Introduction

Neuroblastoma (NB) is a neural crest-derived childhood tumor that predominantly arises in the adrenal gland (Seeger and Reynolds, 1993). These tumors are generally very heterogeneneous with neuronal and glial components, yet highly metastatic tumors exhibit an undifferentiated morphology with a high mitotic index (Shimada et al., 1984). The N-myc gene, a close relative of the cell cycle regulating c-myc gene, is amplified in a subset of aggressive NB tumors and amplification correlates with poor prognosis (Brodeur, 2003). Interestingly, spontaneous tumor regression occurs in some NB, despite high N-MYC expression and widespread metastasis (Westermann and Schwab, 2002), suggesting the preservation of functional differentiation promoting signal pathways in these malignancies. Identification of these signaling pathways may lead to new treatment modalities for patients with malignant NB.

The development of the nervous system is guided by the interactions of trophic signals with cell surface receptors, thereby controlling cell determination and cell fate (Anderson et al., 1997). These receptor–ligand interactions activate complex intracellular pathways, which may regulate cell proliferation, cell differentiation and/or cell death. Many trophic factors, including neurotrophins (NTs) and the GDNF (glial cell line-derived neurotrophic factor) family ligands (GFLs) have been identified and their role(s) in the development of the central (CNS) and the peripheral (PNS) nervous systems has been demonstrated. The family of NTs includes nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), neurotrophin 3 (NT3) and NT4/5. These NTs interact with two classes of membrane bound receptors, namely the common NT receptor p75NTR, a member of the tumor necrosis factor (TNF) receptor gene family that is expressed in neuronal and non-neuronal tissues (Dechant and Barde, 2002; Hempstead, 2002; Roux and Barker, 2002), and members of the TRK gene family, which are expressed in the developing and mature nervous system (Huang and Reichardt, 2001; Lee et al., 2001). The functions of p75NTR include, but are not limited to, induction of apoptosis or survival, regulation of axon myelination and axon regeneration (Barrett, 2000; Beattie et al., 2002; Cosgaya et al., 2002; Dechant and Barde, 2002; Wang et al., 2002). The role of p75NTR in NB is unclear, yet the receptor can complex with TRKA to strengthen the NGF signal, which may contribute to NB differentiation.

The biological consequences of TRK receptor activation range from cell proliferation to cell differentiation and maintenance of the differentiated state, depending on the target cell analysed (Bothwell, 1995). Interestingly, trkA gene expression is lost in highly malignant NB tumors and directly correlates with poor patient prognosis, implicating this differentiation pathway in NB progression (Nakagawara et al., 1993; Miyake et al., 1994; Brodeur, 2003). However, activation of a genetically reconstituted TRKA receptor by NGF resulted in terminally differentiated HTLA230 NB cells, demonstrating that a functional TRKA/NGF signal pathway may override the genetic consequences of N-myc gene amplification (Matsushima and Bogenmann, 1993a). Ciliary neurotrophic factor (CNTF), a trophic factor that promotes cell survival and/or differentiation in a variety of neuronal cells including sensory, sympathetic and motor neurons (Sleeman et al., 2000) activates transcription of the silenced trkA gene in HTLA230 cells, yet only a fraction of tumor cells differentiate in the presence of NGF, suggesting the requirement for additional trophic signals to further NB differentiation. (Bogenmann et al., 1998)

GDNF, neurturin (NTN), artemin and persephin belong to the GFL family, a distant relative of the transforming growth factor (TGF) gene family (Airaksinen and Saarma, 2002). GFLs promote survival and differentiation of various neurons through activation of the membrane tyrosine kinase receptor, RET (Takahashi, 2001; Airaksinen and Saarma, 2002). Germline mutations that generate an activated form of RET can be oncogenic in derivatives of neural crest cells leading to thyroid carcinoma and pheochromocytoma among others (Takahashi, 2001), whereas inactivating mutations lead to Hischsprung's disease (Manie et al., 2001). Mice that lack functional RET are embryonic lethal with loss of sympathetic and enteric neurons as well as a complete absence of kidney formation (Moore et al., 1996; Pichel et al., 1996). RET receptor activation only occurs after complexing of a GFL with a GDNF-family receptor- (GFR) receptor bound to the plasma membrane by glycosyl-phosphatidylinositol (GPI linked) (Treanor et al., 1996). Four different GFR receptors (GFR1–4) have been identified that determine the specificity of GFL binding to the RET receptor, although crosstalk of GFLs and the different GFR receptors has been reported (Creedon et al., 1997).

Ret gene expression is present in most NB tumors and GDNF induces neuronal differentiation in low-grade, but not high-grade NB cells (Nakamura et al., 1994; Hishiki et al., 1998). Interestingly, increased ret expression occurs prior to neurite out growth in NB cells induced to differentiate with retinoic acid (Bunone et al., 1995), and ret expression steadily increases upon differentiation of NB cells by NGF/TRKA signaling, suggesting that the RET signal pathway is associated with the process of NB differentiation (Matsushima and Bogenmann, 1993a). Based on these data we hypothesized that functional collaboration of RET and TRKA signal pathways is required for differentiation of malignant NB cells.

Here we demonstrate that GDNF-mediated RET activation arrests NB cells in the G₀/G₁ phase of the cell cycle and enhances CNTF-induced TRKA expression. Collaboration of GDNF with CNTF and NGF downregulates expression of N-MYC in various malignant NB cell lines and induces neuronal differentiation. Thus, trophic factors regulating normal neuronal maturation can be used to promote NB differentiation, thereby overcoming the malignant phenotype.

Material and methods

Cell culture

HTLA230 cells were isolated from a patient with metastatic stage IV NB tumor. These cells exhibit high levels of N-myc gene amplification and show growth in athymic nude mice (Matsushima and Bogenmann, 1992a; Bogenmann, 1996). GOTO cells were a gift from Matsushima (Jikei University, Japan). Lan-1 and Lan-6 were obtained from Dr Seeger (Childrens Hospital Los Angeles, Los Angeles, USA). SK-N-BE(2) were obtained from Dr Biedler (Biedler and Spengler, 1976) and IMR32 were purchased from American Type Culture Collection (Rockville, MD, USA). Cells were grown in Dulbecco's modified eagle's medium (DMEM) (Irvine Scientific, Irvine, CA, USA) supplemented with 10% fetal bovine serum, glutamine and penicillin/streptomycin (GIBCO/BRL, Grand Island, NY, USA). Although data presented here were obtained with cells grown in medium supplemented with 10% FBS, experiments performed in 1% FBS or serum-free medium yielded similar results. Recombinant human NGF was a gift from Genentech, (South San Francisco, CA, USA) and CNTF was purchased from Promega (Madison, WI, USA). GDNF was obtained from R&D Systems (Minneapolis, MN, USA), whereas NTN and ActD were obtained from PeproTech Inc. (Rocky Hill, NJ, USA) and Sigma (St Louis, MO, USA), respectively.

