Abstract
Angiogenesis is essential for tumor growth and metastasis and depends on the production of angiogenic factors. Mechanisms regulating the expression of angiogenic factors in tumor cells are largely unknown. High expression of the neurotrophin receptor TrkA in neuroblastomas (NBs) is associated with a favorable prognosis, whereas TrkB is mainly expressed on aggressive, MYCN-amplified NBs. To investigate the biological effects of TrkA and TrkB expression on angiogenesis in NB, we examined the expression of angiogenic factors in the human NB cell line SY5Y and its TrkA and TrkB transfectants. In comparison with parental SY5Y cells, mRNA and protein levels of the examined angiogenic factors were significantly reduced in SY5Y-TrkA cells, whereas SY5Y-TrkB cells did not demonstrate a significant change. Conditioned medium of TrkB transfectants and parental SY5Y cells induced endothelial cell proliferation and migration, but this effect was completely absent in SY5Y-TrkA cells. TrkA expression also resulted in severely impaired tumorigenicity in a mouse xenograft model and was associated with reduced angiogenic factor expression and vascularization of tumors, as determined by immunohistochemistry and an in vivo Matrigel assay. TrkA expression inhibits angiogenesis and tumor growth in SY5Y NB cells by down-regulation of angiogenic factors, whereas expression of TrkB does not down-regulate the production of these angiogenic factors. The biologically different behavior of TrkA- and TrkB-expressing NBs may be explained in part by their effects on angiogenesis.
INTRODUCTION
Neovascularization is essential for tumor growth and metastasis formation (1) . The development of new blood vessels in tumors depends on the production of angiogenic factors released from the tumor cells and/or stromal cells. Of the known angiogenic factors, two well-characterized peptides, VEGF 3 ,(2) and bFGF (3) , have been implicated in the neovascularization of a wide variety of tumors (4 , 5) . Other angiogenic factors such as VEGF B, VEGF C, Ang1 and Ang2, and TGF-α have also been shown to induce angiogenesis in vitro and in vivo (6 , 7) . One environmental condition known to enhance VEGF expression is hypoxia (8) . Recent studies have implicated activation of the tyrosine kinase activities of the src family of proto-oncogenes and expression of mutant H- or K-ras oncogenes as important in the induction of VEGF (9 , 10) . However, mechanisms that regulate expression of angiogenic factors in tumor cells are still not well understood, and the signal transduction pathways that regulate VEGF production have yet to be elucidated.
Expression of different members of the Trk family (tyrosine kinase receptors for neurotrophins) plays an important role in the heterogeneous biological and clinical behavior of NB, the most common extracranial malignant solid tumor of childhood (11 , 12) . Observations from several independent studies suggest that high expression of TrkA is present in the subset of NB with favorable biological features and is highly correlated with patient survival (13 , 14) , whereas expression of TrkB is correlated with unfavorable, aggressive NB (12) . Expression of TrkA has been shown to mediate differentiation in vitro (15 , 16) , whereas TrkB has been associated with enhanced survival of NB cells (12 , 17) . However, no differences in signal transduction pathways used by TrkA and TrkB have been reported, and hence the molecular events underlying the different behavior of TrkA- and TrkB-expressing NBs in vivo are still unknown.
High-level expression of angiogenic factors is associated with advanced tumor stage in human NB (18) , and evidence suggests that higher vascularity correlates with metastasis, MYCN amplification, unfavorable histology, and poor outcome of NB (19) . The purpose of this study was to determine whether expression and biological activity of important angiogenic factors such as VEGF and bFGF were influenced by expression of TrkA or TrkB in NB cells in vitro and in vivo.
MATERIALS AND METHODS
Cell Culture and Transfection.
