¹Department of Cell Biology, Neurobiology and Anatomy, University of Cincinnati College of Medicine, 231 Bethesda Avenue, Cincinnati, Ohio, OH 45267-0521, USA

²Institut für Genetik, Forschungszentrum Karlsruhe, Postfach 3640, 76021 Karlsruhe, Germany

Correspondence to: Larry Sherman, Department of Cell Biology, Neurobiology and Anatomy, University of Cincinnati College of Medicine, 231 Bethesda Avenue, Cincinnati, Ohio, OH 45267-0521, USA

The neu/erbB2 protooncogene is overexpressed in numerous human cancers and is mutationally activated in N-ethyl-N-nitrosourea (ENU)-induced rodent tumors of the Schwann cell lineage. We investigated whether expression of activated neu in Schwann cells is sufficient to initiate their immortalization and transformation. Clones of embryonic dorsal root ganglia cells infected with a retrovirus bearing activated neu (NID cells) were selected based on their expression of Schwann cell-specific markers. Compared to embryonic Schwann cells infected with a virus encoding empty vector, we found that NID cells have altered shapes and disorganized cytoskeletons, grow in the absence of growth factors required for normal Schwann cell survival and proliferation, and can be repeatedly passaged. Furthermore, NID cells are invasive in an in vitro matrix invasion assay and form metastatic tumors when injected into syngeneic animals. The neu-induced growth and invasive phenotypes could be reversed by drugs that inhibit Ras and Src activity. Interestingly, later stage Schwann cells infected with activated neu failed to become immortalized. These findings indicate that constitutive activation of erbB2 is sufficient to initiate the immortalization and transformation of immature Schwann cells, and support the notion that Schwann cells have particular developmental stages during which they are susceptible to immortalizing and transforming events.

Schwann cells; neu; erbB2; transformation; immortalization

Introduction

Tumor progression is believed to be a multi-step process, whereby cells acquire new functions as a result of mutations in particular genes (reviewed by Bishop, 1987; Weinberg, 1989; Vogelstein and Kinzler, 1993). In in vitro models of tumor progression, one of these steps is cellular immortalization, in which cells gain unlimited growth potential, morphological changes and chromosome abnormalities (Namba et al., 1996; Newbold et al., 1993; McCormick and Maher, 1988). Immortalized cells can then undergo neoplastic transformation following subsequent genetic alterations (reviewed by Rhim et al., 1990).

The rat proto-oncogene neu and its human homologue erbB2 (also referred to as HER-2) encode a cell surface tyrosine kinase receptor (erbB2) whose expression is frequently elevated in human tumors of epithelial origin (reviewed in Stancovski et al., 1994; Dougall et al., 1994; Hynes and Stern, 1994; Tzahar and Yarden, 1998). While erbB2 has no known ligands of its own, it becomes activated following heterodimerization with erbB1, erbB3, or erbB4, which bind members of the EGF family and neuregulins (Goldman et al., 1990; Wada et al., 1990; Pinkas-Kramarski et al., 1996; Tzahar et al., 1996; Graus-Porta et al., 1997). In immortalized cell lines, erbB2 overexpression has potent transforming activity (Di Fiore et al., 1987; Hudziak et al., 1987; Di Marco et al., 1990), likely due to its ability to cooperate with other erbB receptors.

Activating neu mutations occur with high frequency in chemically induced rodent tumors. In particular, immature cells of the rat Schwann cell lineage exposed transplacentally or at early neonatal stages to N-ethyl-N-nitrosourea (ENU) reproducibly have a specific activating transversion mutation at the transmembrane region of neu (Bargmann and Weinberg, 1988; Weiner et al., 1989; Nikitin et al., 1991). Schwann cells are the major supportive cell population in the peripheral nervous system and ensheath all peripheral axons. Axon-derived neuregulins are required for Schwann cell survival and proliferation (Marchionni et al., 1993; Morrissey et al., 1995; Dong et al., 1995; Levi et al., 1995) and function in Schwann cells by way of erbB2-erbB3 heterodimers (Cohen et al., 1992; Jin et al., 1993; Carroll et al., 1997; Vartanian et al., 1997). However, ENU-induced schwannoma cells survive and proliferate independent of neuregulins and form malignant tumors.

It is unclear from the ENU studies described above whether activating neu mutations are sufficient for Schwann cell transformation. Indeed, Perantoni and co-workers (1987) demonstrated that ENU-induced schwannomas can also have N-ras mutations. However, expression of a wild-type neu transgene under the control of a Schwann cell-specific promoter reduces ENU-induced oncogenesis in vivo and suppresses the tumorigenic phenotype of homozygous mutant neu schwannoma cells in vitro (Nikitin et al., 1996). It is therefore possible that while ENU can induce mutations in other oncogenes, activating mutations in neu are the single event required for the initiation of Schwann cell immortalization and transformation.