Cell cycle analysis

HTLA230 cells were grown for 4 days in the presence or absence of GDNF (10 ng/ml), trypsinized, washed with PBS and fixed in 70% ethanol overnight at 4°C. Cells were washed, treated with RNAse A (40 g/ml) for 30 min at 37°C and stained with propidium iodide (PI) (40 g/ml). Cells were analysed on a FACScan flow cytometer (Becton Dickinson, San Jose, CA, USA), and cell cycle distribution was determined using ModFit LT software (Verity Software House, Topsham, ME, USA).

Western blotting and immunoprecipitation (IP)

Cells were prepared for Western blot analysis and IP as previously described (Torres and Bogenmann, 1996). Briefly, cell extracts (25 g/lane for Western blots) were resolved on polyacrylamide gels (Invitrogen, San Diego, CA, USA), transferred to nitrocellulose membrane (Schleicher and Schuell, Keene, NH, USA) and incubated with appropriate antibodies. The following antibodies were employed in this study, i.e. anti-TRK (C-14), anti-N-MYC (C-19), anti-PCNA (PC10), anti-RET (C-19), and they were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). The anti-phosphotyrosine (PY) was obtained from Upstate Biotechnology Inc. (Lake Placid, NY, USA) and the anti-p75NTR was obtained from Promega. The anti--tubulin antibody was obtained from Sigma. Antibodies were employed according to the manufacturer's instructions and antibody complexes were visualized via an enhanced ECL system (Amersham Co., Arlington Heights, IL, USA), using horseradish peroxidase-conjugated goat anti-mouse or goat anti-rabbit antibodies (Kirkegaard and Perry, Inc., Gaithersburg, MD, USA) as described previously. Blots were stripped according to instructions provided by the manufacturer.

Northern blot analysis

Cells were incubated in complete medium for the indicated time, washed twice with PBS, and total cellular RNA was harvested and processed for Northern blotting as previously described (Matsushima and Bogenmann, 1992a). RNA samples (20 g/lane) were hybridized with the following DNA probes: a 0.5 kb Xba1–BamH1 fragment of the pmp34.1 N-myc clone, a 1.8 kb fragment of the pk-1 clone (human -tubulin), a 1.9 kb fragment of the scg10 gene, the human ret proto-oncogene as described in Matsushima and Bogenmann (1993a). Each probe was radiolabeled to a specific activity of greater than 10⁹ c.p.m./g of DNA, using an oligonucleotide labeling cocktail (Pharmacia LKB Biotechnology, Uppsala, Sweden).

FACS analysis

HTLA230 cells were cultured for 4 days in the presence or absence of GDNF (10 ng/ml), harvested using 5 mM EDTA in PBS and incubated with blocking solution (5% BSA, 1% goat serum in PBS) for 30 min. All incubations were done on ice. Cells were then stained for 1 h either with the secondary antibody (fluorescein isothiocyanate-(FITC)-conjugated affinity-purified goat anti-mouse antibody, ICN Biomedicals, Aurora, OH, USA, dilution of 1 : 50 in blocking solution) alone, or stained with the primary antibody (anti-human GFR1, Transduction Labs, Lexington, KY, USA) at a dilution of 1 : 100 in blocking solution. Fluorescence intensity was determined by FACS analysis.

Data analysis

Western and Northern blots were quantitated using Labworks software (UVP, Upland, CA, USA) and data was standardized based on housekeeping gene expression levels (-tubulin for Western blots, -actin or -tubulin for Northern blots). Relative expression levels obtained from different experiments are presented as the means.d. around the mean. Student's t test (two sided, unpaired, homoscedastic) was used to evaluate statistical significance, and P-values are presented either in the text or the figures.

Results

Functional activation of the RET receptor by GFLs

The biological effects of GDNF on HTLA230 cells that express low levels of the ret gene were investigated (Matsushima and Bogenmann, 1993a). Cells were grown for 2 days in either medium alone (Figure 1a) or medium supplemented with 10 ng/ml of GDNF (Figure 1b) and cultures were analysed microscopically. HTLA230 cells grown in control medium generally form big multicellular clusters of cells with large areas of space between such islands of growing cells. The cells are rather round with some exhibiting a more spread morphology (Matsushima and Bogenmann, 1992b). Cells treated with GDNF showed very subtle morphological changes. However, following GDNF treatment, cultures of HTLA230 cells lacked large multicellular clusters with cells evenly covering the entire Petri dish in a single cell layer. Interestingly, GDNF seemed to markedly reduce cell proliferation and hence, we investigated the effect of GDNF on cell cycle progression. Cells were grown for 4 days in the presence or absence of GDNF, stained with PI and cell cycle distribution was determined by FACS analysis (Figure 1c). Cells grown in control medium demonstrated the DNA profile of a fast growing cell population with 488% of the cells in the G₀/G₁ phase and 388% of the cells in the S phase of the cell cycle. In contrast, GDNF treatment increased the number of cells in the G₀/G₁ phase to 721% (P<0.02), while decreasing the number of cells in the S phase to 140.5% (P<0.02). The number of cells in the G₂/M phase was similar (130.5%) in control and GDNF-treated cells. Hence, we concluded that GDNF treatment inhibits cell proliferation in HTLA230 NB cells. To further substantiate this result, we investigated the effect of GDNF on N-MYC expression by Western blot analysis, since expression of this gene is a marker for proliferating HTLA230 cells. Cells were grown for 4 days in medium alone or medium supplemented with GDNF, and total cell lysates were analysed with an anti-N-MYC antibody (Figure 1d). A lysate from cells grown in the presence of NGF was used as an internal negative control, since HTLA230 cells are NGF unresponsive and thus no changes in N-MYC were expected. Western blots were quantiated as described in Material and methods. High levels of N-MYC proteins were present in control cells and in cells grown in the presence of NGF. In contrast, GDNF induced a greater than fourfold reduction in the abundance of the N-MYC protein, consistent with the observed cell cycle arrest. Furthermore, Western blotting with an antibody to proliferating cell nuclear antigen (PCNA), a widely used marker gene of proliferating cells, also demonstrated a greater than twofold reduction in the abundance of PCNA protein in GDNF-treated cells when compared to either control cells or cells grown in the presence of NGF. Thus, we concluded that GDNF induced cell cycle arrest is accompanied by downregulation of proliferation-associated genes.

Figure 1.

