicles. By 6 days, a time when metanephric cells have extensively migrated from explants and cover the majority of the substratum, DRG axons were present only in the form of large fascicles (Fig. 2B). Consistent with the live imaging observations, growth cones and axons failed to reveal any inhibitory effects on metanephric cell advance, and metanephric cells were observed beneath axons in immunostained cultures (Fig. 2C). The bundles of axons that originated from the DRG and traversed the territory occupied by metanephric cells terminated in nonoccupied territory, and the axons forming the bundles splayed out and grew as individuals (not shown). Thus, metanephric cells caused axons to reorganize into fascicles while axons and growth cones did not affect metanephric cell advance. These observations demonstrate that the loss of available territory results in the bundling of DRG axons into nerve-like structures in vitro.

Nerve-Derived Schwann Cells Inhibit Metanephric Cell Migration in a Contact-Dependent Manner

The observation that metanephric cells inhibit DRG axon extension suggests a mechanism for preventing DRG axons from inappropriately entering metanephric tissue. To determine whether there are mechanisms that prevent the metanephric tissue from invading nerves, we cocultured explants of E10 metanephros with E10 sciatic nerve. We observed that migration of metanephric cells was inhibited on the side of the explants facing the nerve explants (n = 6 cocultures), from which numerous non-neuronal cells had migrated (Fig. 3A). Although the cells migrating from nerve explants exhibited morphologies distinct from those of metanephric cells, these experiments did not allow us to specifically differentiate metanephric and nerve cells. Therefore, we labeled dissociated metanephric cells by using CellTracker and plated the labeled cells at high density over freshly plated nerve explants (4 hr in culture) at a time when explants had attached but cells had not yet migrated out (not shown). Cultures were incubated overnight and examined the next day after staining with phalloidin to reveal all cells (n = 6). Consistent with the observation that metanephric cells and nerve explant cells exhibit different morphologies, phalloidin staining of nerve explant cells and metanephric cells was also strikingly different. Nerve explant cells did not exhibit stress fibers and stained significantly dimmer with phalloidin than metanephric cells (Fig. 2D). These experiments corroborated the previous study by demonstrating that metanephric cells were largely excluded from the halo of cells migrating from the nerve explant (Fig. 2D).

Figure 3.

Schwann cells inhibit metanephric cell migration in a contact-dependent manner by altering lamellipodial dynamics. A: One day in vitro coculture of embryonic day (E) 10 nerve (N) and E10 metanephros (M). Note the distance the metanephric cell front has migrated on the side opposite to the nerve explant (similar to metanephric explant alone, not shown). On the side of the metanephric explant facing the nerve explant metanephric cell migration has halted and the nerve explant and metanephric cells form a border (location approximated by two black arrowheads). B: Time-lapse sequence of the interaction between a nerve-derived Schwann cell (S) and a DiI-labeled metanephric cell (M; both E10). The DiI label of the metanephric cell has accumulated in the perinuclear region, likely due to endocytotic activity (top left panel is an overlay of the phase image and the DiI staining). The box in the top left panel shows the region where the Schwann cell and metanephric interaction occurs. Numbers in panels reflect seconds lapsed from the first frame in the sequence. By 60 sec, the Schwann cell has extended a lamellipodium overlapping the metanephric lamellipodium (arrows). Over the next 30–60 sec, the metanephric cell retracts the lamellipodium and by 90–240 sec, the Schwann cell lamellipodium occupies the vacated region where the interaction occurred and the metanephric lamellipodium is not re-extended in the direct of the Schwann cell. C: Quantification of the dynamics of the metanephric lamellipodium before and after contact with Schwann cells. Five interactions were analyzed in which the metanephric cell's leading edge perpendicularly approached a Schwann cell, thus standardizing the angle of interaction to facilitate analysis. Frames were acquired every 30 sec. In each case, the behavior of the lamellipodium was quantified for 5 min before and after contact (see Experimental Procedures section for details on the quantification of the behavioral data). Comparisons within category and across time (before and after) with the paired t-test. *P < 0.000001, **P < 0.0001, and ***P = 0.53. Scale bars = 200 μm in A, 10 μm in B.