In the present study, we directly determined whether immature rat Schwann cells could undergo malignant transformation following retroviral introduction of activated neu. We found that rat embryo Schwann cells expressing activated neu did not establish normal interactions with axons from sensory neurons and grew in the absence of axon-derived growth factors. Furthermore, these cells demonstrated altered cytoskeletal organization, were immortalized and invasive in vitro, and formed slow-growing malignant tumors in syngeneic animals. These data indicate that activating neu mutations are sufficient to initiate Schwann cell tumorigenesis, and suggest that activation of a single oncogene can result in both cellular immortalization and transformation.

Results

neu-infected Schwann cells have altered morphology and do not adhere to axonsM

To investigate whether overexpression of mutant erbB2 is sufficient for Schwann cell immortalization and transformation, we infected dissociated rat E16 embryonic dorsal root ganglia (DRG) cultures (containing neurons, Schwann cells, and fibroblasts) with a retrovirus encoding activated neu. Twenty clones of G418-resistant neu-infected DRG (NID) cells and five clones infected with vector alone were selected and assayed for expression of active erbB2 by Western blotting with an antibody specific for the phosphorylated form of erbB2. Three clones having similar levels of erbB2 activity were selected for further analysis and compared to vector controls (Figure 1a). These three clones and controls were each positive for Schwann cell markers S100 protein and the low affinity nerve growth factor receptor (LNGFR) (Figure 1b - e), confirming that they were derived from the Schwann cell lineage.

In culture, normal Schwann cells demonstrate a spindle-shaped morphology and have a limited amount of well-organized F-actin (Figure 2a,b). NID cells, however, had a more variable morphology and generally appeared flattened and spread-out compared to normal Schwann cells, with numerous small filipodia and occasionally longer extensions (Figure 2c,d). NID cells were also larger than normal Schwann cells and demonstrated numerous actin stress fibers (Figure 2c,d).

A characteristic feature of normal Schwann cells is that they selectively adhere to neurites in Schwann cell/neuron co-cultures (Salzer et al., 1980; Kleitman et al., 1991). To test whether NID cells similarly adhere to neurites, we cultured either normal rat Schwann cells or NID cells in the presence of purified cultures of rat DRG neurons. As shown in Figure 2e,f, whereas normal Schwann cells had aligned along fascicles of neurites in these cultures, NID cells adhered preferentially to the culture dish. These data indicate that neu-infected Schwann cells have lost the ability to form normal contacts with axons.

NID cells are immortalized and grow in the absence of added growth factors

In vivo, normal Schwann cells require signals from axons for survival and for proliferation (Dong et al., 1995; Salzer et al., 1980; Grinspan et al., 1996). Members of the neuregulin protein family, including glial growth factor (GGF), are among these signals, and can stimulate Schwann cell proliferation and survival in vitro (Marchionni et al., 1993; Morrissey et al., 1995; Dong et al., 1995; Levi et al., 1995). Since erbB2 is phosphorylated in response to neuregulins, we tested whether NID cells, in which erbB2 is constitutively phosphorylated, grow continuously in the absence of added growth factors. After 24 h in serum and growth factor-free medium, NID cells and vector controls were cultured for an additional 24 h in the presence and absence of 10 ng/ml of rh-GGF2. Cells were pulsed with [³H]thymidine for the final 6 h in culture, then analysed for [³H]thymidine incorporation using a scintillation counter. Schwann cells infected with vector alone incorporated little or no [³H]thymidine in the absence of rh-GGF2 while all three NID clones (each at passage number four) incorporated significant amounts of [³H]thymidine (Figure 3). Under these conditions, NID cells have a doubling time of approximately 24 - 28 h (data not shown). Although there were likely to be twice as many NID cells as Schwann cells at the time of pulsing, the levels of [³H]thymidine uptake by NID cells were more than tenfold greater than for Schwann cells. Consistent with previous findings (Kim et al., 1997a) vector-infected Schwann cells demonstrated a significant (P<0.0001) increase in [³H]thymidine incorporation in the presence of rh-GGF2, compared to untreated controls (Figure 3). NID cells, however, responded variably to rh-GG2 (P<0.08; Figure 3). These data indicate that NID cells can grow in the absence of factors normally required for Schwann cell proliferation and survival, even though they can still respond to these factors.

We next determined if NID cells could be repeatedly passaged in the absence of GGF. We observed extensive cell death in E16 Schwann cell cultures grown in medium containing 10% fetal bovine serum by passage 3. NID clones, however, could be continually passaged under the same culture conditions without any signs of cell death (not shown). As shown in Figure 3, one NID clone (NID clone 2) examined at passage 37 incorporated high levels of [³H]thymidine in the absence of rh-GGF2. The levels of [³H]thymidine incorporation were similar to that of early passage NID2 cells (Figure 3). At present, all three of the NID lines have been passaged >50 times. These findings indicate that NID cells are immortalized.