(a,b). HTLA230 cells were grown in the absence (a) or presence (b) of GDNF (10 ng/ml) for 2 days and analysed by light microscopy. Cells grown in control medium only were attached to the culture dish in large cell aggregates and showed a rather round morphology. In contrast, cells grown in the presence of GDNF had a slightly flatter morphology and were uniformly dispersed over the entire culture dish. (magnification 125). (c) Cell cycle distribution analysis. HTLA230 cells were grown for 4 days in the absence (open bars) or presence of GDNF (10 ng/ml) (closed bars), and the cell cycle distribution was determined as described in Materials and methods. Cells grown in control medium showed a DNA histogram consistent with that of an asynchronously growing cell population with a high proportion of cells in the G₀/G₁ and the S -phase of the cell cycle. In contrast, cells grown in the presence of GDNF showed a significant increase in the number of cells present in the G₀/G₁ phase of the cell cycle, concomitant with a greater than twofold decrease in the number of cells in the S phase. Each column represents the mean percentage of cells in each cell cycle phase (s.d., n=2). ^*P<0.02 comparing untreated cells with GDNF-treated cells via Student's t-test. (d) HTLA230 cells were grown in control medium or medium supplemented with either NGF (50 ng/ml) or GDNF (10 ng/ml) for 4 days and total cell lysates were analysed by Western blots using antibodies specific for N-MYC, PCNA and -tubulin. The level of N-MYC protein was high in control cells as well as in NGF-treated cells, whereas the N-MYC level was decreased in GDNF-treated cells. Similarly, the level of PCNA protein was reduced in the presence of GDNF when compared to control cells. Representative blots are shown. (e) Western blots were quantitatively analysed using Labworks software (see Materials and methods), and the data obtained demonstrated that GDNF reduced the level of N-MYC protein (open bars) by 781.8% compared to untreated cells. Similarly, PCNA levels (closed bars) were reduced by 555% in GDNF-treated cells. Each column represents the mean pecentage of N-MYC or PCNA over -tubulin (s.d., n=3). t-Test comparing N-MYC and PCNA levels in GDNF-treated cells with untreated cells ^*P<0.005

Full figure and legend (191K)

The mechanism by which GDNF exerts its effect on HTLA230 cells was investigated next. Ret expression was analysed by Northern blots in HTLA230 cells grown for 4 days in medium alone, or medium supplemented with NGF or GDNF (Figure 2a). Several ret mRNA species with approximate molecular weights of 4.0, 4.5 and 9.0 kb were present at low abundance in untreated HTLA230 cells and NGF treatment did not alter ret gene expression. In contrast, the presence of GDNF increased the steady-state level of all ret mRNA species by fourfold (Figure 2b). To further characterize the mechanism by which GDNF induces ret expression, we exposed HTLA230 cells for 16 h to control medium with or without GDNF (10 ng/ml) in the presence and absence of the RNA synthesis inhibitor actinomycin D (ActD) at 0.5 g/ml. Subsequent Northern blot analysis demonstrated that ActD treatment did not affect ret mRNA steady-state levels in control cells, while the presence of the inhibitor reduced the GDNF-mediated increase in ret mRNA expression (Figure 2c), suggesting that GDNF induces ret gene expression by a transcriptional mechanism, rather than by stabilization of the existing ret mRNA pool.

Figure 2.

(a) HTLA230 cells were grown for 4 days in medium alone or medium supplemented with either NGF (50 ng/ml) or GDNF (10 ng/ml). Total RNA was isolated and analysed by Northern blots with probes specific for ret and -tubulin genes. A low level of ret mRNA was present in cells grown in medium alone or medium supplemented with NGF, whereas GDNF induced a significant increase in the steady-state level of ret mRNA. The blot was rehybridized with a probe for -tubulin to ascertain equal RNA loading. Data from representative experiments are shown (n=3). (b) Quantitative analysis of the Northern blots demonstrated that a statistically significant (^*P<0.006) fourfold increase in the steady-state level of ret mRNA occurred in the presence of GDNF when compared to control cells. Each column represents the mean fold increase in ret/-tubulin mRNA (s.d.) (n=3). (c) HTLA230 cells were grown for 16 h in control medium alone or medium supplemented with GDNF (10 ng/ml) in the absence or presence of the RNA synthesis inhibitor ActD (0.5 g/ml). Total RNA was isolated and analysed by Northern blots with probes specific for ret and -tubulin. Quantitation of blots was done as above. The GDNF-mediated increase (3.5-fold) in steady-state levels of ret mRNA was reduced to 1.8-fold in the presence of ActD when compared to control cells, suggesting that the increase in ret expression occurs by a transcriptional mechanism

Full figure and legend (133K)

The effect of various GDNF concentrations on RET protein expression was analysed next. HTLA230 cells were grown for 4 days in increasing concentrations of GDNF (0, 1, 5, 10, 25 ng/ml) and total cell lysates were Western blotted with an anti-RET antibody (Figure 3a), and receptor expression was quantitated as above (Figure 3b). The anti-RET antibody detected two low level proteins with molecular weights of 150 and 170 kDa in cells grown in medium alone or medium supplemented with 1 ng/ml of GDNF, whereas 5 ng/ml of GDNF noticeably increased RET expression by sevenfold. Maximal expression of total RET protein was achieved by treatment with 10 ng/ml of GDNF, which resulted in an approximate 10-fold increase of RET protein, when compared to untreated cells. Importantly, every cell line tested from a panel of human NB (GOTO, Lan-1, IMR-32, Lan-6, SK-N-BE(2)) showed an increase in total RET expression when treated for 4 days with 10 ng/ml of GDNF, albeit to various degrees (Figure 3c), suggesting that this is a general response to GDNF treatment.

Figure 3.

(a) HTLA230 cells were grown for 4 days in either control medium alone or medium supplemented with increasing concentrations of GDNF. Western blot analysis of total cell lysates demonstrated a dose-dependent increase in total RET protein. Maximal levels of RET protein were seen in the presence of 10 ng/ml GDNF. (b) Quantitative analysis of Western blots demonstrated that RET expression was increased on an average of 10-fold (1.5) in the presence of 10 ng/ml of GDNF when compared to untreated cells. Columns represent the fold increase in RET/-tubulin protein in untreated cells compared to GDNF (10 ng/ml) treated cellss.d. (n=4). The GDNF-mediated increase in RET protein is statistically significant (P<0.05). (c) Various human NB cell lines were grown in the absence or presence of GDNF (10 ng/ml) for 4 days and RET expression was analysed by Western blots. Total RET protein levels were increased in all cell lines tested, albeit to various degrees, suggesting that this is a common mechanism to regulate RET expression. Data from representative experiments are shown (n=2). (d) HTLA230 cells were grown for various periods of time in either control medium or medium supplemented with GDNF (10 ng/ml). The RET protein was immunoprecipitated from total cell lysates and Western blots were analysed using an anti-RET antibody. Designated cultures were also stimulated for 5 min with GDNF (10 ng/ml) prior to harvest of cells to investigate RET receptor phosphorylation using an anti-PY antibody. The continuous presence of GDNF in the medium results in a steady increase in total RET receptor that is present as 150 and 170 kDa molecules. Low RET phosphorylation was present in control cells, while GDNF stimulation induced strong tyrosine phosphorylation of the receptor. A sustained, albeit declining, receptor phosphorylation occurred in the continuous presence of GDNF over a 72 h time period, after which RET was still amenable to strong phosphorylation by short-term GDNF treatment. Data from representative experiments are shown (n=3). (e) HTLA230 cells were grown for 4 days in the absence or presence of increasing concentrations of NTN and total cell lysates (25 g/lane) were immunoblotted with an anti-RET antibody. The continuous presence of NTN induced a dose-dependent increase in total RET protein, although higher concentrations were required when compared to that seen with GDNF. Representative blots are shown (n=2). (f) HTLA230 cells were grown in the absence or presence of GDNF (10 ng/ml) for 4 days and FACS analysis of GFR-1 surface expression was performed. Control and GDNF-treated cells when stained with the FITC-conjugated secondary antibody alone (peak I) showed a similar fluorescence histogram, whereas staining of untreated cells with both the anti-GFR-1 antibody and the secondary antibody showed specific surface staining (peak II). GDNF-treated cells showed a shift in the fluorescence histogram, consistent with increased GFR-1 surface expression (peak III). Thus, both components of the RET receptor complex are regulated by the ligand

