Neuroblastoma Progression Correlates with Downregulation of the Lymphangiogenesis Inhibitor sVEGFR-2

A key step in the malignant progression of tumors in adults is the angiogenic switch, the induction of blood vessels by hypoxic tumors (1, 2). Additionally, tumor-induced lymphangiogenesis correlates frequently with the dissemination of tumor cells via lymphatic vessels. Hemangiogenesis and lymphangiogenesis are robustly regulated by members of the vascular endothelial growth factor (VEGF) family: VEGF-A being (predominantly) hemangiogenic, and VEGF-C and -D are lymphangiogenic (3–5). In adults, hemangiogenesis and lymphangiogenesis are associated with progressed tumor stages. Tumors of infants present a much broader spectrum of heterogeneity than those of adults, especially exemplified by the biology of neuroblastoma.

Neuroblastoma is derived from sympatho-adrenal progenitor cells that migrate from the neural crest into target regions of the embryo. Neuroblastoma is mostly located along the sympathetic trunk ganglia and in the adrenal medulla. The spectrum of disease ranges from complete spontaneous regression, which can be immediately observed in the “special” neuroblastoma stage IVs—but may as well occur in early stages—to malignant progression into stage IV, with 5-year survival rates of <30%. Partial differentiation into ganglioneuroblastma is another developmental pathway (6). The most critical molecular predictor for the behavior and treatment of neuroblastoma is the MYCN proto-oncogene. Amplification (up to 150×) of MYCN characterizes highly aggressive tumors and poor outcome despite intensive treatment (7). Staging of neuroblastoma is performed according to the International Neuroblastoma Staging System. Stage I and II neuroblastomas are localized tumors, which have grown across the midline in stage II. Metastasis to regional and systemic lymph nodes characterizes stages III and IV, respectively (8), indicating active interactions with the lymphovascular system. Additionally, high vascularity is characteristic for the progressed tumor stages (9, 10), indicating an influence of blood capillaries on neuroblastoma cell behavior and their typical dissemination into the bone marrow.

The effect of VEGFs on tumor hemangiogenesis and lymphangiogenesis has been shown in numerous studies of adult tumors, and several investigations in recent years have postulated a similar function for VEGFs in neuroblastoma. Blood vessels and lymphatics are present in neuroblastoma, and expression of the ligands VEGF-A, -C, and -D as well as their receptors has been found (11–13). Increased expression of VEGF-A has been described in neuroblastoma stages III and IV (14), whereas VEGF-C has been identified as a risk factor in stage IV neuroblastoma (15). High levels of VEGF-A have been found in stage IV neuroblastoma, but neither VEGF-A nor VEGF-C correlated with age, MYCN copy number or lymph node metastasis (16). Moreover, it has been observed that MYCN amplification correlates strongly with dense vascular supply, tumor dissemination, and poor survival (9). According to Kang et al. (17), MYCN upregulates VEGF-A and MYCN-amplified neuroblastomas exert changes in the vascular pattern of the chick chorioallantoic membrane (18). Some authors have suggested anti–VEGF-A treatment with bevacizumab for high-risk neuroblastoma (19). There are, however, results that challenge the unequivocal functions of VEGFs for the progression of neuroblastoma. Vessel density was not predictive of survival in a cohort of patients with neuroblastoma (20). Anti–VEGF-A treatment did not result in any reduction of experimental neuroblastoma growth in mice (21, 22), and some authors have emphasized the heterogeneity of angiogenesis stimulators and inhibitors in neuroblastoma (23, 24). We have therefore reinvestigated the expression of VEGFs and their receptors in a cohort of 49 neuroblastoma patients and in 24 neuroblastoma cell lines. Additionally, we studied the effect of MYCN on the expression of VEGFs in primary neuroblastomas and in neuroblastoma cell lines. In contrast to previous studies, we did not observe a positive correlation between tumor progression and the expression of VEGF-A and -C. Of note, we found significant downregulation of the VEGF-C inhibitor sVEGFR-2 in the progressed stages III, IV, and IVs, indicating a positive correlation between lymphangiogenesis and lymph node metastases. Additionally, we observed increased expression of VEGF-A and -D, as well as reduced expression of the inhibitors VEGFR-1 and sVEGFR-2 in MYCN-amplified stage IV neuroblastoma, a finding which was recapitulated by MYCN-transfected SH-EP cells, but not generally observed in MYCN-amplified neuroblastoma cell lines. Our data show that upregulation of the hemangiogenesis and lymphangiogenesis activators VEGF-A and VEGF-D, and downregulation of the hemangiogenesis and lymphangiogenesis inhibitors VEGFR-1 and sVEGFR-2, act in concert during neuroblastoma progression. Cell lines do not necessarily reflect the in vivo situation. In addition to the downregulation of hemangiogenesis inhibitors (25, 26), downregulation of lymphangiogenesis inhibitors is an alternative mechanism for the increased vascularization and metastasis formation of progressed neuroblastoma.