The ability of ENU to induce rat schwannomas depends on the developmental stage at which the cells are exposed. A variety of central and peripheral nervous system tumors, including schwannomas, will form in rats following transplacental or direct treatment with ENU between embryonic day 15 and early postnatal stages (Nikitin et al., 1991; Perantoni et al., 1987; Rajewsky et al., 1983; Schubert et al., 1974; Druckrey, 1973). We tested the possibility that the capacity of activated erbB2 to immortalize and transform Schwann cells is similarly dependent on the developmental stage of the cells. We infected dissociated rat P3 dorsal root ganglia (DRG) with the activated neu retrovirus, then selected clones in the presence of G418 as described above. The resulting clones grew in the absence of rh-GGF2 (data not shown). However, unlike the infected embryonic Schwann cells, none of the 20 neonatal clones selected survived past passage six, and most died by passage four (Table 1). These data are consistent with the notion that activated erbB2 is only sufficient to initiate the transformation and immortalization of immature Schwann cells.

NID cells are invasive in vitro

A crucial step in the metastatic process is when tumor cells invade tissues and break away from their local microenvironment. We tested whether NID cells are invasive using a quantitative in vitro invasion assay (Hennigan et al., 1994; Lamb et al., 1997). In this assay, a 90 m thick, solidified layer of growth factor-depleted Matrigel is placed in the well of a transwell filter insert. Cells are plated on the bottom of the filter, which has 8 m pores. Motile cells travel upwards through filter pores to the top of the filter. Non-invasive cells are impeded by the Matrigel at this point, while invasive cells penetrate and migrate into the Matrigel. Cells that are more invasive travel further into the Matrigel, as measured by laser confocal microscopy. The percentage of cells at a given optical slice is determined by dividing the number of cells at that level by the total number of cells in all optical slices counted in that microscopic field of the transwell.

We compared the invasive potential of NID cells with vector-infected control Schwann cells and with the ENU-induced RT4-D6P2T rat schwannoma cell line (Tomozawa and Sueoka, 1978). After 4 days, a large proportion of each type of cell had migrated across the filter (14.7±4% of Schwann cells; 23.3±4.8% of NID cells, see Figure 4a,b,e; 27.7±2.7% of RT4-D6P2T cells; see Figure 4e). However, unlike the vector control Schwann cells, the RT4-D6P2T and NID cells invaded deep into the Matrigel (Figure 4c - e). No normal Schwann cells were detected between 30 m and 90 m into the Matrigel. In contrast, 13.6±1.6% of NID cells and 24.9±2.5% of RT4-D6P2T cells were found at these deeper layers. Similar results were obtained in three separate experiments. These findings indicate that both ENU-induced schwannomas and neu-infected Schwann cells are invasive.

NID cells are tumorigenic in syngeneic animals

To test whether NID cells are tumorigenic in vivo, male BDX rats were injected subcutaneously with 5´10⁵ NID (clone 2) cells. After a long latency period, tumors began to grow slowly at the site of injection. Twelve weeks after the injection, the average volume of the tumors was 1326±499 mm³ (s.e.). After 16 weeks, two animals had opened their tumors and were sacrificed. The average tumor volume of the remaining six animals was 5227±2137 mm³ (s.e.). Four additional animals eventually opened their tumor and had to be sacrificed. One animal's tumor grew over the German legal limit and was therefore killed 17 weeks after receiving tumor cells. The remaining animal's tumor had a volume of about 17 800 mm³ for several weeks and was killed 26 weeks after being injected with tumor cells. No overt signs of metastatic disease were detectable in any of the animals. Histologically, the NID tumors contained encapsulated groups of elongated cells that often formed fascicles (Figure 5a). Similar results were obtained with two additional NID clones. Cells re-isolated from these tumors maintained expression of S100 protein (data not shown), consistent with the cells still being part of the Schwann cell lineage.

To determine the stability of NID cell tumorigenicity, cells from one of the NID tumors were re-introduced into tissue culture under neomycin selection, and after two passages 5´10⁵ cells were injected subcutaneously into BDX rats as before. The kinetics of tumor growth in these animals was essentially similar to the first experiment, with a latency period of 6 weeks and an average tumor volume of 977 mm³±162 mm³ (s.e.) 11 weeks after tumor cell injection. These data demonstrate that the tumorigenic potential of NID cells is stable and that more aggressive cell lineages were not selected as a result of animal passaging. In these experiments all of the animals eventually opened their tumor before the tumor reached the German legal limit and were sacrificed. Again, no overt signs of metastatic disease were detectable in any of the animals.

NID cells are metastatic in vivo

Upon autopsy, the lymph nodes draining the primary NID tumor were enlarged in some animals. Such enlargement could result from either slow growing metastatic deposits or infection following self-inflicted tumor wounding. We found small lymph node metastases in three out of nine animals from the second experiment described above. Unlike the primary tumors, the histologic appearance of these lymph node tumors was a highly disorganized meshwork of cells and matrix (Figure 5b). These data are consistent with our finding that NID cells are invasive in vitro.