Full figure and legend (206K)

The GDNF-mediated increase in RET protein as a function of time was analysed by IP of the receptor with an anti-RET antibody followed by Western blotting with a RET antibody (Figure 3d). Tyrosine phosphorylation of the immunoprecipitated receptor was analysed by immunoblotting Western blots with an anti-PY antibody. Growth of HTLA230 cells in the continuous presence of GDNF (10 ng/ml) steadily increased the abundance of total RET protein over a period of 3 days, thus mirroring the increase seen by Northern blots. Activation of the receptor with GDNF (10 ng/ml) for 5 min prior to harvest of the cells induced strong tyrosine phosphorylation in control cells, whereas the continuous presence of GDNF resulted in a sustained, albeit steadily declining phosphorylation of RET. To test whether long-term GDNF-treated cells express surface bound RET receptors amenable to activation by GDNF, we continuously incubated cells for 3 days with GDNF followed by a 5 min incubation with GDNF (10 ng/ml) prior to harvest, and the cells were analysed as above. Data obtained demonstrated that the GDNF-mediated increase in total RET protein leads to surface bound RET receptors, since GDNF induced strong tyrosine phosphorylation of the receptor.

NTN, another GFL, activates the RET receptor complex in the presence of GFR2, although crosstalk with other members of the GFR family has been demonstrated (Creedon et al., 1997). Hence, we investigated whether continuous treatment with NTN also leads to an increase in total RET protein. Cells were exposed to increasing concentrations of NTN for 4 days and Western blots were analysed with the anti-RET antibody (Figure 3e). Treatment with 10 ng/ml of NTN induced only a marginal (twofold) increase in total RET protein, whereas incubation with 25 ng/ml of NTN raised the level of RET protein by 12-fold when compared to untreated cells. Thus, activation of RET and its signal pathway increases RET expression, regardless of the GFL used.

Functional activation of RET requires the presence of GFR coreceptors and hence we investigated whether the continuous presence of GDNF also affects expression of GFR-1, the preferred coreceptor for GDNF (Airaksinen and Saarma, 2002). Cells were treated with or without GDNF (10 ng/ml) for 4 days and FACS analysis was performed with an anti-GFR-1 antibody (Figure 3f). Similar background staining was seen in control cells or cells treated with the GFL when stained with a FITC-conjugated secondary antibody only (peak I). Staining of untreated cells with the anti-GFR-1 antibody and the secondary antibody (peak II) demonstrated specific GFR-1 surface expression which was twofold increased (mean fluorescence intensity) in GDNF-treated cells (peak III) as indicated by a shift in fluorescence intensity. Taken together, activation of the RET receptor signal pathway increases the expression of both components of the RET receptor complex.

Cooperation of the RET and TRKA signal pathways

We have previously demonstrated that CNTF, a cytokine involved in maturation of neuronal precursor cells (Ip and Yancopoulos, 1996), renders HTLA230 cells NGF responsive by inducing transcription of the silenced trkA gene (Bogenmann et al., 1998). Yet, only a small fraction of CNTF-treated NB cells differentiated in the presence of NGF, suggesting the requirement for additional trophic factors. We hypothesized that GDNF collaborates with CNTF to enhance the NGF response by increasing the expression of TRKA receptors. Thus, HTLA230 cells were treated for various time intervals with CNTF or GDNF or a combination of the two factors and the TRK receptor was immunoprecipitated from equal amounts of total cell protein and analysed by Western blotting (Figure 4). The TRK protein was essentially undetectable in untreated HTLA230 cells, while CNTF induced a 14-fold increase in the level of receptor protein, consistent with our previous data (Bogenmann et al., 1998). In contrast, GDNF treatment increased TRK expression only by 5 fold, whereas treatment with both the trophic factors enhanced the total level of TRK protein by 35-fold. Subsequent experiments demonstrated that the TRK protein upregulated by CNTF and GDNF could be tyrosine phosphorylated by NGF, but not by BDNF or NT3, consistent with the presence of TRKA (data not shown). Thus, CNTF and GDNF synergize to increase TRKA expression and ultimately the NGF-mediated signal.

Figure 4.

HTLA230 cells were grown for 4 days in control medium alone, or medium supplemented with either CNTF (25 ng/ml) or GDNF (10 ng/ml) or a combination of both factors. The TRKA protein was immunoprecipitated from equal amounts of total cell lysates and Western blotted using an anti-TRK antibody. Quantitative analysis was performed as above. TRKA was essentially undetectable in untreated cells, whereas CNTF induced a 14-fold increase in the level of TRKA protein, while the presence of GDNF resulted in an approximate fivefold stimulation of TRKA expression. However, growth of cells in the presence of both CNTF and GDNF increased the level of TRKA by more than 30-fold, suggesting a synergistic effect of CNTF and GDNF on TRKA expression. Data represent the analysis of four independent experiments and a representative Western blot is shown. ^*P<0.005 comparing untreated cells with CNTF-treated cells. ^**P<0.02 comparing CNTF-treated cells with CNTF- and GDNF-treated cells

Full figure and legend (84K)

We next investigated whether addition of NGF to cultures treated with GDNF and CNTF would morphologically differentiate the majority of HTLA230 cells as predicted by our hypothesis. Thus, cells were grown for 6 days in either medium alone (Figure 5a) or medium supplemented with GDNF (10 ng/ml), CNTF (25 ng/ml) and NGF (50 ng/ml) and cultures were analysed by light microscopy (Figure 5b). Cells grown in medium alone showed a round and undifferentiated morphology typical of proliferating NB cells, whereas cells treated with the three growth factors underwent dramatic morphological changes. An extensive network of thick bundles with an 'axon-like' morphology formed within 2–4 days of treatment, similar to that seen with trkA-transfected NB cells grown in the presence of NGF (Matsushima and Bogenmann, 1993a). The axons and small neurites also contained many tiny bifurcations, thereby establishing a fine network of cell-to-cell connections, typical of normal differentiated neurons.

Figure 5.

