INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Neuroblastoma is a predominantly pediatric neoplasia with nearly all cases occurring in children younger than 10 years of age (1). It is the major form of extracranial solid tumor in children and accounts for 10-15% of childhood cancer-related deaths (2). Based on the early onset of this disease and the expression of specific cellular and developmental markers, neuroblastoma appears to arise from sympathoadrenal precursors that fail to undergo terminal differentiation and/or cell death during fetal development (2-4). Significant advances have been made in prognostic markers for this disease (5-8). In particular, diploid DNA content, N-MYC amplification and chromosome 1p deletion are highly associated with rapid tumor progression and poor outcome. Conversely, near triploidy and absence of 1p deletions are linked to a favorable outcome in infants and greater responsiveness to chemotherapy (2, 9). However, the outcome for neuroblastoma patients remains generally poor despite multitherapy strategies (2, 10). An interesting and characteristic feature of neuroblastoma is the occurrence of spontaneous regression or differentiation into benign ganglioneuromas in a minority of patients, independent of treatment (2, 11). A critical question is what regulatory mechanisms characterize these spontaneously regressing neuroblastomas and distinguish them from the more aggressive, unfavorable forms?

TrkA is a member of the neurotrophin tyrosine kinase receptor family that also includes TrkB and TrkC. TrkA specifically mediates signaling for nerve growth factor (NGF)¹ (although it can also be activated by NT-3) (12) and is critical for both survival and terminal differentiation of sympathetic and a subset of sensory neurons (12-14). Within the central nervous system, TrkA signaling is important for basal forebrain cholinergic neurons, among other functions (14). Recent findings have demonstrated that abundant expression of TrkA is strongly correlated with favorable prognosis for neuroblastoma, while low or absent expression is linked to a poor outcome (15-19). TrkA expression within neuroblastomas occurs specifically in neuroblasts and differentiated ganglion cells, with the highest levels occurring in the latter cells (20). In addition, forced expression of TrkA in neuroblastoma cells lacking this receptor converts them into NGF-responsive cells that undergo growth arrest and terminal differentiation in the presence of NGF (17, 21-23). These findings have led to the suggestion that TrkA expression in neuroblastoma tumors actively promotes their growth arrest and differentiation into a regressed or benign state (16, 24). Alternatively, TrkA expression may be associated with a generally more differentiated state in neuroblastoma cells that are predisposed to growth arrest and further differentiation.

Whether an active participant in neuroblastoma tumor arrest or simply a marker for a more differentiated state, elucidating the mechanisms responsible for TrkA expression (and its absence) in neuroblastoma cells is potentially of great importance for understanding and treating this disease. In particular, it may ultimately reveal regulatory mechanisms that distinguish favorable from poor prognosis tumors that could serve as future therapeutic targets. The existence of human neuroblastoma cell lines that express TrkA at low or moderate to high levels provides an opportunity to explore the determinants of its variable expression in this tumor cell type. To this end, we have examined the role of gene transcription in the differential expression of TrkA in various human neuroblastoma cell lines and have characterized the human TrkA promoter in these cells using transient transfection. Our findings indicate that proximal regulatory sequences play a critical role in both cell-specific and variable TrkA promoter activity in neuroblastoma cells.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cell Culture-- The human neuroblastoma cell lines SH-SY5Y, IMR32, LA-N-6, SK-N-SH, SMS-KCN (25-29) and kidney K293 cells were cultured in complete RPMI 1640 medium (Life Technologies, Inc.) supplemented with 10% heat-inactivated fetal bovine serum (Sigma), 50 units/ml sodium penicillin, and 50 units/ml streptomycin sulfate at 37 °C in humidified 5% CO₂.

Preparation of Total RNA and Northern Blotting-- Total RNA was extracted from fresh or frozen cell pellets using the guanidinium isothiocyanate/CsCl₂ method (30). Twenty µg of total RNA were separated on formaldehyde gels and electrophoretically transferred to GeneScreen Plus membranes (NEN Life Science Products). Membranes were hybridized with a 1.5-kb BamHI-KpnI fragment from the human TrkA cDNA pLM6 (31) that was labeled using random primers. Variation in the loading of total RNA was normalized by hybridization with a 1.2-kb PstI fragment derived from a human glyceraldehyde-3-phosphate dehydrogenase cDNA (pHcGAP) (32).