The phenotype of NID cells is reversed by inhibitors of Ras and Src

ErbB2 can activate the Ras-MAP kinase cascade through interactions with the adapter proteins Grb2 and Sos (Janes et al., 1994). Activated ErbB2 can also directly interact with and activate members of the Src family of protein tyrosine kinases (Muthuswamy and Muller, 1995; Luttrell et al., 1994; Muthuswamy et al., 1994). It is therefore likely that signaling cascades involving Ras and Src are involved in maintaining the phenotype of neu-infected Schwann cells. We tested this notion using either an inhibitor of farnesyl protein transferase (L-739,749) to block cellular Ras activity (Kim et al., 1997b; Prendergast et al., 1994; James et al., 1993), or herbimycin A to block src activity (Uehara et al., 1989; Fukazawa et al., 1991). Cells were grown in the presence and absence of either drug for 48 h then assayed for cell proliferation and invasion as described above. We found that both drugs significantly inhibited NID cell [³H]thymidine incorporation (Figure 6a) and cell invasion (Figure 6b) at concentrations previously shown to block Ras processing (Kim et al., 1997b) and Src kinase activity (Sobko et al., 1998) in Schwann cells. Both drugs inhibited cell invasion to a similar degree, while herbimycin A also inhibited cell motility as indicated by the small number of cells that reached the bottom of the filter (Figure 6b). Herbimycin A was also a more potent inhibitor of cell proliferation, reducing NID [³H]thymidine incorporation by approximately 83% while L-739,749 only blocked NID proliferation by 37%. These data suggest that activated neu promotes cell invasion and proliferation at least in part through Ras- and Src-dependent signaling cascades.

Discussion

Immortalization and transformation are two steps in tumor progression that likely require distinct genetic alterations in affected cells. Such alterations could, however, be the consequence of a single event, such as a mutation in an oncogene or tumor suppressor gene. Primary cultures of mouse connective tissue cells, for example, will become immortalized and transformed following retroviral infection with the v-fos oncogene (Jenuwein et al., 1985). We have demonstrated that constitutive activation of a single oncogene, neu, is sufficient to initiate the immortalization and malignant transformation of immature Schwann cells. Schwann cells from embryonic dorsal root ganglia infected with a retrovirus carrying activated neu demonstrated abnormal cell shape and cytoskeletal organization, grew independent of added growth factors, and could be repeatedly passaged without apparent phenotypic alterations. Furthermore, NID cells were invasive in vitro and formed malignant tumors in vivo.

The transformed phenotype of NID cells was reversed by inhibitors of Src and Ras. While herbimycin A and L-739,749 may affect additional signaling molecules, they were used at concentrations previously shown to block Src kinase activity and Ras farnesylation (which is required for Ras to become active), respectively, in Schwann cells (Sobko et al., 1998; Kim et al., 1997b). Both drugs were effective in reversing the ability of NID cells to invade matrix and to proliferate in the absence of added growth factors. Our findings agree with previous studies indicating that Ras and Src are essential for erbB2 signaling in mammary carcinoma cells (Luttrell et al., 1994; Janes et al., 1994), and further support the idea that constitutive erbB2 activity is necessary to maintain the transformed phenotype of rat schwannoma cells.

The ability of neu to immortalize and transform Schwann cells may depend on the state of Schwann cell differentiation at the time the active gene is introduced. While we obtained multiple immortalized clones of neu-infected cells from embryonic DRG, none of the clones from neonatal DRG cells became immortalized. Interestingly, the neurooncogenic effect of ENU in rats is also dependent on the developmental stage at which the drug is administered, with the maximum yield of schwannomas occurring following drug exposure during early postnatal development (Druckrey, 1973; Rajewsky et al., 1977; Perantoni et al., 1987; Nikitin et al., 1991). It is possible that the levels of endogenous erbB2 or erbB3 dictate whether or not Schwann cells become transformed following exposure to ENU. The developmental window in which ENU exposure results in the maximum yield of malignant schwannomas corresponds with the time of maximum neu gene transcription (Jin et al., 1993). In contrast to the in vivo studies of ENU-induced schwannomas, our data suggest that transformation by directly introducing activated erbB2 into Schwann cells either does not occur during neonatal development or only occurs with very low frequency. Additional alterations may therefore be required for the transformation of more mature Schwann cell populations. However, our findings clearly show that erbB2 activation is sufficient to initiate immature Schwann cell carcinogenesis, and suggest that the developmental state of Schwann cells may determine their susceptibility to tumor-initiating alterations in signal transduction pathways.