HTLA230 cells were grown for 6 days in either growth medium alone (a) or medium supplemented with GDNF (10 ng/ml), CNTF (25 ng/ml) and NGF (50 ng/ml) (b) and cultures were analysed microscopically. Cells grown in control medium were generally organized within large cellular aggregates and showed a rather round morphology. In contrast, growth factor-treated cultures were composed of an extensive intertwined network of axons and neurites derived from differentiated NB cells (magnification 125)

Full figure and legend (496K)

Neuronal differentiation is accompanied by growth cessation, downregulation of cell cycle control genes and expression of neuron-specific genes. To demonstrate that the observed morphorphological differentiation is consistent with neuronal differentiation, we cultured HTLA230 cells in either growth medium alone or medium supplemented with the indicated growth factors, cell numbers were determined from triplicate cultures at the designated time intervals and a Student's t-test analysis was performed for day 7 (Figure 6a). Cell doubling time for each culture condition was determined between day 5 and day 7 (Figure 6b). Cells grown in medium alone or medium supplemented with CNTF and NGF showed similar growth kinetics and the performed t-test confirmed that these were not statistically different. Indeed, control cells showed an approximate doubling time of 41 h similar to that determined for cells grown in the presence of CNTF and NGF (36 h) (Figure 6b). However, consistent with the observed cell cycle arrest presented in Figure 1c, cells grown in the presence of GDNF had a significantly reduced final cell density (P=0.009), concomitant with an increased doubling time of 58 h. The presence of all three growth factors further reduced the final cell density (P=0.0004) and proliferation was substantially suppressed between days 5 and 7, thereby increasing the doubling time to approximately 108 h. Thus, morphological neuronal differentiation was concurrent with growth cessation.

Figure 6.

(a) HTLA230 cells were grown in medium alone or medium supplemented with the indicated growth factors and cell numbers were determined from triplicate cultures at the indicated time points. Cells grown in medium alone or medium supplemented with both CNTF (25 ng/ml) and NGF (50 ng/ml) showed a similar rapid growth rate, whereas the growth of cells in the presence of GDNF (10 ng/ml) was significantly (P<0.009) reduced, consistent with the cell cycle data (see Figure 1c). Treatment with a combination of GDNF, CNTF and NGF reduced cell proliferation even more dramatically (P<0.0004). (b) The population doubling times for cells grown under the above conditions were determined for the time period between days 5 and 7 of the treatment. Untreated cells as well as cells grown in the presence of NGF and CNTF had similar population doubling times of approximately 40 h, whereas the continuous presence of GDNF prolonged the doubling time to 58 (8) h. However, the presence of the three trophic factors increased the doubling time by approximately 2.5-fold, when compared to untreated cells, consistent with the observed growth cessation

Full figure and legend (122K)

HTLA230 cells transfected with a trkA cDNA express several neuronal markers (i.e., SCG10, NF-68, GAP43) when differentiated with NGF and exhibit downregulation of the N-myc gene (Matsushima and Bogenmann, 1993b). Thus, we investigated whether cells differentiated with GDNF, CNTF and NGF show a similar biological phenotype and we chose scg10, a protein present in the mature neurite, as a suitable marker (Stein et al., 1988). The expression of scg10, N-myc and -actin was determined by Northern blots in cells exposed for various time periods to the growth factors (Figure 7a). The scg10 mRNA is essentially undetectable in control cells, yet its expression is readily detectable within 12 h of growth factor treatment, and a steady increase in its abundance occurs resulting in approximately a 20-fold increase after 4 days when compared to untreated cells. This increase in scg10 expression is concomitant with the establishment of an abundant network of axons and neurites (see Figure 5b). In contrast, the N-myc gene is abundantly expressed in parental HTLA230 (Matsushima and Bogenmann, 1992b), but a greater than 50-fold decrease in the steady-state level of N-myc expression occurred as a result of a 6-day treatment, mirroring the observed growth arrest. These data were further substantiated by Western blotting of cell lysates isolated at the same experimental time points with antibodies to RET, p75NTR, N-MYC and PCNA (Figure 7b). Neuronal differentiation was accompanied by an increase in the abundance of the RET and p75NTR proteins. Quantitation of these expression data demonstrated a greater than 50-fold increase in total RET protein, whereas p75NTR expression was augmented by 20-fold (Figure 7c). In contrast, expression of N-MYC and PCNA were downregulated by growth factor treatment, and quantitive data demonstrated a greater than 12-fold decrease for total N-MYC protein, whereas total PCNA protein declined approximately fourfold (Figure 7c). This biological response was not unique to HTLA230 NB cells. Investigation of N-MYC expression in other cell lines with high N-myc gene amplification (i.e., GOTO, IMR32) demonstrated that the abundance of this marker of proliferating NB cells was 8–10-fold reduced following 6 days of treatment with GDNF, CNTF and NGF (Figure 7d), suggesting that the RET and TRKA signal pathways collaborate in various highly tumorigenic NB cells to modulate the expression of proliferation regulating genes.

Figure 7.

(a) HTLA230 cells were grown for the indicated time intervals in the presence of GDNF, CNTF and NGF and total RNA was isolated to analyse expression of the scg10 and N-myc genes. Expression of the scg10 gene was essentially undetectable in untreated cells, while a steady increase in scg10 mRNA occurred in the presence of growth factors, mirroring the outgrowth of axons and neurites. In contrast, N-myc expression was high in untreated cells, while growth factor treatment induced a continuous decline in steady-state levels of the N-myc mRNA pool, consistent with the observed cessation of cell proliferation. The fold increase of scg10 or decrease of N-myc was determined and standardized based on the level of -actin mRNA. (b) HTLA230 cells were grown in the presence of the growth factors for the indicated time intervals and total cell lysates were prepared for Western blotting with antibodies specific for RET, p75NTR, N-MYC, PCNA and -tubulin. Expression of RET and p75NTR receptors was low in untreated cells, but steadily increased in the presence of growth factors. In contrast, expression of N-MYC and PCNA was high in untreated cells but decreased during differentiation. (c) Quantitation of protein expression after 6 days of growth factor treatment. The data is expressed as fold increase or decrease in treated compared to untreated cells. Each column represents the mean fold increase or decrease for the indicated protein/-tubulin (s.d.) (n=3). (d). GOTO and IMR32 NB cells were grown for 6 days in either medium alone or medium supplemented with GDNF (10 ng/ml), CNTF (25 ng/ml) and NGF (50 ng/ml) and total cell lysates were Western blotted with an anti-N-MYC and an anti--tubulin antibody. High levels of N-MYC protein were present in untreated cells, yet growth factor treatment significantly reduced N-MYC expression levels in both cell lines. Quantitative analysis was performed as before (n=2)

Full figure and legend (208K)