Isolation of Nuclei and Nuclear Run-on Analysis-- Cell nuclei were isolated using a modification of the method of Greenberg and Ziff (33). Approximately 2-3 × 10⁷ cells were lysed and homogenized in hypotonic buffer (10 mM Tris-HCl, pH 7.4, 70 mM NaCl, 3 mM MgCl₂, 1% Nonidet P-40, 0.5 mM phenylmethylsulfonyl fluoride, and 3 mM dithiothreitol) at 4 °C. The nuclei were pelleted and resuspended in 200 µl of storage buffer (25% glycerol, 50 mM Tris-HCl, pH 7.4, 5 mM MgCl₂, 1 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, 3 mM dithiothreitol, and 1 µg/ml each of the proteinase inhibitors pepstatin A, bestatin, aprotinin, and leupeptin) at a concentration of 5-7.5 × 10⁷ nuclei/ml.

Isolation of Human TrkA Genomic Sequences-- A 3.1-kb HindIII fragment was isolated from an EMBL-3 genomic library derived from human placental DNA using a 236-bp EcoRI-SmaI fragment of the human TrkA cDNA as well as an oligonucleotide complementary to sequences 22-48 of the TrkA cDNA (5-TCTGTGCGCTCCCAGCTGCAGCTGCCA-3) as probes (31). This genomic sequence contained the 5-flanking region as well as exon I and a portion of the first intron of the human TrkA gene based on dideoxy sequencing and comparison with the human TrkA cDNA sequence (31). It was cloned into the HindIII site of pBluescript (KS) to generate pJD1.

Primer Extension and RNase Protection Assays-- Primer extension was performed essentially as described elsewhere (37). An end-labeled primer (5-CGATGTAGAGCTCAGTCAGGTTCTCTGCGCCGGGCAGGTGGTGGAGG-3 corresponding to sequences +174 to +221 relative to the main translation start codon for human TrkA) (31) was hybridized to total RNA in hybridization buffer (150 mM KCl, 10 mM Tris-HCl, pH 8.3, and 1 mM EDTA) at 65 °C for 4 h. Reverse transcription was carried out with SuperScript II RNase H-reverse transcriptase (Life Technologies, Inc.) at 55 °C for 40 min. Reactions were phenol/chloroform-extracted, precipitated, and analyzed on 8% polyacrylamide-8 M urea sequencing gels.

Generation of TrkA Promoter-Luciferase Contructs-- The 3.1-kb HindIII fragment from pJD1 was cloned into the HindIII site of pGL3 to generate pTrkLuc3.1 and pTrkLuc3.1AS (antisense). The pGL3basic vector was modified prior to this insertion by digestion with MluI, treatment with T4 DNA polymerase, and religation to remove the MluI site. pTrkLuc2.8 and pTrkLuc2.7 were made by partial digestion of pJD1 with NarI and treatment with Klenow fragment, followed by complete digestion with SpeI. The 2.7- and 2.8-kb fragments were then inserted into the HindIII and NheI sites of pGL3. pTrkLuc2.6 was produced by isolating a 2.6-kb KpnI-AflIII fragment from pTrkLuc3.1 and inserting it into pGL3 at the KpnI and HindIII sites. pTrkLuc1.2 was generated from pTrkLuc2.6 by digestion with NheI and SacII and religation of the 6.0-kb plasmid fragment. pTrkLuc1.0 and pTrkLuc0.7 were produced by digestion of pTrkLuc2.6 with BglII and SacI, respectively, followed by religation of the vector fragments. pTrkLuc0.2 was made from pTrkLuc2.6 by digestion with BamHI and NheI and self-ligation of a 5.1-kb vector fragment, and pTrkLuc0.14 was generated from pTrkLuc0.2 by digestion with SacI and ApaI and self-ligation of the 5.0-kb vector fragment.