SH-SY5Y is a neuronal subclone from the SK-N-SH NB cell line and has been described previously (20) . The NB cell line NB69 was obtained from the cell line bank of The Children’s Hospital of Philadelphia. Full-length TrkA cDNA and K538N cDNA (a TrkA mutant defective in the ATP-binding site of the Trk kinase domain) were a generous gift from David Kaplan and were cloned into the retroviral expression vector pLNCX (21) . The TrkB cDNA was cloned into the plasmid vector pcDNA3 (Invitrogen, Carlsbad, CA). All cells were grown at 5% CO2 in RPMI 1640 (Life Technologies, Inc.) supplemented with 10% fetal bovine serum, l-glutamine, and gentamicin (all from Life Technologies, Inc.). Medium was changed twice a week. The TrkA-WT and K538N constructs were transfected into the packaging cell line Bing by electroporation, and virus-containing supernatant (10 ml) from these cells was added to LipoTaxi (Stratagene, La Jolla, CA) and used to infect SH-SY5Y cell cultures (22) . Transfection of the TrkB construct was carried out using N-[1-(2,3-dioleoyloxyl)propyl]-N,N,N-trimethylammoniummethyl sulfate (Sigma Chemical Co., St. Louis, MO). Transfected cells were subjected to selection with 500 μg/ml G418 (Sigma Chemical Co.). Surviving, drug-resistant SY5Y cells were subcloned by limited dilution to obtain single-cell clonal lines. As negative controls, SY5Y cells were infected with a retrovirus bearing the pLNCX empty retroviral vector (SY5Yvec). For a comparison of different cell lines, the TrkA-WT construct and the empty vector were also transfected into NB69 NB cells as described above. The identity of all transfectants was confirmed by sequencing after transfection.
RT-PCR.
The mRNA Expression of TrkA, TrkB and angiogenic factors was analyzed by RT-PCR using specific biotinylated primers (sequences available upon request). Total RNA was extracted with the RNeasy Kit (Qiagen, Valencia, CA), reverse transcribed and amplified for 20 cycles on a PTC-100 Programmable Thermal Controller (MJ Research, Inc., Watertown, MA) using the Superscript amplification system (Life Technologies, Inc.). The PCR products were run on a 6% polyacrylamide gel and transferred to a nylon membrane (Hybond N+, Amersham, IL). Biotinylated signals were detected using the Southern Light Detection system (Tropix) and exposed to X-ray film. Target gene expression was normalized to the coamplified housekeeping gene GAPD as described (23) .
Determination of VEGF and bFGF Protein Levels in CM.
Cell cultures were grown at 80% confluency for 24 h in serum-free RPMI medium and, where indicated, 100 μm desferoxamine mesylate to mimic hypoxia (24) . Cell supernatants were collected, centrifuged at 3000 rpm for 10 min and stored at −80°C. Concomitantly, cell pellets were harvested and cell number was determined by counting the cells in a hemocytometer. The amount of VEGF and bFGF protein in the CM was determined with ELISA kits (R&D Systems, Minneapolis, MN). VEGF and bFGF were expressed as pg protein/106cells/24 h.
Endothelial Proliferation and Migration Assay.
HUVECs (Clonetics Corp., Walkersfield, MD) were plated in 24-well tissue culture plates at a concentration of 5.0 × 104 cells/well in medium containing optimal concentrations of growth factors (EGM medium; Clonetics Corp.). After starving the cells overnight in basic medium containing no growth factors (EBM medium; Clonetics Corp.), HUVECs were incubated at 37°C for 96 h in CM of parental and transfected SY5Y cell clones (50-fold concentrated supernatant, diluted 1:3 in EBM + 1% fetal bovine serum). After incubation, HUVECs were stained and counted as described previously (25) . To evaluate their capacity to block EC proliferation induced by SY5Y-CM, specific antibodies against VEGF and bFGF (Santa Cruz Biotechnology, Santa Cruz, CA) were preincubated with SY5Y-CM. Preincubation with normal mouse IgG (Vector Laboratories, Burlington, CA) was used as control. To determine the ability of NB CM to activate migration of ECs, we used Transwell inserts (Costar) containing a polycarbonate membrane with 8-μm pores as described previously (26) . Migration was quantitated by counting the cells that migrated to the bottom of the membrane in 10 high power fields at ×320 magnification.
In Vivo Tumor Growth.
SY5Y, SY5Y-TrkA, and SY5Y-TrkB cells were cultured to 80% confluence, harvested, and resuspended in Matrigel (Collaborative Biomedical Products, Bedford, MA). Four-week-old female athymic NCR (nu/nu) mice (National Cancer Institute, Frederick, MD) were inoculated s.c. in the flank with 2.5 × 107 cells/inoculate as described previously (27) . Tumors were measured two times a week, and calculations of tumor volume were made by multiplying the product of three dimensions by π/6. Harvested tumors were cut into two fragments: one part was fixed in formalin for immunohistochemistry; and one part was snap-frozen in liquid nitrogen to obtain RNA.
