Netrin-1 acts as a survival factor for aggressive neuroblastoma

Abstract

Neuroblastoma (NB), the most frequent solid tumor of early childhood, is diagnosed as a disseminated disease in >60% of cases, and several lines of evidence support the resistance to apoptosis as a prerequisite for NB progression. We show that autocrine production of netrin-1, a multifunctional laminin-related molecule, conveys a selective advantage in tumor growth and dissemination in aggressive NB, as it blocks the proapoptotic activity of the UNC5H netrin-1 dependence receptors. We show that such netrin-1 up-regulation is a potential marker for poor prognosis in stage 4S and, more generally, in NB stage 4 diagnosed infants. Moreover, we propose that interference with the netrin-1 autocrine loop in malignant neuroblasts could represent an alternative therapeutic strategy, as disruption of this loop triggers in vitro NB cell death and inhibits NB metastasis in avian and mouse models.

Figure 1.

Netrin-1 up-regulation is detected in aggressive NB. (A) Netrin-1 mRNA levels in 102 stage 4S (n = 24), [1yr ⁻] stage 4 (n = 12), and [1yr⁺] stage 4 (n= 66) NBs measured by Q-RT-PCR. HPRT housekeeping gene was used as a control. Mean netrin-1 mRNA expression value for each subgroup is indicated by an “m” value. Mean netrin-1 mRNA levels in stage 4S and [1yr⁻] stage 4 were, respectively, compared with the mean netrin-1 detected in [1yr⁺] stage 4. The data were statistically determined using Student's t test compared with levels of [1yr⁺] stage 4. *, P < 0.05; **, P < 0.01. Each sample was assessed in two independent experiments. (B) Representative netrin-1 immunohistochemistry on one [1yr⁺] stage 4 and one stage 4S tumor. Insets show control without primary antibody. Bars, 50 µm. T, tumor cells; S, stromal cells. Netrin-1 antibody specificity is further shown in Fig. 2 D and Fig. S1 B. Immunohistochemistry was performed on four [1yr⁺] stage 4 and four stage 4S tumors. (C) Mean UNC5H mRNA levels in the different stage 4 NBs. Q-RT-PCR using UNC5H1–4-specific primers was performed. Mean UNC5H1–4 mRNA levels in [1yr⁻] stage 4 and [1yr⁺] stage 4 were, respectively, compared with the mean UNC5H1–4 levels detected in stage 4S. Error bars indicate SEM. The data were statistically determined using Student's t test compared with levels of stage 4S. *, P < 0.05; **, P < 0.01. Samples were analyzed in duplicates for each gene. (D) Representative UNC5H1 and UNC5H4 immunohistochemistries on [1yr⁺] stage 4 and stage 4S tumors. Insets show control without primary antibody. Bars, 50 µm. Immunohistochemistry was performed on four stage 4 [1yr⁺] and four stage 4S tumors. (E) Netrin-1 up-regulation is a marker of poor prognosis in stage 4S NB. Overall survival in a panel of 24 infants diagnosed with stage 4S NB with primary tumors showing either netrin-1–low (gray) or netrin-1–high (black) levels. The data was statistically determined using the Kaplan-Meier method. P-value is indicated. (F) Netrin-1 up-regulation is a marker of poor prognosis in infants with NB. Data are presented as in E, with a panel of 30 infants bearing NB.

Figure 2.