Transient Transfection and Reporter Gene Assays-- Supercoiled plasmid DNAs were prepared on Qiagen columns (Qiagen, CA), and constructs were transfected in duplicate. Co-transfected pCMV--galactosidase DNA was used for normalizing transfection efficiency and each construct was tested a minimum of three times. Nonspecific promoter activity was determined in each experiment using the antisense construct pTrkLuc3.1AS and its value was subtracted from the activity of each sense promoter construct after normalization. Nonspecific activities were approximately 10, 22, and 27% of total promoter activities for SMS-KCN, SH-SY5Y, and K293 cells, respectively. Cells were seeded 12-24 h prior to transfection in 35-mm culture dishes at 0.3-0.8 × 10⁶ cells/dish. 1.5 µg of pTrkLuc construct, and 0.75 µg of pCMV--galactosidase were combined with 6 µl of LipofectAMINE (Life Technologies, Inc.) in 200 µl of RPMI 1640 medium (antibiotic- and serum-free) and incubated for 30 min at room temperature. The DNA mixture was diluted with 800 µl of RPMI 1640 medium and then added to cells for 5 h at 37 °C. Cells were then washed with culture medium containing serum and antibiotics and cultured for an additional 43 h. Transfected cells were washed with phosphate-buffered saline, lysed with 200 µl 1 × lysis buffer (Promega, Madison, WI), and supernatants were stored at 80 °C.

Gel Shift Analysis-- Protein extracts were prepared from freshly isolated nuclei in 20 mM Hepes, pH 7.9, 400 mM NaCl, 1 mM EGTA, 1 mM EDTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and 1 µg/ml each of the proteinase inhibitors pepstatin A, bestatin, aprotinin, and leupeptin. Protein was determined using a Bio-Rad assay kit, and relative concentrations in different extracts were verified on SDS-polyacrylamide gels and Coomassie Blue staining.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Relative Transcription Rates of the TrkA Gene in Human Neuroblastoma Cell Lines-- Neuroblastoma cell lines exhibit varying degrees of TrkA expression (41-43) and thus can be used to examine the basis for both its cell-specific and variable expression in this tumor cell type. Gene transcription often plays a primary role in cell-specific gene expression (44). To determine whether this was the case for TrkA in human neuroblastoma cells, Northern blots were initially performed using different cell lines (Fig. 1). SK-N-SH, IMR32, and SMS-KCN cells contained relatively high levels of a 2.9-kb TrkA mRNA, with IMR32 being the most enriched. In contrast, SH-SY5Y cells contained low levels of the 2.9-kb transcript, and LA-N-6 cells expressed intermediate levels. TrkA mRNA was undetectable in human kidney K293 cells (data not shown).

View larger version (42K): [in this window] [in a new window]   Fig. 1.   Northern blot analysis of TrkA mRNA levels in different neuroblastoma cell lines. Total RNA (20 µg) was examined from the following cell lines: SK-N-SH (lane 1), SMS-KCN (lane 2), SH-SY5Y (lane 3) LA-N-6 (lane 4), SMS-KCN (lane 5), IMR32 (lane 6), and SH-SY5Y (lane 7). The size of the human TrkA mRNA is indicated in kilobases. Membranes were subsequently stripped and rehybridized with a probe to human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (shown below).

View larger version (55K): [in this window] [in a new window]   Fig. 2.   Transcriptional analysis of the human TrkA gene in different cell lines. Equivalent amounts of 32P-labeled RNAs were generated from K293, SH-SY5Y, SMS-KCN, and IMR32 cell nuclei and then hybridized to the indicated plasmid DNAs.