Immunohistochemistry.
Three different randomly chosen xenografts of SY5Y and SY5Y-TrkB as well as the two available TrkA xenografts were analyzed by immunohistochemistry. Tumor specimens were sectioned (5 μm) and processed for standard immunohistochemical staining. Anti-VEGF rabbit polyclonal antibody A-20 (Santa Cruz Biotechnology) was used at a 1:400 dilution, anti-bFGF rabbit polyclonal antibody (Sigma Chemical Co.) was used at a 1:200 dilution, and anti-Ki-67 rabbit polyclonal antibody NCLki67p (Novocastra Laboratories, New Castle, United Kingdom) was used at a 1:1600 dilution. After incubation with an antirabbit secondary antibody, the immune complex was visualized with the StrAviGen Super Sensitive kit (BioGenex Laboratories, San Ramon, CA) using diaminobenzidine tetrahydrochloride as a chromogen. Sections were counterstained with hematoxylin. Microvessel staining was performed by EC labeling with GSI lectin (Sigma Chemical Co.) at a 1:10 dilution as described previously (28) . Evaluation of immunostained sections was carried out independently by two observers. The intensity of VEGF, bFGF, and GSI lectin staining in representative tissue sections was graded on a scale of 0–3+. The proliferation index (number of Ki-67-positive cells/100 neoplastic cells) was determined by counting the percentage of Ki-67-positive cells in at least 1000 tumor cells at ×400 magnification in the most densely stained region.
Matrigel Assay for Angiogenesis.
In vivo angiogenesis assays were carried out by injecting athymic NCR (nu/nu) mice with 0.5 ml of Matrigel (Collaborative Biochemical Products) alone or mixed on ice with 107 SY5Y, SY5Y-TrkA, or SY5Y-TrkB cells as the angiogenic stimulus. Matrigel was injected s.c. in the abdominal midline on day 0 in all experiments as described previously (29) . Matrigel plugs were harvested on day 11 for hemoglobin assay. Quantitation of hemoglobin was performed by the Drabkin method (30) .
RESULTS
Trk Expression and Autophosphorylation of Stable SY5Y Transfectants.
We stably expressed full-length constructs of TrkA, TrkB, a kinase-inactive TrkA mutant (K538N), and the empty vector control (Vec) in the human NB cell line SH-SY5Y. Expression of functional Trk receptors was examined by RT-PCR and immunoblotting. The SY5Y-TrkA and SY5Y-TrkB cell lines demonstrated high and comparable levels of expression of the inserted receptor on mRNA and protein levels (data not shown). Each receptor was autophosphorylated at a low level even before the addition of ligand. A strong increase in autophosphorylation of the receptor was seen after addition of the appropriate ligand NGF or brain-derived neurotrophic factor to SY5Y-TrkA and SY5Y-TrkB cells, respectively (data not shown). In contrast, there was no detectable Trk expression or autophosphorylation in SY5Y-Vec or untransfected SY5Y cells. SY5Y-K538N showed high expression of the TrkA receptor on mRNA and protein levels, but no autophosphorylation was demonstrated.
Decrease in Angiogenic Factor mRNA Levels due to Expression of TrkA.
We analyzed the effect of TrkA and TrkB on expression of VEGF, VEGF B, VEGF C, bFGF, Ang1, Ang2, and TGF-α by semiquantitative RT-PCR. SY5Y-TrkA cells demonstrated a clear decrease in mRNA expression of the most abundant VEGF isoforms, VEGF165 and VEGF121, and Ang1, in comparison with parental SY5Y and SY5Y-Vec cells. Even more striking was the effect of TrkA transfection on mRNA expression of bFGF and Ang2. Whereas parental SY5Y and SY5Y-Vec cells expressed high levels of bFGF and Ang2 mRNA, we could not detect these mRNAs in SY5Y-TrkA cells (Fig. 1) ⇓ . VEGF C was not expressed in any of the cell clones, and VEGF B was expressed at similar moderate levels in all of the cell clones examined (data not shown). SY5Y-TrkB cells expressed slightly decreased mRNA levels of Ang2 and slightly increased mRNA levels of VEGF121 and Ang1 (Fig. 1) ⇓ , but no other changes in angiogenic factor expression were observed. Transfection of SY5Y with the kinase-inactive TrkA mutant K538N resulted in decreased expression of Ang2, but expression of all other angiogenic factors was comparable with that seen in parental SY5Y cells.