RNA samples of 50 primary, untreated tumors, were kindly provided by the Tumorbank of the German Neuroblastoma Studies Group, Drs. F. Berthold, B. Hero, and J. Theissen, Children's Hospital University of Cologne, Cologne, Germany. Tumor specimens were prepared according to a standard protocol. Two representative areas were dissected out of the tumor and each was divided into four parts. One part was fixed in formalin and three parts were snap-frozen in liquid nitrogen. Snap-frozen specimens were used in our study. RNA was isolated with Trizol (Invitrogen). The samples were tested with Bioanalyzer 2100 (Agilent Technologies). One sample failed the test and was discarded. The remaining 49 samples were allocated to the neuroblastoma stages as follows: stage I (n = 8), stage II (n = 6), stage III (n = 5; 2 were MYCN amplified), stage IV (n = 20; 10 were MYCN amplified), and stage IVs (n = 10; 1 was MYCN amplified).

All 24 human neuroblastoma cells lines (Table 1; see ref. 27 for a well-arranged review of 113 neuroblastoma cell lines) were maintained in a humidified incubator at 37°C and 5% CO₂ atmosphere using RPMI 1640 (Lonza) with 10% fetal bovine serum (Biochrome) and 1% penicillin/streptomycin (Invitrogen). The neuroblastoma cell line SH-EP was stably transfected to overexpress the MYCN oncogene as described previously (28). The transfected cell line was designated WAC2. Transfected cells were continuously selected by adding G418 (100 μg/mL; Invitrogen) to the medium. PC-3 human prostate carcinoma cells were purchased from DSMZ, and used as reference for VEGF-A and VEGF-C ELISAs.

Cells were rinsed twice with PBS and RNA was isolated directly from the culture plate using Trizol (Invitrogen) as recommended by the supplier. Quality of RNA samples was analyzed with NanoDrop spectrophotometer (NanoDrop Products) and ethidium bromide staining on agarose gels.

We prepared cDNA from 2 μg total RNA using Omniscript reverse transcriptase (Qiagen). Real-time PCR was performed with Opticon2 thermal cycler (MJ Research), using SYBR Green JumpStart Taq ReadyMix (Sigma-Aldrich). Primers used are listed in Table 2. All primers were designed to produce fragments, which span exon-intron boundaries, to exclude amplification of genomic DNA. The primers were also designed to detect all known splice variants of the respective gene of interest. For sVEGFR-2, the reverse primer recognizes the intron 13 motif, which is specific for the truncated transcript variant of this secreted form of VEGFR-2 (29). The probe was measured against two different β-actin probes. Both measurements revealed significant downregulation in metastatic neuroblastoma (only one is shown in Fig. 1G).

Sandwich ELISA was performed to measure VEGF-A and VEGF-C protein in cell culture supernatants of neuroblastoma cell lines SH-EP and SH-IN in comparison with the human prostate carcinoma PC-3 cell line, using methods and tools described recently (30, 31). Cells were cultured for 4 d in RPMI 1640. Supernatant was collected and compared with control (day 0) supernatant of PC-3 cells. Experiments were performed twice.