Differentiation of normal dorsal root ganglion cells with NGF results in mature nerve cells dependent on trophic factors (Crowley et al., 1994; Smeyne et al., 1994) and we wondered whether differentiated HTLA230 would demonstrate a similar phenotype. Hence, HTLA230 cells were differentiated with GDNF, CNTF and NGF and the medium was replaced with medium from which growth factors were omitted. Morphological analysis demonstrated that the intertwined network of differentiated cells detached within 2 days upon removal of the growth factors from the growth medium, suggesting that maintenance of differentiated NB cells in vitro required the presence of trophic factors. The remaining attached cells were predominantly single cells with a rather flat morphology. A time-dependent analysis of marker gene expression (i.e., RET, TRKA, p75NTR, N-MYC) in attached cells demonstrated a loss of high level RET, p75NTR and TRKA expression, whereas levels of N-MYC protein rapidly increased within 48 h following withdrawal of the three trophic factors (Figure 8, lanes 2–5). These data together with the morphological analysis suggested that differentiated NB cells are lost from the cultures following withdrawal of trophic support, whereas residual attached cells assumed the phenotype of proliferating NB cells. To investigate which of the growth factor(s) is required for maintenance of differentiated NB cells, we supplemented the medium with individual growth factors or combinations of them and determined that the presence of CNTF and NGF was sufficient to sustain cultures of differentiated NB cells. Indeed, under these conditions, differentiated cells maintained high levels of RET, p75NTR and TRKA expression, while the level of N-MYC protein was low (Figure 8, lane 6). Interestingly, the level of TRKA protein was three times lower than that in the presence of all three trophic factors, yet it was 30 times higher than in undifferentiated cells. These data are also consistent with our model whereby GDNF increases CNTF-mediated TRKA expression. Thus, high level TRKA expression, mediated in part by GDNF, is required to promote differentiation of immature progenitors, but is no longer necessary for maintenance of the differentiated state.

Figure 8.

HTLA230 cells were grown for 7 days in medium alone (lane 1) or medium supplemented with GDNF, CNTF and NGF (lane 2). Similarly, cells were grown in the presence of the growth factors for 6 days followed by growth for 12 h (lane 3), 24 h (lane 4) or 48 h (lane 5) in medium from which the three trophic factors were omitted. In addition, medium from differentiated cultures was replaced for 48 h with medium containing only CNTF and NGF (lane 6) and total cell lysates prepared from attached cells were Western blotted with antibodies specific for RET, p75NTR, TRKA, N-MYC and -tubulin. High level RET and p75NTR expression in differentiated cells rapidly declined upon growth factor withdrawal (lanes 3–5), but was maintained in the presence of only CNTF and NGF (lane 6). Similarly, TRKA expression also declined within 48 h upon removal of trophic factors, yet its expression was maintained, albeit at a lower level, in medium supplemented with CNTF and NGF. In contrast, the low level of N-MYC expression in differentiated cells was quickly reversed upon withdrawal, but the presence of CNTF and NGF not only maintained differentiated cells, but also sustained low N-MYC expression. Thus, differentiated cells can be maintained by functional NGF/TRKA signaling

Full figure and legend (96K)

Discussion

Human NB cells have long been utilized as a model to study tumor cell differentiation, and pharmacological concentrations of retinoic acid and other agents induce differentiation in vitro and in vivo, although the stability of the achieved state of differentiation remains a matter of debate (Thiele et al., 1985; Matsushima and Bogenmann, 1992a). Here, we demonstrate that the malignant phenotype of aggressive NB cells carrying a highly amplified N-myc gene is reversed by collaborative signals from the endogenous RET and TRKA pathways, thereby overriding the consequences of genetic alterations associated with NB oncogenesis.

The ret gene was originally identified as an oncogene and constitutively active RET has since been associated with many human cancers derived from the neural crest, although the mechanism by which RET activation transforms cells remains elusive. Intriguingly, NB is of neural crest origin, yet no ret mutations seem to be associated with NB oncogenesis (Santoro et al., 1990; Goodfellow, 1994; Hofstra et al., 1996). In fact, ret expression occurs in most NB tumors independent of their clinical staging; however, GFL-mediated morphological differentiation is substantially attenuated in tumor cells derived from advanced-stage NB (Hishiki et al., 1998). Activation of RET induces cytoskeletal reorganization mediated by the activities of c-Src and the Rho GTPases (Barone et al., 2001; Encinas et al., 2001; Fukuda et al., 2002) and interestingly, loss of FYN signaling, a member of the Src kinase family, correlates with a more aggressive phenotype of NB (Berwanger et al., 2002). HTLA230 cells, isolated from a bone marrow metastasis, showed only subtle morphological changes in the presence GDNF or NTN, when compared to control cells. Yet, GFL treatment induced scattering of cells from large cell aggregates, resulting in cultures with evenly dispersed cells, perhaps reminiscent of the effect of GDNF on cell migration (Barnett et al., 2002; Natarajan et al., 2002). In addition, GDNF-treated cells were arrested in the G₀/G₁ phase of the cell cycle and substantially downregulated N-MYC expression. These data are entirely consistent with the previously observed GDNF-mediated increase in expression of the CDK inhibitor p27^kip1, a negative regulator of cell cycle progression (Baldassarre et al., 2002), perhaps suggesting that activation of the RET pathway induces exit from cell cycle, thereby enabling trophic signals to promote neuronal differentiation. Consistent with this idea, GFLs are members of the TGF- superfamily and thus may share some growth regulatory functions with TGF-. Indeed, functional cooperation of GDNF and TGF- signaling has been demonstrated (Krieglstein et al., 1998).

Interestingly, exposure of HTLA230 cells to either GDNF or NTN dose dependently increased levels of ret mRNA and RET protein. The increase in ret gene expression was sensitive to ActD treatment, thus required active transcription, rather than stabilization of the message, consistent with the previous data (Bunone et al., 1995). Importantly, the GDNF-mediated increase in total RET protein was not unique to HTLA230 cells as five other tested NB cell lines showed a similar biological response, perhaps suggesting that RET activation creates a feed back response that enhances the GFL-mediated signal. However, RET receptor activation by GFLs generally requires GFR coreceptors and several members of the GFR family can be expressed in NB tumors (Hishiki et al., 1998). HTLA230 cells express high levels of GFR-1 by Western blot analysis (data not shown) and FACS data demonstrated an increase in surface expression of GFR-1 coreceptor in the presence of GDNF. Thus, expression of both components of the RET receptor complex are upregulated by GDNF. Such a feedback mechanism has been observed in GDNF and NTN-null mutant mice, which showed decreased expression of GFR-1 and GFR-2, respectively (Airaksinen and Saarma, 2002). A ligand-mediated increase in RET expression that strengthens the trophic signal of the ligand is consistent with the observed prolonged RET receptor activation in the continuous presence of GDNF. This sustained RET signal thus may facilitate the arrest in the G₀/G₁ phase of the cell cycle in asynchronously growing cells until a differentiation signal is received (Bunone et al., 1995).