Expression of TrkA down-regulates angiogenic factors on mRNA levels in SH-SY5Y cells. A representative example of semiquantitative RT-PCR showing expression levels of the angiogenic factors VEGF (VEGF165 and VEGF121), bFGF, Ang1, Ang2, and TGF-α in the indicated cell clones is shown. Expression of the housekeeping gene GAPD was used as internal control. Lower expression levels were observed of all analyzed angiogenic factors in TrkA-transfected SY5Y cells. Expression levels shown are the representative results of three independent experiments with at least two different cell clones.
Significant Reduction of VEGF and bFGF Protein Levels in Cell Supernatants of SY5Y-TrkA Cells.
Parental SY5Y and SY5Y-Vec cells produced high quantities of VEGF (∼700 pg/106 cells/24 h) but less bFGF protein (∼3 pg/106 cells/24 h). SY5Y-TrkB cells produced increased amounts of VEGF and bFGF protein in their supernatants in comparison with parental cells. Interestingly, a 4.6-fold decrease in the production of VEGF protein was observed in the supernatants of SY5Y-TrkA cells compared with parental SY5Y cells (Fig. 2) ⇓ . Secretion of bFGF protein into CM of TrkA transfectants was virtually undetectable (data not shown). The presence of exogenous NGF did not significantly change the protein levels of VEGF or bFGF produced by SY5Y-TrkA cells, suggesting that activation of TrkA by exogenous NGF might not be required for the effect of TrkA on angiogenic factor expression. Corresponding to the mRNA expression levels, VEGF and bFGF protein levels in the kinase-inactive TrkA mutant K538N were comparable with the levels seen in parental SY5Y cells, indicating that a functional TrkA receptor molecule is required for the inhibiting effect on angiogenic factor production.
Expression of TrkA in SY5Y cells reduces VEGF protein levels in the supernatant. The indicated cell lines were cultured with serum-free medium for 24 h under normoxic and hypoxic conditions. TrkA- and TrkB-transfected SY5Y cells were cultured in the absence and presence of their specific ligands (NGF and brain-derived neurotrophic factor) as indicated. VEGF concentrations in the CM were determined by specific VEGF ELISA and are expressed as pg/106 cells/24 h. Shown are the representative results of one of three different clones of transfected cells tested. Data points are the mean of triplicates. SDs were <10%.
, normoxic conditions; ▪, hypoxic conditions.
Because hypoxic stimulation of VEGF has been described previously in NB (31) , we examined the response to hypoxia in parental and transfected SY5Y cell clones. Hypoxia-like conditions (mimicked by the addition of desferoxamine mesylate) resulted in a 6–10-fold up-regulation of VEGF protein production in all cell clones (Fig. 2) ⇓ including the TrkA transfectants. In SY5Y-TrkA cells, the hypoxia-induced up-regulation of the very low basal VEGF protein levels resulted in measurable amounts of VEGF, but the corresponding CM of SY5Y and SY5Y-TrkB cells still contained about 6-fold more immunodetectable VEGF. Up-regulation of protein production was even more striking for bFGF. Whereas SY5Y, SY5Y-Vec, SY5Y-TrkB, and SY5Y-K538N demonstrated a 20–30-fold induction under hypoxia-like conditions, SY5Y-TrkA cells only produced an 8–10-fold induction of basal bFGF protein levels (data not shown). This resulted in 20–25-fold reduced bFGF protein levels in TrkA supernatants under hypoxic conditions. Thus, hypoxic CM can also induce VEGF and bFGF protein secretion in SY5Y-TrkA cells, but the magnitude of the effect is markedly reduced in comparison with that seen in parental SY5Y cells.
SY5Y-TrkA Cells Have Diminished Ability to Stimulate HUVEC Proliferation and Migration in Vitro.