Materials and methods

Selection of neu-infected DRG (NID) cell clones

Dorsal root ganglia were aseptically removed from either E16 or P3 BDX or Sprague-Dawley rats. Between 150 - 200 ganglia were dissociated in trypsin-EDTA for 30 min at 37°C, triturated, pelleted by centrifugation at 1000 r.p.m. for 5 min, then resuspended in DMEM+10% heat-inactivated FBS. Cells were plated on 60 mm tissue culture dishes and incubated at 37°C in a humidified 5% CO₂ atmosphere. After 2 h, medium was replaced with conditioned medium from the retrovirus producer cell lines NeuINA -2 or IFN -2 (as described by Bradbury et al., 1993) plus 4 g/ml of polybrene. After 6 h the conditioned medium was replaced with fresh DMEM with serum. Infected cells were cultured for 48 h then treated with 700 g/ml G418. For neu-infected cells, 20 stable clones were selected, expanded, and tested for the presence of constitutively activated erbB2 by Western blotting, using an anti-phospho-erbB2 antibody (Upstate Biotechnology) as described in the manufacturer's instructions. For cells infected with the INA vector control virus, between 5 - 10 clones were selected and expanded for each of the experiments described below.

Immunocytochemistry and actin staining

Immunocytochemical labeling of cells with the Schwann cell markers S100 and LNGFR, and actin staining, were performed as previously described (Pelton et al., 1998). Briefly, cells were fixed with 4% paraformaldhyde in PBS for 20 min at room temperature. For S100 protein, cells were permeabilized by blocking in PBS with 10% normal goat serum and 0.2% Triton X-100 for 30 min. For LNGFR, cells were blocked without detergent. Cells were incubated with S100 antibody (S100, 1 : 6000, Dako) or LNGFR antibody (217c, 1 : 200, Peng et al., 1982) for 1 h at room temperature. Cells were then washed three times with blocking buffer and incubated for 30 min with a 1 : 200 dilution of FITC-conjugated-goat-anti-rabbit IgG (for S100) or FITC-conjugated-goat-anti-mouse-IgG (for LNGFR) (Jackson Labs). To visualize actin filaments cells were permeabilized as described above then incubated with 4 U/ml of BODIPY-phalloidin (Molecular Probes) for 30 min at room temperature. After the final antibody and phalloidin incubations, cells were washed three times with blocking buffer then mounted in Fluoromount G (EM Sciences). Cells were viewed and photographed on a Zeiss Axiophot microscope.

Neuron-Schwann cell co-cultures

Dissociated rat E15 DRG were cultured in the presence of anti-mitotic drugs to kill dividing cells as previously described (Kleitman et al., 1991). Neurons were maintained in DMEM plus 10% human placental serum and 50 ng/ml NGF (Harlan). Approximately 1´10⁵ vector control Schwann cells or NID cells were seeded with the established neurons. After 2 days, cultures were analysed by microscopy to determine whether the seeded cells preferentially bound to neurites.

[³H]thymidine incorporation assay

Vector transfected Schwann cells or NID cells were plated at 5´10⁴ cells/well in triplicate in DMEM without serum in 24 well tissue plates (Costar) for 24 h and treated with growth factors or left untreated for an additional 24 h. Cells were pulsed with 1 Ci of [³H]thymidine for the final 6 h in culture. Labeled cells were washed three times in PBS and solubilized in 200 l of 0.2 M NaOH as previously described (Sherman et al., 1997). Mean CPM were then determined in a scintillation counter.

Drug treatments

Cells were treated with 5 M herbimycin A (Sigma; Sobko et al., 1998) for 16 - 24 h or with 10 M L739,749 (Merck; Kim et al., 1997b) for 48 h prior to proliferation and invasion assays.

In vitro invasion assay

Cell invasion was measured as previously described (Hennigan et al., 1994; Lamb et al., 1997). Briefly, 80 l of growth factor-depleted Matrigel (Becton-Dickinson) diluted 1 : 1 with sterile, ice cold PBS was plated into 6.5 mm diameter, 8 m pore size transwell filter inserts (Costar). Transwells were then incubated at 37°C to solidify the Matrigel, inverted, and a 100 l drop containing 1´10⁶ cells/ml was plated on the top of the inverted filter. After 1 h, the transwells were placed in 24 well plates with serum free DMEM and incubated for 3 days. Transwells were then rinsed in PBS and incubated in ice cold methanol for 15 min at -20°C, dried, washed with PBS, and stained with 10 g/ml propidium iodide for 15 min at room temperature. Stained cells were then washed and analysed using a Bio-Rad MRC-600 laser confocal microscope. Invasion was quantified by measuring the number of propidium ioide labeled nuclei on the top and bottom of the transwell filter, and at 10 m intervals within the Matrigel. For each test condition, data are represented as the means of three microscopic fields´three transwells. The percentage of cells at any given optical slice ws determined as the total number of cells observed at that level divided by the total number of cells counted at all levels of a particular microscopic field of the transwell.