Identification of the Transcription Initiation Region for the Human TrkA Gene in Neuroblastoma Cells-- A combination of RNase protection and primer extension analysis was used to accurately localize the TrkA promoter region employed in human neuroblastoma cells. RNase protection was performed with a riboprobe (pTrk-NarI) complementary to sequences spanning the 5-end of the human TrkA cDNA originally isolated from K562 erythroleukemia cells (31) (Fig. 3A). Analysis of total RNA from SH-SY5Y, SMS-KCN, and IMR32 cells revealed two prominent protection products of ~110 and 125 nucleotides (nt) in each case (Fig. 3B). Additional products ~150-160 nt in length sometimes were observed upon longer exposure (e.g. in IMR32 cells, Fig. 3B). No specific products were detected for TrkA-negative K293 kidney cells. The two major start sites predicted by these results lie ~60-70 bp downstream of the 5-end for the TrkA cDNA originally isolated from K562 cells (Fig. 3A).

View larger version (40K): [in this window] [in a new window]   Fig. 3.   Mapping the transcription start sites for the human TrkA gene. A, schematic representation of the human TrkA genomic fragment showing the transcription start sites determined by RNase protection (asterisks) and primer extension (square). An additional start site sometimes observed by RNase protection is shown in parentheses. Locations of the primer used for extension analysis (a and arrow), translation initiation site (ATG and arrow), and 5-end of the TrkA cDNA from K562 cells (inverted arrowhead) are also shown. The pTrkA-NarI riboprobe used in RNase protection assays is indicated by the bar below. B, mapping of start sites by RNase protection. Total RNA (20 µg) was analyzed from SH-SY5Y (lane 1), SMS-KCN (lane 2), IMR32 (lane 3), and K293 (lane 5) cell lines. Yeast tRNA (lane 4) served as a negative control. C, mapping of start sites by primer extension. Total RNA (50 µg each except lane 1, 80 µg) was examined from SH-SY5Y (lane 1), SMS-KCN (lane 2), IMR32 (lane 3), and K293 cells (lane 4). The specific extension product is indicated by an asterisk. In both A and B, the sizes of DNA markers are shown in nucleotides (nt).

Localization of Neuroblastoma Cell-specific Elements within the Human TrkA Promoter-- To identify regions within the TrkA promoter that are important for expression in neuroblastoma cells, a series of promoter constructs were generated in the luciferase-containing plasmid, pGL3basic (Fig. 4). The largest of these, pTrkLuc2.8, consisted of a 2.8-kb HindIII-NarI fragment containing ~2.6 kb of 5-flanking sequence and a portion of exon I. Additional versions consisted of 5- and/or 3-deletions of pTrk2.8 as well as an antisense construct containing a 3.1-kb genomic fragment that served as a control for nonspecific activity. The low TrkA-expressing SH-SY5Y line and TrkA-negative K293 cells were transiently transfected with these various constructs to determine the presence of neuroblastoma cell-specific regulatory sequences within the human TrkA promoter. Transfection of SH-SY5Y cells yielded comparable promoter activity for all constructs tested (Fig. 5A). The smallest of these, pTrkLuc0.14, consisted of a 138-bp proximal 5-flanking sequence defined by ApaI and AflIII sites lying just upstream of the transcription initiation region (see Fig. 4). While the activities of various constructs were generally constant in K293 cells as well, their absolute levels were ~10-fold lower than in SH-SY5Y cells (Fig. 5B). Thus, the 138-bp ApaI-AflIII region is sufficient to mediate neuroblastoma cell-specific regulation of the TrkA promoter. Based on the various constructs tested, no significant enhancer or repressor sequences are apparent lying upstream or downstream of this region.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Schematic diagram of human TrkA genomic sequences used to generate pTrkLuc promoter constructs. The 3.1-kb HindIII fragment derived from the human TrkA gene is shown at the top along with relevant restriction sites. The 5-end of exon I (shaded box) is based on primer extension analysis. Genomic fragments used in various pTrkLuc constructs are indicated below and were cloned into the pGL3basic reporter plasmid.

View larger version (39K): [in this window] [in a new window]   Fig. 5.   Cell specificity of TrkA promoter activity in transiently transfected cells. A, luciferase activities for different pTrkLuc constructs in SH-SY5Y neuroblastoma cells. B, reporter activities of the same constructs in TrkA-negative human kidney K293 cells. For comparison, the activity of the pTrkLuc2.6 construct in SH-SY5Y cells is also shown. Bars indicate the standard errors of duplicate samples.