To determine whether the cells that migrated from E10 nerve explants were Schwann cells, we stained these cells with an antibody to a Schwann cell–specific marker (P0) expressed at all stages of chicken Schwann cell development (Bhattacharyya et al., 1991). This study revealed a strong correlation between the morphology of the cells and the expression of P0. Schwann cells in vitro are described as exhibiting elongated leading processes and lamellipodia but no stress fibers (Uziyel et al., 2000). These morphological features distinguish Schwann cells from the polarized and stress fiber–containing fibroblasts. We observed P0 staining only in cells exhibiting the morphology of Schwann cells (Fig. 2E). A total of 83% of cells (n = 56) in E10 dissociated nerve cultures were characterized as Schwann cells based on combined P0-positive staining and morphological criteria. Thus, although some dissociated cells from E10 nerves were fibroblasts, presumably derived from the developing epineurium that is difficult to fully remove during dissection, cell morphology was a good predictor of Schwann cells in our culture system.

To determine whether the inhibition of metanephric cell migration by nerve-derived Schwann cells was dependent on contact-mediated inhibition, we cocultured metanephric explants and dissociated Schwann cells obtained from E10 sciatic nerve. The metanephric cells were labeled with DiI to distinguish them from the nerve-derived cells during live imaging. Schwann cells were identified by the lack of DiI staining and characteristic morphology (as described above). Nerve-derived cells that exhibited a fibroblast-like morphology were not considered in these experiments. These studies revealed that Schwann cells inhibited metanephric cell lamellipodial advance (Fig. 3B,C). The behavior of the metanephric cell lamellipodium was compared before and after contact with the Schwann cell. After contact with Schwann cells, the lamellipodium of metanephric cells stopped advancing and either remained stable or underwent retraction. Individual metanephric lamellipodia were observed to undergo repeated attempts at contact with the Schwann cell surface, each time undergoing contact-mediated repulsion. A total of 29 interactions were monitored. Quantitative analysis of the details of five interactions between metanephric cells and Schwann cells provided direct support for the qualitative observations (Fig. 3C). Contact with Schwann cells decreased the percentage of metanephric lamellipodium undergoing protrusion by 75%. Conversely, after contact the percentage of perimeter remaining stable increased by 195%. The percentage of perimeter undergoing retraction was only modestly increased by 22%. These interactions were selected for quantitative analysis based on the criterion that the lamellipodia of the metanephric cell perpendicularly approached the Schwann cell surface, providing a standardized geometry of interaction. However, the inhibition of metanephric lamellipodial protrusion was observed regardless of the angle of cell–cell interaction. These observations indicate that the response of the metanephric cell lamellipodium is localized to the site of contact with the Schwann cell. Qualitative observations of contact between metanephric cells did not reveal signs of repulsion and cells were observed to extend lamellipodia over/under one another.

Metanephric Membranes Collapse DRG Growth Cones and Halt Axon Extension

To further develop in vitro assays for the analysis of the axon extension-inhibitory signal present on metanephric cells, we prepared crude plasma membrane particles from E10 metanephric cells. The membrane particles were subsequently spotted on polylysine-coated culturing surfaces and E10 DRG explants placed in the vicinity of the border of the membrane spots (n = 16 explants). After overnight culturing, we determined whether DRG axons were able to extend into the membrane spot. These experiments demonstrated that DRG axons avoided the E10 membranes (Fig. 4). Only on rare occasions were axons observed to extend into the membrane spot, consistent with the low percentage of E10 DRG axons that extend into territory occupied by living metanephric cells (Fig. 1C). When axons grew onto the membranes, the distance they extended into the membranes spot was minor relative to the outgrowth on the other side of the explant not contacting the membranes (Fig. 4). E10 DRG explants placed directly on the membrane carpet failed to extend any axons (n = 3). Live imaging of the responses of E10 DRG growth cones to contact with the membrane carpet revealed an overall similar response as growth cones (n = 8 growth cones) contacting living metanephric cells. Upon contact growth cones collapsed and axons stalled. We did not observe retraction of axons on polylysine. This finding is not surprising as polylysine provides strong attachment of the axon to the substratum effectively countering retraction. Regardless of this minor substratum-dependent difference between the responses of growth cones to membranes on polylysine relative to living cells on laminin, the membranes halted axon extension.

Figure 4.