Statistical analyses were performed using SAS Software v. 9.1 (SAS Institute).

We have studied primary neuroblastoma from a cohort of 49 patients by real-time reverse transcription-PCR (RT-PCR) using SYBR Green and the ΔΔCt method for relative quantification of VEGFs and VEGF receptors. None of the patients had received radiotherapy or chemotherapy before tumor resection. Staging of neuroblastoma was performed according to the International Neuroblastoma Staging System and the specimens were allocated to the stages as follows: stage I (n = 8), stage II (n = 6), stage III (n = 5; 2 of which were MYCN amplified), stage IV (n = 20; 10 of which were MYCN amplified), and stage IVs (n = 10; 1 of which was MYCN amplified). We studied expression of VEGF-A, VEGF-C, VEGF-D, VEGFR-1 (FLT1), VEGFR-2 (KDR), and VEGFR-3 (FLT4), as well as a soluble splice variant of VEGFR-2 (sVEGFR-2), which has very recently been shown to act as an endogenous inhibitor of lymphangiogenesis (29). Expression of all VEGFs and their receptors varied greatly. We did not observe a statistically significant difference between locoregional tumors (stages I and II) and metastasized neuroblastoma (stages III, IV, and IVs) for VEGF-A, -C, and -D (Fig. 1A–C). Also, VEGFR-1, -2, and -3 were not significantly regulated (Fig. 1D–F); however, we observed significant downregulation of the lymphangiogenesis inhibitor sVEGFR-2 in metastatic stages III, IV, and IVs (Fig. 1G).

We have then compared the expression levels of VEGFs in MYCN-amplified stage IV neuroblastoma with those that had normal MYCN status (each 10 per group). We did not find statistically significant differences; however, there were tendencies for an increase in the expression of VEGF-A and -D and for downregulation of VEGFR-1 and sVEGFR-2 in MYCN-amplified stage IV tumors (Fig. 2). This suggests increased hemangiogenic and lymphangiogenic potential.

In a further approach, we isolated RNA from 24 human neuroblastoma cell lines and studied the expression of VEGF-A, VEGF-C, VEGF-D, VEGFR-1, VEGFR-2 and VEGFR-3, and sVEGFR-2. Relative expression in neuroblastoma cell line CHLA20 was used as a reference and was set as 1 (Fig. 3). Only a few neuroblastoma cell lines showed robust expression of VEGF-A. We observed 15-fold expression in CHLA90 and almost 40-fold expression in SH-IN (Fig. 3A). To analyze the biological significance of this observation, we inoculated SH-IN (high VEGF-A), GI-MEN and CHP134 (low VEGF-A) on the chorioallantoic membrane of chick embryos. The three cell lines produced solid tumors after 7 days, but only SH-IN were densely vascularized, whereas GI-MEN and CHP134 produced almost avascular tumors (data not shown). VEGF-C was highly expressed in SK-N-SH and SH-EP (Fig. 3B), and VEGF-D was high in IMR5, KCN, NLF, SH-IN, IMR32, and SK-N-AS (Fig. 3C). Expression of VEGF receptors was low, and for VEGFR-3, was almost undetectable (data not shown). We found very weak expression of VEGFR-1 in various neuroblastoma cell lines (Fig. 3D), elevated expression of VEGFR-2 only in NB 69 (Fig. 3E), but considerable expression of sVEGFR-2 in a number of cell lines, most prominently in NLF, LAN6, SK-N-AS, LAN1, and LAN2 (Fig. 3F). For two cell lines, SH-EP and SH-IN, we measured VEGF-A and VEGF-C at protein level with sandwich ELISA (Table 3). The data show that RNA and protein expression correlate very well, e.g., SH-IN, which has the highest VEGF-A mRNA expression, secretes high amounts of VEGF-A (2,005 pg/mL), whereas VEGF-C protein is not measurable (compare Fig. 3A and B). SH-EP secretes VEGF-C (140 pg/mL), which is comparable to the prostate carcinoma cell line PC-3, and has considerably high levels of VEGF-C mRNA.