One of the initial steps of tumor-induced neovascularization is EC proliferation and movement toward the chemoattractive stimuli produced by the tumor cells. Therefore we sought to determine whether the expression of TrkA affected the ability of SY5Y cells to activate endothelial proliferation and migration by factors secreted into the CM. The data shown in Fig. 3, A and B ⇓ , indicate that CM from the parental cell line SY5Y, SY5Y-Vec, SY5Y-K538N, or SY5Y-TrkB were highly stimulatory for HUVEC proliferation, whereas CM from TrkA-expressing SY5Y cell clones demonstrated minimal stimulation, regardless of the presence of NGF. For a comparison of different cell lines, we also analyzed the effect of CM obtained from TrkA-transfected NB69 NB cells on EC proliferation. Whereas CM of this cell line transfected with the empty vector control was highly stimulatory for EC proliferation, CM of NB69-TrkA cells did not stimulate EC proliferation (data not shown). The notion that VEGF is the major element in the CM that promoted HUVEC proliferation was tested by including a VEGF-neutralizing polyclonal antibody in the CM of SY5Y cells during the proliferation assay (Fig. 3B) ⇓ . This antibody completely blocked the stimulation of EC proliferation elicited by CM of SY5Y. In contrast, a bFGF-neutralizing monoclonal antibody added to the CM of SY5Y cells only reduced the stimulatory effect on HUVEC proliferation to 50%. No effect was seen by the addition of normal mouse IgG. These data suggest that VEGF is the major component in the CM of NB cells that stimulates HUVEC proliferation and migration. An endothelial migration assay produced results similar to those of the proliferation assay (data not shown).
Effect of NB cell supernatant on the proliferation of HUVECs. Parental and transfected SY5Y cells were cultured in serum-free media for 24 h under normoxic conditions. Supernatants were concentrated 50-fold and diluted 1:3 with basic EBM medium. HUVECs were cultured in the indicated CM, and cells were counted after 96 h. A, representative example of HUVECs cultured with SY5Y CM (bottom left) and CM of TrkA transfectants (bottom right) in comparison with full EGM medium (top left; containing EC growth factors) and basic EBM medium (top right; containing no growth factors). Phase-contrast microscopy, ×320. B, HUVEC proliferation in relation to control media. HUVEC counts cultured in full EGM medium as positive control were determined to be 100%, and HUVEC counts cultured in basic EBM medium were determined to be 0%. HUVEC proliferation of the tested CM is expressed as a percentage of HUVEC proliferation in full medium. For additional positive controls, recombinant human VEGF (100 ng/ml) or recombinant human bFGF (10 ng/ml) was added to basic EBM medium. SY5Y CM was preincubated with specific anti-VEGF or anti-bFGF monoclonal antibodies to test their neutralizing effect on HUVEC proliferation.
TrkA Mediates Inhibition of in Vivo Tumor Growth and Vascularity.
Having demonstrated that down-regulation of angiogenic factor production by expression of TrkA had substantial in vitro effects on HUVEC proliferation and migration, we next sought to determine the effects on in vivo behavior. We implanted parental SY5Y cells as well as TrkA- and TrkB-transfected SY5Y cells s.c. in the flank of athymic nude mice. Cells from parental SY5Y and SY5Y-TrkB clones formed tumors with similar kinetics and of similar size 15–20 days after injection, whereas tumor growth was remarkably suppressed for SY5Y-TrkA cells (Fig. 4) ⇓ . Injection of up to 5 × 108 SY5Y-TrkA cells into 24 mice resulted in the growth of only 2 small tumors (tumor take, 8%), and these started to grow only after a prolonged latency period of about 90 days (data not shown). One of these tumors (TrkA1) demonstrated a very slow growth behavior, remained small for several weeks, and appeared macroscopically less vascular than TrkB and SY5Y tumors when it was harvested after 120 days. Four additional SY5Y-TrkA xenografts formed small nodules that were barely palpable and remained in a dormant state for >6 months. All other TrkA xenografts failed to form tumors altogether (18 of 24 mice). In comparison, SY5Y and SY5Y-TrkB cells had a tumor take of 83% and 100%, respectively.
Tumor growth in nude mice. SY5Y, SY5Y-TrkA, and SY5Y-TrkB cells (2.5 × 107) in Matrigel were injected s.c. in the flank of nude mice (6 mice each for SY5Y and TrkB; 24 mice for TrkA). Tumor volumes were estimated at the indicated times after implantation, and data are shown as the mean of each group. The experiment was terminated when tumors reached a volume of 300–400 mm3.