Expression of the endogenous trkA gene is lost in highly malignant NB tumors, thus NB progression is characterized by retention of the GFL/RET signal pathway controlling cell cycle progression and loss of the NGF/TRKA pathway that promotes cell differentiation. Activation of a genetically reconstituted NGF/TRKA pathway restores NGF responsiveness and reverses the malignant phenotype, while CNTF-induced expression of the endogenous trkA gene enables only a fraction of HTLA230 cells to differentiate in the presence of NGF (Bogenmann et al., 1998), indicating the requirement for additional trophic factors. Our data now demonstrate that CNTF-mediated TRKA expression is synergistically enhanced by the presence of GDNF, a novel mechanism by which to strengthen the NGF differentiation signal. Intriguingly, crosstalk between the RET and TRKA signal pathways exists in developing sympathetic neurons, whereby increased ret expression and RET phosphorylation takes place as a consequence of TRKA activation by NGF (Tsui-Pierchala et al., 2002). Taken together, our observations with malignant NB cells suggest that RET receptor activation inhibits cell cycle progression and enhances responsiveness to NGF; thus, NB cell differentiation requires the collaboration of functional RET and TRKA signal pathways.

Differentiated HTLA230 cells exhibited multiple biological properties characteristic of mature neurons. First and foremost, cultures of differentiated NB cells essentially ceased to proliferate, consistent with a substantially increased doubling time. Growth cessation was also accompanied by a near complete downregulation of the N-myc gene, which was not unique to HTLA230 cells, since IMR32 and GOTO cells, both of which express high levels of N-MYC, showed a similar biological response to growth factor treatment. Furthermore, HTLA230 cells that were differentiated by the signals of the endogenous TRKA pathway expressed high levels of SCG10, a protein only present in mature neurites. This high level of SCG10 expression in addition to other markers such as NF-68, GAP43 and peripherin was also seen in HTLA230 cells that were differentiated upon activation of a genetically reconstructed NGF/TRKA pathway (Matsushima and Bogenmann, 1993a). Thus, growth factor-induced NB differentiation recapitulates the differentiation pathway of normal neuronal precursors. Lastly, maintenance of differentiated NB cells in culture was dependent on functional NGF/TRKA signaling, similar to that seen with normal neurons (Smeyne et al., 1994). Withdrawal of the three growth factors rapidly resulted in loss of the intricate network of neurons, which could be maintained in the presence of CNTF and NGF, while GDNF was no longer required. These data suggest that GDNF is necessary for promoting differentiation of immature progenitor cells by inducing growth arrest and enhancing CNTF-mediated TRKA expression, while differentiated NB cells no longer require high levels TRKA. Indeed, differentiated cells maintained in the presence of CNTF and NGF alone had lower levels of TRKA (Figure 8, lane 6) than cells grown in the presence of all three trophic factors, reminiscent of the decrease in TRKA expression that occurs in various cells of the nervous system following maturation (Molliver and Snider, 1997).

Treatment of advanced-stage NB involves chemotherapy together with radiotherapy and more recently bone marrow transplantation followed by treatment with retinoic acid (Matthay et al., 1999). Despite these intense treatment modalities, survival of patients with recurrent NB remains dismal. Relapse after complete remission may reflect the ability of tumor cells to hide in sanctuaries (micrometastases) and elimination of these hidden tumor cells represents a major challenge of tumor therapy. An NGF-based treatment modality has previously been attempted with little success (Kumar et al., 1970); however, our data now demonstrate that a network of trophic signals is needed to achieve NB differentiation. Indeed, preliminary experiments with HTLA230 cells grown in nude mice demonstrated that NB tumors responded to growth factor treatment, since N-MYC expression was substantially decreased after treatment with GDNF, CNTF and NGF (data not shown), although the effect on micrometastases remains to be determined.