Metanephric membrane particles inhibit axon extension. A: A carpet of E10 metanephric membranes (MM) prevents embryonic day (E) 10 dorsal root ganglia (DRG) axon extension. Note the axons stopped at the membrane border. Axon extension on the opposite side of the explant continued unabated (not shown). B: Example of an E10 DRG explant cultured at the border with MM (dotted line). Note the extensive outgrowth of axons on the right side of the border. Few axons have extended on the MM (arrows), and they are much less dense and less than 20% in length of the axons not growing on the MM (not shown). C: Time-lapse sequence of E10 DRG axons approaching and interacting with MM. The axons stop at the MM border and the growth cones collapse. Numbers in panels reflect minutes into the sequence. Insets at 22–105 min show the stalling behavior of two growth cones contacting the MM. D: Example of an E10 sympathetic ganglion cultured at the border of MM (similar to the E10 DRG explant in B). Note that axons extended on MM to a similar distance and at similar density as on polylysine alone (two-headed arrows, dotted line denotes the MM border). Scale bars = 20 μm in A (applies to A,B), 50 μm in C, 200 μm D.

E10 sympathetic and E6 DRG axons are relatively insensitive to the inhibitory signal presented by metanephric cells (Figs. 1C, 2A). Therefore, we tested whether these neuron types could extend axons on metanephric membrane particles. Experiments were set up as for the E10 DRG explants and cultured for 2 days. Both E6 DRG (n = 9 explants) and E10 sympathetic explants (n = 9) gave rise to axons regardless of placement relative to the membranes. For example, axons could readily extend across the membrane border and extend to a similar distance as the control axons from the same explants growing on polylysine alone (Fig. 4D). In addition, both E6 DRG and E10 sympathetic explants generated axons even when cultured directly on the membrane carpets (n = 2 explants in each case; not shown), in striking contrast to the complete absence of axon extension exhibited by E10 DRG axons. These experiments demonstrate that the axon inhibitory activity present on E10 metanephric membranes is specific for E10 DRG axons, in a manner not distinguishable from the living metanephric cells and that failure of E10 DRG axons to cross the border into the membrane carpet or extend axons on the carpet is not due to nonspecific factors associated with the membranes.

DISCUSSION

Inhibition of growth cone extension has been shown to have a fundamental role in the formation of the nerve trunks as axons emerge from the neural tube and migrate through the somites (Tosney, 1988a, b). However, little is known about the role of contact-mediated inhibition in the development of nerves at later stages. In this study, we present in vitro evidence of bidirectional inhibitory contact-mediated interactions between the nerves and metanephros of the developing chicken embryo. We suggest these interactions are important to the maintenance of nerve structure during a period when the metanephros expands and invades the lumbosacral region by ensuring that extending growth cones do not invade the metanephric tissue and the metanephric tissue does not invade the nerve.

Prolonged coculturing of metanephric explants and E10 DRG explants resulted in metanephric cells occupying the majority of the culturing substratum and the displacement of axons from that territory. The loss of territory available to axons resulted in axonal fasciculation and the formation of large nerve-like structures. Overall, this observation is consistent with the results of Snow et al. (2003), who demonstrated that axons fasciculate as they extend onto a substratum that contains low levels of axons extension inhibitory molecules (chondroitin sulfate proteoglycans). The formation of nerve-like structures in our experiments was likely due to axons selecting other axons as a preferred substratum in the absence of available culturing substratum due to occupation of the territory by metanephric cells. Similarly, Schwann cells prefer axonal surfaces to the culturing substratum (Seilheimer et al., 1989) and, thus, preferentially would become associated with the axons as the axons form fascicles. This model system may be regarded as an in vitro self-organizing system between different cell types and could be used to better understand the dynamics of interactions between multiple cell types at the system level as previously demonstrated for epithelial, mesenchymal, and tumor cells (Honda et al., 1996; Jiang et al., 1999; Deisboeck et al., 2001).

Contact of metanephric cells with growth cones or axons did not negatively affect metanephric cell migration. However, contact of metanephric cell leading edges with the surfaces of nerve-derived Schwann cells resulted in an inhibition of metanephric protrusive activity. Protrusion of the leading edge is a required step in cell migration (Ridley et al., 2003). Thus, Schwann cells block metanephric cell migration in a contact-dependent manner by inhibiting the first step in the migration process, lamellipodial protrusion. We suggest this inhibitory mechanism operates in vivo during the period of metanephric expansion into the lumbosacral region to prevent metanephric cell invasion of the lumbosacral nerves. A requirement for a mechanism to prevent metanephric cell invasion of nerves is emphasized by the lack of a mature epineurium during the developmental times when metanephric cells have the opportunity to invade nerves (Du Plessis et al., 1996).