Fig. 5 ⇓ shows the results of immunostaining for VEGF and bFGF and staining for ECs in representative tumor sections of SY5Y-, TrkA-, and TrkB-transfected cells. SY5Y and TrkB transfectants are highly positive for VEGF and bFGF protein. In contrast, the slowly growing TrkA1 tumor is largely negative for VEGF and bFGF. Consistent with this pattern, a reduction in the number of microvessels in the TrkA1 tumor versus the SY5Y and TrkB tumors was also observed (Fig. 5) ⇓ . The proliferation index was determined by Ki-67 staining. Ki-67-immunoreactive cells were scattered throughout the tumors, and the Ki-67 proliferation indices were similar in SY5Y, SY5Y-TrkA, and SY5Y-TrkB xenografts (data not shown). To confirm the results obtained by immunostaining, an alternative method was used to assess tumor angiogenesis. Matrigel implants were used as an in vivo assay to quantitate vascularization by measuring the hemoglobin content of the Matrigel plugs. Matrigel implants without additives were pale and unvascularized 11 days after implantation. Inclusion of SY5Y or SY5Y-TrkB cells in the Matrigel implant provided an angiogenic stimulus that made the implant visibly vascularized and resulted in a hemoglobin content of 8–12 mg/g Matrigel (Fig. 6) ⇓ . On the other hand, inclusion of SY5Y-TrkA cells in Matrigel implants resulted in significantly less vascularization (∼4 mg hemoglobin/g Matrigel), indicating that SY5Y-TrkA cells are less angiogenic in vivo.
Immunohistochemistry of NB xenografts. Immunohistochemistry of xenografts formed in nude mice by parental SY5Y cells (a, e, and i), SY5Y-TrkB cells (b, f, and j), SY5Y-TrkA1 cells (c, g, and k) and revertant SY5Y-TrkA2 cells (d, h, and l). Staining with anti-VEGF (a–d), anti-bFGF (e–h), and GSI lectin (i–l) is shown. Magnification, ×100.
Vascularization of Matrigel implants. Mice were injected s.c. with 0.5 ml of Matrigel either alone or mixed with 107 tumor cells of SY5Y, TrkA transfectants, or TrkB transfectants, as indicated. Eleven days later, the Matrigel implants were harvested, and their hemoglobin content was assayed. Each group contained five mice whose implants were assayed separately. Bars, SDs. The hemoglobin content of Matrigel plugs with TrkA transfectants is significantly less than the hemoglobin content of Matrigel plugs with parental SY5Y cells or TrkB transfectants.
Revertant Cells Regain Their Angiogenic and Tumorigenic Properties after Loss of TrkA Expression.
One mouse in the group that had been inoculated with the TrkA-transfected cells developed a tumor (TrkA2) that grew with kinetics similar to SY5Y and TrkB tumors after a prolonged latency period of 90 days. RT-PCR revealed an absence of TrkA expression in this tumor, comparable with parental SY5Y tumors, indicating that the cells had genetically reverted and lost their TrkA expression. Interestingly, VEGF and bFGF mRNA levels (Fig. 7) ⇓ and protein levels (Fig. 5) ⇓ had returned to levels comparable with those seen in parental SY5Y cells (Fig. 7) ⇓ . Also, microvessel density was similar to SY5Y and TrkB tumors in this revertant TrkA tumor. Thus, these tumor cells recovered their tumorigenicity and angiogenic capacity concomitantly with the loss of TrkA expression.
Expression of TrkA and angiogenic factors on mRNA levels in mouse xenografts. Representative example of semiquantitative RT-PCR shows expression levels of TrkA in comparison with the angiogenic factors VEGF (VEGF165 and VEGF121) and bFGF in mouse xenografts of parental SY5Y cells, TrkB transfectants, and TrkA transfectants. Expression of the housekeeping gene GAPD was used as internal control. Low expression levels of VEGF and bFGF were observed in the xenograft with high TrkA expression. Lost TrkA expression in the second TrkA xenograft corresponds to restored high mRNA expression levels of VEGF and bFGF.
DISCUSSION
Expression of TrkA and TrkB is generally thought to contribute to NB tumor biology primarily through effects on cell differentiation and proliferation (12 , 13) . However, the molecular events underlying the different behaviors of TrkA- and TrkB-expressing NBs in vivo are largely unknown. Aggressive NBs with poor outcomes are associated with higher vascularity in the tumors, and therefore angiogenesis is presumably controlled by genetic alterations in oncogene and tumor suppressor gene expression. We hypothesized that expression of TrkA or TrkB might not only affect differentiation and proliferation of NB cells but might also be involved in the regulation of tumor angiogenesis.