Tumorigenicity assay

Eight 12-week-old male BDX rats (from the animal facility at the Institut für Genetik, Forschungszentrum Karlsruhe) were injected subcutaneously in the left flank with 5´10⁵ NID cells suspended in 100 l PBS. Animals were regularly monitored for tumor growth and tumor volume was measured using calipers. When rats became moribund they were sacrificed and autopsied. Where described, tumor material was aseptically taken and homogenized in PBS using a cell dissociation sieve (Sigma). The resuspended tumor cells were then cultivated in DMEM+10% FBS containing 700 g/ml G418. Tumor, lymph nodes and lung were also taken and fixed in 4% buffered paraformaldehyde and prepared for histochemistry.

Acknowledgements

We thank Tanja Schüler and Saikumar Karyala for technical assistance, Norma Howells for assistance with animal breeding, and Helmut Ponta for helpful discussions and ideas. We are also grateful to Paul Edwards from the Department of Pathology at the University of Cambridge for the -2 pINA-neu and -2 pINA cell lines, to Mark Marchionni from Cambridge Neuroscience for the rh-GGF2, to Jackson Gibbs and Nancy Krohl from Merck Sharp and Dohme Research Laboratories for the L-739,749. This work was supported by NIH NS28840 to N Ratner, NRSA NS10297 to L Sherman and He551/8-2 from the Deutsche Forschungsgemeinschaft to P Herrlich.

Bargmann CI and Weinberg RA. (1988). EMBO J. 7, 2043-2052. MEDLINE

Bishop JM. (1987). Science 235, 305-311. MEDLINE

Bradbury JM, Arno J and Edwards PA. (1993). Oncogene 8, 1551-1558. MEDLINE

Carroll SL, Miller ML, Frohnert PW, Kim SS and Corbett JA. (1997). J. Neurosci. 17, 1642-1659. MEDLINE

Cohen JA, Yachnis AT, Arai M, Davis JG and Scherer SS. (1992). J. Neurosci. Res. 31, 622-634. MEDLINE

Di Fiore PP, Pierce JH, Kraus MH, Segatto O, King CR and Aaronson SA. (1987). Science 237, 178-182. MEDLINE

Di Marco E, Pierce JH, Knicley CL and Di Fiore PP. (1990). Mol. Cell. Biol. 10, 3247-3252. MEDLINE

Dong Z, Brennan A, Liu N, Yarden Y, Lefkowitz G, Mirsky R and Jessen KR. (1995). Neuron 15, 585-596. MEDLINE

Dougall WC, Qian X, Peterson NC, Miller MJ, Samanta A and Greene MI. (1994). Oncogene 9, 2109-2123. MEDLINE

Druckrey H, Ivankovic S and Gimmy J. (1973). Z. Krebsforsch. Klin. Onkol. 79, 282-297.

Fukazawa H, Li PM, Yamamoto C, Murakami Y, Mizuno S and Uehara Y. (1991). Biochem. Pharmacol. 42, 1661-1671. MEDLINE

Goldman R, Levy RB, Peles E and Yarden Y. (1990). Biochemistry 29, 11024-11028. MEDLINE

Graus-Porta D, Beerli RR, Daly JM and Hynes NE. (1997). EMBO J. 16, 1647-1655. Article MEDLINE

Grinspan JB, Marchionni MA, Reeves M, Coulaloglou M and Scherer SS. (1996). J. Neurosci. 16, 6107-6118. MEDLINE

Hennigan RF, Hawker KL and Ozanne BW. (1994). Oncogene 9, 3591-3600. MEDLINE

Hudziak RM, Schlessinger J and Ullrich A. (1987). Proc. Natl. Acad. Sci. USA 84, 7159-7163. MEDLINE

Hynes NE and Stern DF. (1994). Biochim. Biophys. Acta 1198, 165-184. MEDLINE

James GL, Goldstein JL, Brown MS, Rawson TE, Somers TC, McDowell RS, Crowley CW, Lucas BK, Levinson AD and Marsters Jr JC. (1993). Science 260, 1937-1942. MEDLINE

Janes PW, Daly RJ, deFazio A and Sutherland RL. (1994). Oncogene 9, 3601-3608. MEDLINE

Jenuwein T, Muller D, Curran T and Muller R. (1985). Cell 41, 629-637. MEDLINE

Jin JJ, Nikitin AY and Rajewsky MF. (1993). Cell Growth Differ. 4, 227-237. MEDLINE

Kim HA, DeClue JE and Ratner N. (1997a). J. Neurosci. Res. 49, 236-247.

Kim HA, Ling B and Ratner N. (1997b). Mol. Cell. Biol. 17, 862-872. MEDLINE

Kleitman N, Wood P and Bunge RP. (1991). In: Culturing Nerve Cells. G Banker and K Goslin. (eds). The MIT Press: Cambridge, MA, pp. 251-278.