Promoter Sequences Mediating Variable TrkA Expression in Neuroblastoma Cells-- Northern and nuclear run-on analyses demonstrated that TrkA transcription occurs to varying degrees in human neuroblastoma cell lines, similar to what is observed in primary tumors. To explore the nature of the regulatory elements responsible for elevated transcription of the TrkA gene in neuroblastoma cells, promoter constructs also were transiently transfected into SMS-KCN cells. This neuroblastoma cell line exhibits a severalfold higher level of TrkA mRNA and transcription than occurs in SH-SY5Y cells (see Figs. 1 and 2). As with the other two cell lines examined, the activity of the pTrkLuc0.14 promoter construct was very similar to the longer versions tested (Fig. 6). In addition, absolute promoter activity was ~5-fold greater than observed in SH-SY5Y cells. This indicates that enhanced TrkA promoter activity is recapitulated in transiently transfected SMS-KCN cells and that the relevant cis-elements are contained within a 138-bp proximal 5-flanking region. It should be noted that further deletion of 102 bp from the 5-end of the ApaI-Af1III sequence resulted in >93% loss of luciferase activity in all three cell lines (data not shown), confirming the critical importance of this region for TrkA promoter activity.

View larger version (48K): [in this window] [in a new window]   Fig. 6.   Promoter activity in neuroblastoma SMS-KCN cells which express TrkA at relatively high levels. Reporter gene activities were determined as outlined in Fig. 5. The activity of pTrkLuc2.6 in transfected SH-SY5Y (low TrkA-expressing) cells is shown for comparison.

Cell-specific Binding of Nuclear Factors to the Neuroblastoma Regulatory Region of the TrkA Promoter-- The above findings indicated that nuclear factor interactions within the 138-bp ApaI-AflIII region mediated both cell-specific and variable expression of the TrkA promoter in neuroblastoma cells. Gel shift experiments were performed to examine these interactions directly and compare them in different cell lines using the ApaI-AflIII sequence as a probe. Four major complexes were detected in nuclear extracts prepared from SMS-KCN cells (Fig. 7, a-d), which express the TrkA gene at relatively high levels. All four complexes were specifically competed by a 50-fold excess of unlabeled homologous DNA but not by an equivalent amount of an unrelated competitor DNA. In SH-SY5Y cells, which express TrkA at low levels, complex a was present in concentrations similar to SMS-KCN cells but complex b was markedly reduced, and complexes c and d were extremely low (Fig. 7). The complexes detected in SH-SY5Y cells also were specifically competed by homologous unlabeled competitor. Further, all four complexes were essentially undetectable in TrkA-negative K293 cells. Thus, multiple DNA binding proteins interact with the proximal TrkA regulatory region mediating cell-specific and variable expression in TrkA-expressing neuroblastoma cells. These factors are extremely low in TrkA-negative cells and at least three of them (complexes b, c, and d) are specifically elevated in neuroblastoma cells exhibiting enhanced TrkA promoter activity.

View larger version (77K): [in this window] [in a new window]   Fig. 7.   Nuclear factor binding to the neuroblastoma regulatory region of the TrkA promoter. Gel shift analyses were performed with an end-labeled ApaI-AflIII genomic fragment. Lane 1, free probe; lanes 2-4, nuclear extracts from SMS-KCN cells; lanes 5-7, extracts from SH-SY5Y cells; lanes 8-10, extracts from K293 cells. Three µg of nuclear protein were used per lane. The presence or absence of a 50-fold excess of homologous or unrelated competitor DNA is indicated above each lane. The positions of the four DNA-protein complexes observed in SMS-KCN cells (a-d) are also indicated.