Expression of TrkA Inhibits Angiogenesis in Human SH-SY5Y NB Xenografts.
The results of our study indicate that TrkA overexpression in human NB cells markedly inhibits angiogenesis. We provide several lines of evidence for this TrkA-mediated effect on angiogenesis. We demonstrate (a) a decrease in the mRNA expression of several angiogenic factors, (b) a corresponding significant reduction of VEGF and bFGF protein secretion under normoxic and hypoxic conditions, (c) a diminished ability of TrkA transfectants to stimulate EC proliferation and migration in vitro, (d) a diminished ability to induce angiogenesis in vivo, and (e) reversion of these effects after loss of TrkA expression in one tumor. However, there are several additional observations and questions raised by our study.
Quantification of Tumor Angiogenesis.
There was only one TrkA tumor available in our experiments that was not revertant and suitable for immunohistochemistry. However, this tumor is most likely not representative because it is the only TrkA-expressing tumor growing beyond a size of 2 mm3 of 24 injected xenografts. To complement the studies regarding quantification of angiogenesis, we used Matrigel implants as an in vivo assay for tumor angiogenesis and quantitated vascularization by measuring hemoglobin content (29) . The results of this assay indicated that significant inhibition of angiogenesis was indeed the cause of different tumorigenicity of SY5Y-TrkA cells in comparison with SY5Y and SY5Y-TrkB cells.
Addition of Exogenous NGF Is Not Required for the Inhibitory Effect of TrkA on Angiogenesis.
The presence of exogenous NGF was not required for the reduction of angiogenic factors in SY5Y-TrkA cells. SY5Y cells express low amounts of endogenous NGF as determined by RT-PCR (data not shown), and we observed a low level of autophosphorylation in TrkA- and TrkB-transfected cells in the absence of exogenous ligands even in serum-free medium. A functional, activated TrkA receptor molecule is required for the reduction of angiogenic factor production because the kinase-inactive TrkA mutant K538N was not capable of mediating an inhibitory effect on angiogenesis. This fact supports the hypothesis that the low level of autophosphorylation seen in TrkA-transfected cells even in the absence of exogenous NGF might be sufficient to cause the inhibitory effects of TrkA on angiogenesis. Another possibility, which remains to be elucidated, is the existence of a different, NGF-independent mechanism by which TrkA down-regulates expression of angiogenic factors.
Is a 4-6-fold Suppression of VEGF a Sufficient Explanation for the Profound Antitumor Effect of TrkA?
We demonstrated a 4-6-fold suppression of VEGF protein in TrkA transfectants and observed a profound antitumor effect. Studies in VEGF knockout mice have shown that disruption of only a single VEGF allele, equivalent to 50% reduction of VEGF protein, is sufficient to block vasculogenesis and angiogenesis to such an extent that embryos die between days 11 and 12 of gestation (32) . Induced suppression of VEGF protein expression by antisense methods in a human glioblastoma by only 3-fold can almost completely obliterate the tumorigenic ability of such cells in nude mice (33) . Thus, relatively small reductions in VEGF can lead to rather profound suppression of angiogenesis. However, we cannot rule out mechanisms other than down-regulation of angiogenic factors by which TrkA might influence angiogenesis and invasiveness of NBs. Changes in the net balance of angiogenic inhibitors and activators directly affect vascularity, tumor growth, and metastasis. The fact that the inverse correlation of TrkA expression and expression of angiogenic factors such as VEGF and bFGF was not significant in 37 primary NBs (18) also suggests that mechanisms other than down-regulation of angiogenic factors are involved in the effect of TrkA on angiogenesis. We are currently analyzing the expression of angiogenic inhibitors in the CM of TrkA-expressing cells. Additional events involved in tumor progression include secretion of matrix-degrading enzymes by tumor (25) ; therefore, this should be examined further. Another mechanism by which TrkA might affect tumor growth is to influence rates of tumor cell proliferation, as shown previously in some in vitro studies. However, this effect of Trk receptor expression on proliferation is not evident in vivo, as shown by similar Ki-67 proliferation indices in SY5Y, SY5Y-TrkB, and SY5Y-TrkA tumors. This suggests that the different effect of TrkA and TrkB on angiogenesis might be more important than proliferation effects for the tumor biology and prognosis of NB.