Lamb RF, Hennigan RF, Turnbull K, Katsanakis KD, MacKenzie ED, Birnie GD and Ozanne BW. (1997). Mol. Cell. Biol. 17, 963-976. MEDLINE

Levi AD, Bunge RP, Lofgren JA, Meima L, Hefti F, Nikolics K and Sliwkowski MX. (1995). J. Neurosci. 15, 1329-1340. MEDLINE

Luttrell DK, Lee A, Lansing TJ, Crosby RM, Jung KD, Willard D, Luther M, Rodriguez M, Berman J and Gilmer TM. (1994). Proc. Natl. Acad. Sci. USA 91, 83-87. MEDLINE

Marchionni MA, Goodearl AD, Chen MS, Bermingham-McDonogh O, Kirk C, Hendricks M, Danehy F, Misumi D, Sudhalter J, Kobayashi K, et al. (1993). Nature 362, 312-318. MEDLINE

McCormick JJ and Maher VM. (1988). Mutat. Res. 199, 273-291. MEDLINE

Morrissey TK, Levi AD, Nuijens A, Sliwkowski MX and Bunge RP. (1995). Proc. Natl. Acad. Sci. USA 92, 1431-1435. MEDLINE

Muthuswamy SK and Muller WJ. (1995). Oncogene 11, 271-279. MEDLINE

Muthswamy SK, Siegel PM, Dankort DL, Webster MA and Muller WJ. (1994). Mol. Cell. Biol. 14, 735-743. MEDLINE

Namba M, Mihara K and Fushimi K. (1996). Crit. Rev. Oncog. 7, 19-31. MEDLINE

Newbold RF, Cuthbert AP, Themis M, Trott DA, Blair AL and Li W. (1993). Tox. Lett. 67, 211-230.

Nikitin AY, Ballering LA, Lyons J and Rajewsky MF. (1991). Proc. Natl. Acad. Sci. USA 88, 9939-9943. MEDLINE

Nikitin AY, Jin JJ, Papewalis J, Prokopenko SN, Pozharisski KM, Winterhager E, Flesken-Nikitin A and Rajewsky MF. (1996). Oncogene 12, 1309-1317. MEDLINE

Pelton PD, Sherman LS, Rizvi TA, Marchionni MA, Wood P, Friedman RA and Ratner N. (1998). Oncogene 17, 2195-2209. MEDLINE

Peng WW, Bressler JP, Tiffany-Castiglioni E and de Vellis J. (1982). Science 215, 1102-1104. MEDLINE

Pinkas-Kramarski R, Soussan L, Waterman H, Levkowitz G, Alroy I, Klapper L, Lavi S, Seger R, Ratzkin B, Sela M and Yarden Y. (1996). EMBO J. 15, 2452-2467. MEDLINE

Perantoni AO, Rice JM, Reed CD, Watatani M and Wenk ML. (1987). Proc. Natl. Acad. Sci. USA 84, 6317-6321. MEDLINE

Prendergast GC, Davide JP, deSolms SJ, Giuliani EA, Graham SL, Gibbs JB, Oliff A and Kohl NE. (1994). Mol. Cell. Biol. 14, 4193-4202. MEDLINE

Rajewsky MF. (1983). Rec. Results Cancer Res. 84, 63-76.

Rhim JS, Yoo JH, Park JH, Thraves P, Salehi Z and Dritschilo A. (1990). Cancer Res. 50, 5653S-5657S. MEDLINE

Salzer JL and Bunge RP. (1980). J. Cell. Biol. 84, 739-752. MEDLINE

Schubert D, Heinemann S, Carlisle W, Tarikas H, Kimes B, Patrick J, Steinbach JH, Culp W and Brandt BL. (1974). Nature 249, 224-227. MEDLINE

Sherman L, Xu HM, Geist RT, Saporito-Irwin S, Howells N, Ponta H, Herrlich P and Gutmann DH. (1997). Oncogene 15, 2505-2509. MEDLINE

Sobko A, Peretz A and Attali B. (1998). EMBO J. 17, 4723-4734. Article MEDLINE

Stancovski I, Sela M and Yarden Y. (1994). Cancer Treat. Res. 71, 161-191. MEDLINE

Tomozawa Y and Sueoka N. (1978). Proc. Natl. Acad. Sci. USA 75, 6305-6309.

Tzahar E, Waterman H, Chen X, Levkowitz G, Karunagaran D, Lavi S, Ratzkin BJ and Yarden Y. (1996). Mol. Cell. Biol. 16, 5276-5287. MEDLINE

Tzahar E and Yarden Y. (1998). Biochim. Biophys. Acta 1377, M25-M37. MEDLINE

Uehara Y, Fukazawa H, Murakami Y and Mizuno S. (1989). Biochim. Biophys. Res. Commun. 163, 803-809.