References

1. Airaksinen MS and Saarma M. (2002). Nat. Rev. Neurosci., 3, 383–394. | Article | PubMed | ISI | ChemPort |
2. Anderson DJ, Groves A, Lo L, Ma Q, Rao M, Shah NM and Sommer L. (1997). Cold Spring Harb. Symp. Quant. Biol., 62, 493–504. | PubMed | ISI | ChemPort |
3. Baldassarre G, Bruni P, Boccia A, Salvatore G, Melillo RM, Motti ML, Napolitano M, Belletti B, Fusco A, Santoro M and Viglietto G. (2002). Oncogene, 21, 1739–1749. | Article |
4. Barnett MW, Fisher CE, Perona-Wright G and Davies JA. (2002). J. Cell Sci., 115, 4495–4503. | PubMed |
5. Barone MV, Sepe L, Melillo RM, Mineo A, Santelli G, Monaco C, Castellone MD, Tramontano D, Fusco A and Santoro M. (2001). Oncogene, 20, 6973–6982. | Article | PubMed |
6. Barrett GL. (2000). Prog. Neurobiol., 61, 205–229. | Article | PubMed | ChemPort |
7. Beattie MS, Harrington AW, Lee R, Kim JY, Boyce SL, Longo FM, Bresnahan JC, Hempstead BL and Yoon SO. (2002). Neuron, 36, 375–386. | Article | PubMed | ISI | ChemPort |
8. Berwanger B, Hartmann O, Bergmann E, Bernard S, Nielsen D, Krause M, Kartal A, Flynn D, Wiedemeyer R, Schwab M, Schafer H, Christiansen H and Eilers M. (2002). Cancer Cell, 2, 377–386. | Article | PubMed | ISI | ChemPort |
9. Biedler JL and Spengler BA. (1976). Science, 191, 185–187. | PubMed | ISI | ChemPort |
10. Bogenmann E. (1996). Int. J. Cancer, 67, 379–385. | PubMed |
11. Bogenmann E, Peterson S, Maekawa K and Matsushima H. (1998). Oncogene, 17, 2367–2376. | Article | PubMed |
12. Bothwell M. (1995). Annu. Rev. Neruosci., 18, 223–253.
13. Brodeur GM. (2003). Nat. Rev. Cancer, 3, 203–216. | Article | PubMed | ISI | ChemPort |
14. Bunone G, Borrello MG, Picetti R, Bongarzone I, Peverali FA, de Franciscis V, Della Valle G and Pierotti MA. (1995). Exp. Cell Res., 217, 92–99. | PubMed |
15. Cosgaya JM, Chan JR and Shooter EM. (2002). Science, 298, 1245–1248. | Article | PubMed | ISI | ChemPort |
16. Creedon DJ, Tansey MG, Baloh RH, Osborne PA, Lampe PA, Fahrner TJ, Heuckeroth RO, Milbrandt J and Johnson Jr EM. (1997). Proc. Natl. Acad. Sci. USA, 94, 7018–7023. | Article | PubMed | ChemPort |
17. Crowley C, Spencer SD, Nishimura MC, Chen KS, Pitts Meek S, Armanini MP, Ling LH, MacMahon SB, Shelton DL, Levinson AD and X. (1994). Cell, 76, 1001–1011. | Article | PubMed | ISI | ChemPort |
18. Dechant G and Barde YA. (2002). Nat. Neurosci., 5, 1131–1136. | Article | PubMed | ISI | ChemPort |
19. Encinas M, Tansey MG, Tsui-Pierchala BA, Comella JX, Milbrandt J and Johnson Jr EM. (2001). J. Neurosci., 21, 1464–1472. | PubMed |
20. Fukuda T, Kiuchi K and Takahashi M. (2002). J. Biol. Chem., 277, 19114–19121. | Article | PubMed |
21. Goodfellow PJ. (1994). Curr. Opin. Genet. Dev., 4, 446–452. | PubMed |
22. Hempstead BL. (2002). Curr. Opin. Neurobiol., 12, 260–267. | Article | PubMed | ISI | ChemPort |
23. Hishiki T, Nimura Y, Isogai E, Kondo K, Ichimiya S, Nakamura Y, Ozaki T, Sakiyama S, Hirose M, Seki N, Takahashi H, Ohnuma N, Tanabe M and Nakagawara A. (1998). Cancer Res., 58, 2158–2165. | PubMed |
24. Hofstra RM, Cheng NC, Hansen C, Stulp RP, Stelwagen T, Clausen N, Tommerup N, Caron H, Westerveld A, Versteeg R and Buys CH. (1996). Hum. Genet, 97, 362–364. | PubMed |
25. Huang EJ and Reichardt LF. (2001). Annu. Rev. Neurosci., 24, 677–736. | Article | PubMed | ISI | ChemPort |
26. Ip NY and Yancopoulos GD. (1996). Annu. Rev. Neurosci., 19, 491–515. | Article | PubMed | ISI | ChemPort |
27. Krieglstein K, Henheik P, Farkas L, Jaszai J, Galter D, Krohn K and Unsicker K. (1998). J. Neurosci., 18, 9822–9834. | PubMed | ISI | ChemPort |
28. Kumar S, Steward JK, Waghe M, Pearson D, Edwards DC, Fenton EL and Griffith AH. (1970). J. Pediatr. Surg., 5, 18–22. | PubMed |
29. Lee FS, Kim AH, Khursigara G and Chao MV. (2001). Curr. Opin. Neurobiol., 11, 281–286. | Article | PubMed | ISI | ChemPort |
30. Manie S, Santoro M, Fusco A and Billaud M. (2001). Trends Genet., 17, 580–589. | Article | PubMed | ISI | ChemPort |
31. Matsushima H and Bogenmann E. (1992a). Int. J. Cancer, 51, 250–258.
32. Matsushima H and Bogenmann E. (1992b). Int. J. Cancer, 51, 727–732.
33. Matsushima H and Bogenmann E. (1993a). Mol. Cell. Biol., 13, 7447–7456. | PubMed | ISI | ChemPort |
34. Matsushima H and Bogenmann E. (1993b). Advances in Neuroblastoma Research 4 Evans A (ed.). John Wiley and Sons, Inc.: New York.
35. Matthay KK, Villablanca JG, Seeger RC, Stram DO, Harris RE, Ramsay NK, Swift P, Shimada H, Black CT, Brodeur GM, Gerbing RB and Reynolds CP. (1999). N. Engl. J. Med., 341, 1165–1173. | Article | PubMed | ISI | ChemPort |
36. Miyake M, Suzuki T, Shimada H, Stram D and Seeger RC. (1994). Prog. Clin. Biol. Res., 385, 163–168. | PubMed |
37. Molliver DC and Snider WD. (1997). J. Comp. Neurol., 381, 428–438. | Article | PubMed | ISI | ChemPort |
38. Moore MW, Klein RD, Farinas I, Sauer H, Armanini M, Phillips H, Reichardt LF, Ryan AM, Carver Moore K and Rosenthal A. (1996). Nature, 382, 76–79. | Article | PubMed | ISI | ChemPort |
39. Nakagawara A, Arima Nakagawara M, Scavarda NJ, Azar CG, Cantor AB and Brodeur GM. (1993). N. Engl. J. Med., 328, 847–854. | Article | PubMed | ISI | ChemPort |
40. Nakamura T, Ishizaka Y, Nagao M, Hara M and Ishikawa T. (1994). J. Pathol., 172, 255–260. | Article | PubMed | ISI | ChemPort |
41. Natarajan D, Marcos-Gutierrez C, Pachnis V and de Graaff E. (2002). Development, 129, 5151–5160. | PubMed | ISI | ChemPort |
42. Pichel JG, Shen L, Sheng HZ, Granholm AC, Drago J, Grinberg A, Lee EJ, Huang SP, Saarma M, Hoffer BJ, Sariola H and Westphal H. (1996). Nature, 382, 73–76. | Article | PubMed | ISI | ChemPort |
43. Roux PP and Barker PA. (2002). Prog. Neurobiol., 67, 203–233. | Article | PubMed | ISI | ChemPort |
44. Santoro M, Rosati R, Grieco M, Berlingieri MT, D'Amato GL, De Franciscis V and Fusco A. (1990). Oncogene, 5, 1595–1598. | PubMed | ISI | ChemPort |
45. Seeger RC and Reynolds CP. (1993). Cancer Med. Holland JF, Frei III E, Bast RC, Kufe DW, Morton DL and Weichselbaum RR (eds). Lea & Febiger: Philadelphia, pp 2172–2184.
46. Shimada H, Chatten J, Newton Jr W, Sachs N, Hamoudi AB, Chiba T, Marsden HB and Misugi K. (1984). J. Natl. Cancer Inst., 73, 405–416. | PubMed | ChemPort |
47. Sleeman MW, Anderson KD, Lambert PD, Yancopoulos GD and Wiegand SJ. (2000). Pharm. Acta Helv., 74, 265–272. | Article | PubMed | ChemPort |
48. Smeyne RJ, Klein R, Schnapp A, Long LK, Bryant S, Lewin A, Lira SA and Barbacid M. (1994). Nature, 368, 246–249. | Article | PubMed | ISI | ChemPort |
49. Stein R, Orit S and Anderson DJ. (1988). Dev. Biol., 127, 316–325. | PubMed | ISI | ChemPort |
50. Takahashi M. (2001). Cytokine Growth Factor Rev., 12, 361–373. | Article | PubMed | ISI | ChemPort |
51. Thiele CJ, Reynolds CP and Israel MA. (1985). Nature, 313, 404–406. | Article | PubMed | ISI | ChemPort |
52. Torres M and Bogenmann E. (1996). Oncogene, 12, 77–86. | PubMed |
53. Treanor JJ, Goodman L, de Sauvage F, Stone DM, Poulsen KT, Beck CD, Gray C, Armanini MP, Pollock RA, Hefti F, Phillips HS, Goddard A, Moore MW, Buj Bello A, Davies AM, Asai N, Takahashi M, Vandlen R, Henderson CE and Rosenthal A. (1996). Nature, 382, 80–83. | Article | PubMed | ISI | ChemPort |
54. Tsui-Pierchala BA, Milbrandt J and Johnson Jr EM. (2002). Neuron, 33, 261–273. | Article | PubMed | ISI | ChemPort |
55. Wang KC, Kim JA, Sivasankaran R, Segal R and He Z. (2002). Nature, 420, 74–78. | Article | PubMed | ISI | ChemPort |
56. Westermann F and Schwab M. (2002). Cancer Lett., 184, 127–147. | Article | PubMed | ISI | ChemPort |

Acknowledgements

This work was supported by a grant from the NIH (RO1 NS36978).