It has been shown that MYCN inhibition by short interfering RNA blocks VEGF-A secretion in MYCN-amplified IMR-32 cells (17). To test the effects of MYCN, we studied stable overexpression of MYCN in SH-EP, which has regular MYCN expression. The stably transfected cell line was designated WAC2, because the cells form colonies in soft agar (28). For VEGF-C, VEGFR-1, -2 and -3, expression in WAC2 was not different from that of the parental cell line SH-EP (Fig. 3B, D, and E); however, VEGF-A and -D were upregulated, and sVEGFR-2 was significantly downregulated, so that it was no longer detectable in WAC2 (Fig. 3A, C, and F). Notably, these effects may promote hemangiogenesis and lymphangiogenesis.

According to published data, we also divided the 24 cell lines into two groups with normal versus amplified MYCN expression and compared VEGF ligands and receptors (Fig. 4A–F). However, the results did not reflect any of the results measured in primary tumors, and underline that in vitro data have to be interpreted with great caution.

In summary, our data show that there is significant downregulation of the lymphangiogenesis inhibitor sVEGFR-2 in metastatic neuroblastoma as compared with localized neuroblastoma stages I and II. In MYCN-amplified stage IV neuroblastoma, there is a tendency for the upregulation of VEGF-A and -D, and downregulation of VEGFR-1 and sVEGFR-2. Similar results could be observed in MYCN-transfected SH-EP cells. Together, upregulation of hemangiogenesis and lymphangiogenesis activators (VEGF-A and VEGF-D, respectively) and downregulation of hemangiogenesis and lymphangiogenesis inhibitors (VEGFR-1 and sVEGFR-2, respectively) may represent cooperative mechanisms during neuroblastoma progression.

Quite some time ago, the importance of VEGF-A for hemangiogenesis and of VEGF-C and VEGF-D for lymphangiogenesis was revealed (32–39). The essential role of VEGF-A in tumor hemangiogenesis is the basis for its targeting in antiangiogenesis therapy (40). Tumor-induced lymphangiogenesis by VEGF-C and -D, and the positive correlation with the lymphogenic spread of tumor cells had been described several years ago (41–44). Dense vascularization is a key feature of malignant tumor progression (1). This holds true for numerous tumor types in the adult, and has also been observed in tumors of infants, such as neuroblastoma.

A vascular index of neuroblastoma specimens (total number of vessels per mm²) has been measured in a cohort of 50 patients, and it was found that an index of >4 correlates strongly with widely disseminated disease, poor survival, and MYCN amplification (9). A similar correlation with angiogenic (integrin α_vβ₃-positive and α_vβ₅-positive) endothelium has been described by Erdreich-Epstein et al. (45). A number of studies have revealed a positive correlation between neuroblastoma progression and VEGFs, however, most of these studies were performed in vitro or in experimental tumors in nude mice, and there are only few data available on primary tumors. Among these, Pavlakovic et al. (46) have shown that VEGF-A levels are not increased (rather slightly reduced) in the serum of neuroblastoma patients as compared with healthy controls. In contrast, elevated levels of VEGF-A, -B, and -C have been measured by RT-PCR (quantified by densitometric analysis, normalized against glyceraldehyde-3-phosphate dehydrogenase expression in a cohort of 37 patients) in neuroblastoma stages III and IV as compared with stages I, II, and IVs (11). By means of ELISA, VEGF-A protein has been measured in five primary neuroblastoma and revealed values of 150 to 2,400 pg/g total protein with the highest value in a stage II neuroblastoma (13). Fakhari et al. (14) studied the expression of VEGF-A, -B, and -C in 37 patients with neuroblastoma (patients that underwent routine radiologic and medical program before surgery) using real-time RT-PCR and VEGF-A in serum with ELISA. They observed a significant upregulation of VEGF-A and -C in stages III and IV, as compared with adrenal control tissue, and elevated serum VEGF-A levels in stage III only. Additionally, they found upregulation of VEGFR-1 and -2 in stage III neuroblastoma. In summary, the published data do not yet provide a consistent picture of VEGFs in neuroblastoma and we have therefore reinvestigated their expression in a cohort of 49 untreated, primary tumors, and in 24 cell lines using real-time RT-PCR. Thereby we also included the soluble splice variant of VEGFR-2 (sVEGFR-2), an endogenous inhibitor of lymphangiogenesis (29), and studied the effect of MYCN on the respective expression patterns.