Comparison with TrkA Data Obtained in a Different Cell Line and in Primary NBs.
We demonstrated previously that high-level expression of angiogenic factors is associated with advanced tumor stage in NB (18) . To examine whether the phenomena observed in SY5Y cells also occur in primary NB tumors, we compared expression levels of angiogenic factors with expression levels of TrkA in that study. Although there was a tendency toward inverse correlation of TrkA expression and angiogenic factor expression, the correlation did not reach statistical significance. This might be due to the small number of primary tumors examined (37) in the study panel. Another explanation may be that not only down-regulation of angiogenic stimulators but also up-regulation of angiogenic inhibitors might be involved in the inhibitory effect of TrkA on angiogenesis. This hypothesis is supported by preliminary data suggesting that EC proliferation in the CM of SY5Y-TrkA cells is still not stimulated in the presence of recombinant VEGF or bFGF (data not shown). We are currently trying to identify angiogenic inhibitors in the CM of SY5Y-TrkA cells. The comparable effects of CM obtained from SY5Y-TrkA and NB69 NB cells on EC proliferation indicate that the observed results are not just phenomena seen in a single cell line.
VEGF Is the Major Angiogenic Factor Produced by NB Cells.
VEGF and bFGF have a potent and synergistic effect on induction of angiogenesis, but VEGF is currently regarded as the major angiogenesis stimulator for most types of human cancer (34) . Our data suggest that VEGF is also the major angiogenic factor in NB cells because the angiogenic effect of NB CM was completely blocked by anti-VEGF, but not by anti-bFGF. These results are in agreement with the observations of Ribatti et al. (25) , who demonstrated the in vivo angiogenic capacity of CM from two NB cell lines, which could also be blocked by a neutralizing anti-VEGF antibody. However, these results need to be confirmed in a larger panel of NB cell lines and tumors.
Further Evidence for the Involvement of Tyrosine Kinase Receptors in Angiogenesis.
Results obtained in other cell systems also are consistent with our data. Stable expression of TrkA decreases the tumorigenicity and invasiveness of highly tumorigenic C6 glioma cells independent of inducing differentiation (35) . Also, Trk receptors are involved in medullary thyroid carcinoma progression and angiogenesis (36) . Finally, neutralizing antibodies against epidermal growth factor and erbB-2/neu receptor tyrosine kinases down-regulate VEGF production by tumor cells in vivo and in vitro (37) . All three independent studies demonstrate the involvement of tyrosine kinase receptors in tumorigenicity and invasiveness of different human tumor cell lines.
Conclusions.
Previous studies showed that TrkA expression and TrkB expression were useful prognostic factors and had an effect on differentiation, proliferation, and survival of NB. This study provides the first in vivo evidence that expression of TrkA and TrkB in NBs growing in a mouse xenograft model also has a profound effect on tumor invasiveness by influencing angiogenesis in vivo. The exact mechanism by which TrkA down-regulates angiogenic factors remains unknown. Additional studies are needed to identify the signal transduction pathways that mediate the inhibitory angiogenic effect and to determine whether inhibition of angiogenesis with more specific tyrosine kinase agonists or antagonists can provide new treatment options in NB.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 Supported by grants from the Deutsche Krebshilfe and the Kind-Philipp Stiftung (to A. E.), the Wolfson Young Investigator Fund (A. E.), the Jeffrey Miller Neurooncology Research Fund (M. A. G.), NIH Grant NS 34514 (to G. M. B.), and the Audrey E. Evans Endowed Chair (G. M. B.).
↵2 To whom requests for reprints should be addressed, at Division of Oncology, ARC, Suite 902D, The Children’s Hospital of Philadelphia, 3516 Civic Center Boulevard, Philadelphia, PA 19104-4318.
↵3 The abbreviations used are: VEGF, vascular endothelial growth factor; NB, neuroblastoma; bFGF, basic fibroblast growth factor; Ang, angiopoietin; TGF-α, transforming growth factor α; RT-PCR, reverse transcription-PCR; HUVEC, human umbilical vein endothelial cell; CM, conditioned medium (media); NGF, nerve growth factor; EC, endothelial cell; GSI, Griffonia simplicifolia I.
- Received September 7, 2000.
- Accepted January 18, 2002.
- ©2002 American Association for Cancer Research.
References
Volume 62, Issue 6