This study revealed that one, or more, inhibitory signals present on the surfaces of Schwann cells blocks metanephric lamellipodial protrusion. Lamellipodial protrusion is the first step in cell migration (Ridley et al., 2003). Thus, the signal present on Schwann cells acts to halt migration at an early stage. This effect is similar to that of growth cone collapsing factors that also block lamellipodial and filopodial protrusion (reviewed in Gallo and Letourneau, 2004). It will be of interest to identify the Schwann cell–derived signals at the molecular level and determine whether they belong to a previously identified class of growth cone collapsing factors (e.g., semaphorins).

DRG growth cones become sensitive to metanephric cell-mediated contact inhibition during the developmental time window when the metanephros expands into the lumbosacral region. On the other hand, metanephric tissue from the earliest developmental time point investigated (E7) is inhibitory to older DRG growth cones (E10), indicating that the metanephros exhibits inhibitory signals from its genesis. These observations suggest an up-regulation of receptors in DRG neurons for the inhibitory signal present on the surfaces of metanephric cells. Inhibition of DRG axon extension by metanephric cells may reflect the existence of a mechanism that prevents inappropriate extension of axons out of nerves into the metanephros at a time when the epineurium has not yet matured (Du Plessis et al., 1996).

This study provides a novel model system for studying the mechanism of cell-cell–mediated contact inhibition of migration. To our knowledge, this is the first report of Schwann cell mediated inhibition of the migration of another cell type. Future investigations will address the molecular nature of the cell surface inhibitory signals present on metanephric and Schwann cells. Semaphorin 3A and 3F are expressed in the developing kidney (Villegas and Tufros, 2002), and semaphorins have well-established repellant guidance roles for DRG axons (Luo et al., 1993). Of interest, chicken DRG increase expression of neuropilins, semaphorin receptors, between E7 and E10 (Pond et al., 2002), an observation consistent with the development of DRG sensitivity to metanephric cells in vitro. However, metanephric cells also repelled E7 retinal ganglion cell axons that are not responsive to semaphorin 3A (Luo et al., 1993). Thus, although a member of the semaphorin family may be involved in DRG responses to metanephric cells, semaphorin 3A is not a likely candidate. An additional argument against semaphorin 3A is that it is predominantly in secreted form (Luo et al., 1993) and, thus, would not be expected to act in a contact-dependent manner. The identification of Schwann cell–expressed molecules that inhibit the migration of metanephric cells may have implications for the treatment of metastatic kidney cancer (Fishman and Antonia, 2003).

EXPERIMENTAL PROCEDURES

Cell and Explant Culture

F12H medium containing additives (Gallo and Letourneau, 1999) was used for all culturing on laminin coated glass coverslips (25 μg/ml in phosphate buffered saline [PBS] overnight incubation at 39°C). Chicken E6–E10 DRG, retinal, sympathetic, and ciliary explant cultures were prepared using standard methods. Metanephric explants were obtained by removing the whole of the metanephros from E7–E10 embryos and by manual sectioning, making square explants approximately 300–400 μm in length. Before sectioning, the metanephric tissue was cleared of overlying tissues not of metanephric origin. Explants of sciatic nerve were obtained by dissection of approximately 5 mm of nerve starting at the plexus region. Tissues overlying the nerve were removed after dissection and explants of nerve cut into segments approximately 300–400 μm in length.

Dissociated cells were obtained from metanephric or sciatic nerve explants by using a standardized dissociation protocol (as described in Gallo and Letourneau, 1999). Briefly, tissues were incubated in calcium-magnesium–free PBS (CMF-PBS) for 10 min, followed by incubation in 0.2% trypsin in CMF-PBS for 12 min. Trypsin was blocked by the addition of medium containing 10% calf serum, and cells were dissociated by mechanical trituration.

Video Microscopy

Time-lapse imaging of live DRG, sciatic nerve, and metanephric cells was performed on either a 135M (DRG/metanephric interactions, ×20 phase objective) or 200M (metanephric/nerve cell interactions, ×40 phase objective) Zeiss inverted microscope. The 135M and 200M microscopes are equipped with AxioCam and Orca ER CCD cameras, respectively. Each microscope and camera is in series with a computer running AxioVision software for image acquisition and analysis. The stage was heated by using an air curtain incubator (NevTek, Burnsville, VA).

The responses of growth cones to contact with cells migrating from metanephric explants were classified as continued extension, stalling, or retraction. Continued extension means that the growth cone continued to extend and grew onto the metanephric cells. Stalling refers to the growth cone stopping after contact and no longer extending or undergoing retraction away from the contact. Retraction means that, after contact with the metanephric cells, the growth cone collapsed and the axon retracted away from the contact. Only growth cones that contacted metanephric cells 20 min before the end of the time-lapse sequence were counted to allow time for responses to develop.