View larger version (21K): [in this window] [in a new window]   Fig. 8.   Sequence of the human TrkA 5-proximal promoter region. Potential binding sites for known DNA binding proteins are shown by boxes. Sequences that were missing in the human TrkA gene sequence reported by Greco et al. (62) are indicated by a dash (-), and two bases that were inverted are shown in lowercase letters above the corresponding sequences.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

While TrkA expression and/or signaling is often deficient in neuroblastoma cells derived from advanced stage tumors (45, 46), its signal transduction pathway was shown to be intact in cells obtained from a favorable primary tumor (16). Forced expression of TrkA in receptor-deficient neuroblastoma cells also leads to growth arrest and differentiation of tumors in nude mice following in vivo treatment with NGF (21). Recent studies have further suggested that Schwann cells are capable of infiltrating favorable neuroblastomas and may provide a local source of NGF (47). These and other studies have led to the hypothesis that TrkA has an active role in the spontaneous maturation/regression of favorable neuroblastoma tumors (16, 24). However, a direct role for NGF and TrkA signaling in this process remains to be established.

The strong correlation between high TrkA expression and a favorable outcome for neuroblastoma (16, 21-23) indicates that understanding the mechanisms responsible for TrkA expression in neuroblastoma cells is likely to provide important insight into this disease and may assist in the development of potentially novel treatments. For example, it could lead to therapies focused on induction of TrkA that, in combination with neurotrophin treatment, could promote tumor differentiation. Even if TrkA expression is simply a marker for a more differentiated tumor state, such analyses will reveal fundamental regulatory pathways that distinguish TrkA-expressing neuroblastoma cells from their TrkA-negative counterparts.

The present studies form a strong basis for this approach by demonstrating a major role for gene transcription in the elevated expression of TrkA in neuroblastoma cells. In contrast, enhanced expression of N-MYC in human neuroblastoma cells is regulated largely at the level of mRNA stability (48). Distinct mechanisms thus determine differential expression of these markers for neuroblastomas having favorable and poor prognoses. Cis elements mediating both cell-specific and elevated TrkA transcription in neuroblastoma cells have been localized to a 138-bp proximal promoter sequence. Differential gene transcription can occur by various mechanisms, including expression of unique transcription factors, elevated levels, and/or novel combinations of more generally expressed factors and epigenetic processes such as chromatin-dependent transcriptional effects or DNA methylation (49-51). The present studies indicate that differential expression of DNA-binding proteins plays an important role in TrkA promoter regulation in neuroblastoma cells. That is, several DNA-protein complexes were identified involving the proximal regulatory promoter region that were common to TrkA-expressing neuroblastoma cells and extremely low or undetectable in TrkA-negative cells. Further, certain complexes were markedly elevated in a neuroblastoma cell line that transcribes the TrkA gene at relatively high levels. Although these findings do not rule out a role for chromatin structure or DNA methylation in TrkA promoter regulation, such mechanisms are not required for differential activation within the proximal 5-flanking region defined here since it is unlikely they contribute to promoter activity determined in transient transfection assays.

Several candidate elements were identified within the proximal regulatory region that may be bound by transcription factors specifically elevated in TrkA-expressing neuroblastoma cells. These include several GC-box sites such as for Sp1, AP2, and Egr-1, as well as sites for ATF and ets-related factors. Sp1 elements have been implicated in cell-specific or differentiation-associated regulation of several promoters, including those for p21/WAF1, hepatocyte growth factor and human KDR/flk-II (52-54). Further, Egr-1 sites are involved in the regulation of certain neuronally expressed promoters, and this factor is expressed and inducible in neuroblastoma cells (55, 56). ATF and AP2 family members also have an important role in neuronal gene expression (55, 57-59). Finally, ets domain proteins have been implicated in cell-specific gene expression and at least some members of this family are restricted to or enriched in neuronal cells, including PEA3 and NERF (55, 60, 61). Thus, the present findings define a region critical for TrkA promoter expression in neuroblastoma cells which contains several candidate regulatory elements. It is also possible that additional, novel regulatory elements exist within the proximal 5-flanking region that remain to be identified. Their definition should provide important insight into the mechanistic basis for TrkA gene regulation in neuroblastoma.