In contrast with previous studies, we have not detected elevated expression of VEGF-A and -C in stage III and IV neuroblastoma. In fact, there is no significant difference between localized stages I and II and the metastatic stages III, IV, and IVs with regards to the expression of VEGF-A, -C, and -D. VEGFR-1 is expressed at almost equal amounts in all stages, and VEGFR-2 and -3 are also detectable in all stages. A significant difference between localized and metastatic stages was found for sVEGFR-2, a secreted endogenous inhibitor of lymphangiogenesis (29). This inhibitor is highly expressed in stages I and II, and barely detectable in stages III, IV, and IVs. Because the regulation of lymphangiogenesis by tumors has an influence on their metastatic behavior (3), our data for the first time indicates that downregulation of an inhibitor of lymphangiogenesis may expedite the formation of lymph node metastases. Neuroblastoma stages III, IV, and IVs are characterized by the spread of tumor cells to local, distant, and dermal lymph nodes, respectively. However, we have to be aware that our measurements, performed at the RNA level, do not necessarily reflect the amount of secreted protein. Downregulation of sVEGFR-2 may be enhanced by the MYCN oncogene, which is a critical clinical predictor for poor outcome. In stage IV neuroblastoma, MYCN amplification does not only downregulate sVEGFR-2, but also the hemangiogenesis inhibitor VEGFR-1. At the same time, the hemangiogenesis and lymphangiogenic factors VEGF-A and -D are upregulated. Identical regulation patterns could be observed after MYCN transfection of SH-EP cells. In these cells, we even observed complete switching-off of sVEGFR-2 expression. Our data support the observation that MYCN amplification promotes the progression of neuroblastoma, and it obviously does so by both the upregulation of activators and the downregulation of inhibitors. Downregulation of other inhibitors of hemangiogenesis in neuroblastoma has been described previously (25, 26).

The concept of high vascularity of progressed tumors has been challenged recently, as a number of data have pointed to a more aggressive behavior and the upregulation of genes associated with poor survival in avascular glioblastomas (47) and hypoxic neuroblastoma (48). In mice, the inhibition of tumor angiogenesis either by targeting the VEGF-A or the VEGFR/PDGFR kinase pathways have induced progression to greater malignancy and increased invasiveness, and decreased overall survival (49, 50). However, our observations in general support the classic view, although differences in the expression of angiogenic factors may not become immediately evident by comparing tumor stages. The proportion between proangiogenic and antiangiogenic factors has to be taken into account, as well as the fact that tumor cell lines do not necessarily reflect the in vivo behavior in the human.

No potential conflicts of interest were disclosed.

We thank S. Schwoch, Ch. Zelent, and B. Manshausen for their excellent technical assistance and Drs. Frank Berthold, Barbara Hero, and Jessica Theissen of the German Neuroblastoma Studies Group, the Children's Hospital University Cologne, Cologne, Germany, for providing tumor samples and data. We are grateful to Dr. L. Schweigerer for leaving us the neuroblastoma cell lines.

Grant Support: DFG grant WI1452/11-1 (J.W. Hing); NIH/NEI, Research to Prevent Blindness Senior Scientific Investigator and Unrestricted Awards, Doris Duke Distinguished Clinical Scientist Award, and the Burroughs Wellcome Fund Clinical Scientist Award in Translational Research (J. Ambati).

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.