The dynamics of the perimeter of metanephric cell lamellipodia interacting with Schwann cell membranes was analyzed by using differential image analysis as described for growth cone lamellipodia in Gallo (1998). Metanephric explants were labeled with DiI and cultured overnight. The next day, freshly dissociated E10 sciatic nerve cells were added to the cultures and time-lapse acquisition begun 3–4 hr later, a time when the nerve-derived cells had attached to the substratum. Briefly, the extent of perimeter contact between the cells was determined by visual inspection, and the stretch of perimeter that made contact was noted and used as a reference frame for the analysis. The same length of perimeter was then analyzed during the 5 min before and after the contact between the cells. The percentage of perimeter undergoing protrusion, retraction, or remaining stable was then determined for each time-lapse interval. For each of six interactions, the 10 measurements for each lamellipodial behavior (protrusion, stable, or retracting) were then collapsed into a mean percentage for the cell, before and after contact, respectively. The means of each category for each cell were then used to perform statistical analysis, yielding a sample size of six cells in a matched before-and-after experiment.

Cell Labeling and Immunocytochemistry

Dissociated metanephric cells were labeled with CellTracker Green (Molecular Probes, Eugene, OR) by incubating the dissociated cells in suspension with 2.5 μM CellTracker in culturing medium for 40 min. Dissociated cells were then pelleted by centrifugation, the CellTracker containing medium was removed, and the cell pellet was washed twice with PBS before plating the cells. Similarly, metanephric explants were labeled with 0.2 μM Vybrant DiI (Molecular Probes) in PBS for 5 min at 39°C and 15 min at 4°C. The explants were then pelleted and resuspended by slight mechanical trituration twice with PBS before culturing. This protocol produced explants on the order of approximately 100–200 μm in size.

Actin filament staining and immunocytochemistry was performed essentially as described in Gallo and Letourneau (1999). To label axons, we used an antibody to a monoclonal neuron-specific isoform of tubulin (βIII; Covance, Berkeley, CA). Cultures were fixed using 0.25% glutaraldehyde for 15 min, washed with PBS, treated with 2 mg/ml sodium borohydride for 15 min, washed again in PBS, and blocked for 30 min in 10% goat serum PBS containing 0.1% Triton X-100. A 1:500 dilution of the primary antibody was applied for 45 min followed by washing in PBS. Fluorescein isothiocyanate–conjugated goat anti-mouse secondary antibody (Sigma) was applied at 1:400 for 45 min. After washing with PBS, coverslips or video dishes were mounted with NoFade mountant and stored at −20°C.

Cells were stained with antibody to a Schwann cell specific marker (P0) after fixation with 8% paraformaldehyde and 10% sucrose in PBS followed by washing in PBS and blocking with 10% goat serum. Undiluted primary antibody (1E8 supernatant, Developmental Studies Hybridoma Bank) was applied for 45 min followed by washing in PBS and treatment with 1:100 secondary fluorescein-conjugated goat anti-mouse (Sigma) for 45 min. Omission of primary antibody gave no staining.

All fluorescent image acquisition was performed by using a 200M Zeiss inverted microscope (as described in the previous sections). Images were stored digitally and prepared for presentation by using Adobe Photoshop.

Preparation of Membrane Particles and Membrane Spots

Crude E10 metanephric membrane particles were generated by using the protocol of deBlaquiere and Burgess (1999). Membranes were routinely prepared by using a total of six metanephros obtained from three embryo dissections. To form spots of membrane particles, we first coated glass coverslips with polylysine overnight (39°C incubation) in borate buffer as previously described (Gallo, 1998), followed by two washes with PBS and one with deionized water. The coverslips were then allowed to air-dry. The membrane preparation (2.5 μl with a protein concentration of 1 μg/μl; determined by using Bio-Rad Protein Assay, Bio-Rad, Hercules, CA) was then applied to the center of the polylysine-coated coverslips (100 μg/ml overnight incubation) and allowed to air-dry. This resulted in a carpet of membranes, attached to the polylysine substratum, with a roughly circular spot-like shape. After drying, culturing medium was added to the coverslips and DRG explants placed in the vicinity (300–600 μm) of the border of the membrane spot or on the membranes, and cultured overnight.

Acknowledgements

We thank Dr. Y.J. Son (Drexel University) for advice with the culturing of Schwann cells. This research was supported by faculty start up funds to G.G. from Drexel University.

Ancillary