Vartanian T, Goodearl A, Viehover A and Fischbach G. (1997). J. Cell Biol. 137, 211-220. MEDLINE

Vogelstein B and Kinzler KW. (1993). Trends Genet. 9, 138-141. MEDLINE

Wada T, Qian XL and Greene MI. (1990). Cell 61, 1339-1347. MEDLINE

Weinberg RA. (1989). Cancer Res. 49, 3713-3721. MEDLINE

Weiner DB, Liu J, Cohen JA, Williams WV and Greene MI. (1989). Nature 339, 230-231. MEDLINE

Figure 1 Schwann cells infected with activated erbB2 maintain expression of Schwann cell markers. (a) Western blot showing expression of phosphorylated erbB2 in neu-infected and vector control embryonic rat DRG cells. Lane 1: a clone infected with vector alone; lanes 2 - 4; three clones of neu-infected DRG cells. Cells were grown in serum-free DMEM for 24 h then 25 g of total protein from each clone was separated by SDS - PAGE and blotted. Blots were probed with a phospho-erbB2 antibody. (b - e) Immunocytochemical staining for (b, c) S-100 protein and (d, e) LNGFR in (b, d) vector infected DRG cells and (c, e) the NID clone shown in lane 2 from (a). Note that both S100 and LNGFR expression are maintained in NID cells, confirming that they are derived from Schwann cells

Figure 2 NID cells have aberrant cytoskeletal organization and fail to make normal contacts with axons. (a - d) Comparison of vector control and NID cell morphology and actin staining. (a, c) Phase contrast photomicrographs of (a) vector control Schwann cells, and (c) NID clone 2 cells. Note that compared to controls, NID cells are large, spread out cells with large nuclei. (b, d) BODIPY-phalloidin staining of the cells in (a) and (c). Note the NID cells have dramatically disorganized actin filaments compared to controls. (e, f) Co-cultures of embryonic DRG neurons with (e) vector control Schwann cells, and (f) NID cells. Arrows indicate locations of control Schwann cells and NID cells. Note that NID cells fail to associate with neurites, while control cells preferentially bind neurites. Similar results were obtained in two additional NID clones

Figure 3 NID cells proliferate in the absence of added growth factors. Vector control Schwann cells (`SC') of NID cells (clone 2) were cultured in serum-free DMEM for 24 h, then in the presence of 15 ng/ml rh-GGF2 for an additional 24 h. Cells were pulsed with [<sup>3</sup>H]thymidine for the final 6 h of culture. Data are expressed as the mean CPM±s.d. of triplicate samples from one assay. This experiment was performed three times using two NID clones with similar results. Data shown here are from a single clone of vector infected Schwann cells and a single clone of neu-infected cells analysed at passage number four and passage number 37 (`NIDp37'). Note that although NID cells had doubled at the time of pulsing (approximate doubling time of 24 - 26 h under these conditions) they incorporated greater than tenfold more [³H]thymidine than Schwann cells in the same culture conditions

Figure 4 NID cells are invasive in vitro. The invasive properties if NID (clone 2) cells were compared with vector control Schwann cells (`SC') and an ENU-induced rat schwannoma cell line (RT4-D6P2T). (a - d) Confocal images (`optical slices') of propidium iodide labeled NID cells after 3 days in culture on the top (a) and bottom (b) of a transwell filter, and at 30 m (c) and 90 m (d) into matrigel within the transwell. Note that labeled NID nuclei can be detected at 30 m but not at 90 m. (e) Mean percentages of cells observed on the top, bottom, and between 20 - 90 m within a Matrigel-containing transwell. Each data point represents the means and standard deviations from three fields each of three transwells. Note that control cells do not invade into the Matrigel, while NID cells and RT4-D6P2T cells are invasive. Similar results were obtained using another NID clone (number 3)

Figure 5 NID cells form metastatic tumors. Tumor tissue from a primary NID tumor (clone 2) (a) and a lymph node secondary tumor (b) were fixed in 4% paraformaldehyde, embedded in paraffin, sectioned, then stained with hematoxylin and eosin using standard histological protocols. The general histological appearance of these tumors is similar to that seen in human schwannomas

Figure 6 Inhibitors of Src kinase activity and Ras farnesylation inhibit NID cell proliferation and invasion. (a) Cell (NID clone 2) were grown in serum-free medium for 48 h in the presence of 5 M herbimycin A (Herb A), 10 M L739,749, or an equivalent volume of vehicle (DMSO) as a control. Cells were pulsed with [³H]thymidine as above, then analysed by scintillation counting. Note that both drugs inhibited NID cells [³H]thymidine incorporation. (b) In vitro invasion assay of herbimycin A and L739,749 treated NID cells. Cells were treated with either drug for 48 h, then plated on transwell filters as described above. Note that herbimycin A-treated cells failed to migrate through the filter pores and did not invade the Matrigel. Cells treated with L739,749 migrated as well as controls to the top of the filter, but also failed to invade the Matrigel

Table 1 Late passage survival of P3 DRG cell clones following infection with activated neu