Netrin-1 up-regulation is detected in NB cell lines. (A) Netrin-1 expression measured by Q-RT-PCR in a panel of 28 NB cell lines. HPRT housekeeping gene was used as a control. The netrin-1 level is indicated as follows: −, not detectable; −/+, barely detectable; + to ++++, moderate to very high expression. Mouse cell lines are in italics. Cell lines outlined and highlighted in grey are, respectively, netrin-1–low and netrin-1–high cell lines further used in the experiments. (B) Netrin-1 receptor expression in IMR32 and CLB-Ge2 cell lines. DCC/UNC5H Q-RT-PCR was performed on netrin-1–expressing (IMR32) or netrin-1–low (CLB-Ge2) cells using specific primers. Ratio of netrin-1 and netrin-1 receptor expression to the HPRT housekeeping gene is presented. (C) Confocal analysis of UNC5H1, UNC5H3, and UNC5H4 receptor immunostaining in human IMR32 cells. Left and right correspond to IMR32 cells transfected with scramble siRNA and specific siRNA UNC5H, respectively. A fluorescence intensity profile corresponding to the white dashed bar is presented under each panel. Bars, 10 µm. (D) Immunostaining on human IMR32 and CLB-Ge2 cell lines using netrin-1 antibody. Bottom insets show control without primary antibody. Top inset: antibody specificity was tested by adding human recombinant netrin-1. Bars, 50 µm. (E) Confocal analysis of netrin-1 immunostaining on IMR32 cells. A fluorescence intensity profile corresponding to the white dashed bar is presented below. Bar, 5 µm. (F) Quantification of netrin-1 protein secreted in IMR32 and CLB-Ge2 cells conditioned medium by sandwich ELISA assay. Quantification in ng/ml was made according to a dose curve done with recombinant human netrin-1. Data are means of three independent experiments. Error bars indicate SEM. *, P < 0.05 using a two-sided Mann-Whitney test compared with level in IMR32 cells. (G) Quantification of NTN1 and MYCN genomic DNA compared with control NAGK genomic DNA by PCR, using genomic DNA specific primers for each gene, in IMR32 and CLB-Ge2 cells. (H) Quantification of NTN1 promoter activity in IMR32 and CLB-Ge2 cells. Both cell lines were transfected with the vector pGL3-NetP-Luc encoding luciferase under the control of NTN1 promoter. Data presented are normalized on luciferase activity in cells transfected with pGL3 empty vector. Data are means of four independent experiments. Error bars indicate SEM. *, P < 0.05 using a two-sided Mann-Whitney test compared with levels in IMR32 cells.

Figure 3.

Down-regulation of netrin-1 autocrine loop by siRNA triggers NB tumor cell death. (A) Analysis of netrin-1 expression using Q-RT-PCR in nontransfected (control) IMR32 cell line or 24 h after transfection with scramble siRNA (siRNA scr.) or with netrin-1 siRNA (siRNA net.). Data are means of three independent experiments. Error bars indicate SEM. *, P < 0.05 using a two-sided Mann-Whitney test compared with levels in control. (B) Immunostaining on IMR32 cell line using netrin-1 antibody in the absence of transfection (control) or 24 h after transfection with scramble siRNA or netrin-1 siRNA. Note that the general caspase inhibitor z-VAD-fmk was added to avoid cell death induced by netrin-1 siRNA. Insets show control without primary antibody. Bars, 50 µm. (C and D) Cell death induction in IMR32 and CLB-Ge2 cell lines was quantified in nontransfected cells (control) or after transfection with either scramble siRNA or netrin-1 siRNA using trypan blue exclusion assay (C) or relative caspase-3 activity assay (D). Data are means of four independent experiments. In C and D, error bars indicate SEM. *, P < 0.05 calculated using a two-sided Mann-Whitney test compared with level of control.

Figure 4.

Disruption of netrin-1 autocrine loop by a decoy receptor fragment triggers NB tumor cell death. (A) Scheme representing netrin-1 and its receptors DCC and UNC5H and the fifth fibronectin type III domain of DCC (DCC-5Fbn) used to induce cell death. The downstream effector DAPK implicated in UNC5H-induced cell death is also represented. (B–D) Quantitative analysis of cell death in IMR32 and CLB-Ge2 cell lines treated with 1 µg/ml DCC-5Fbn, with or without addition of netrin-1 in excess (150 ng/ml) to reverse the effect of DCC-5Fbn. A negative control was also performed by adding an unrelated IL3R peptide produced in the same condition as DCC-5Fbn. Cell death was quantified by trypan blue exclusion assay (B) while apoptosis was monitored by measuring relative caspase-3 activity (C) or by TUNEL staining (D). Bars, 100 µm. In D, TUNEL staining was performed on three independent experiments. (E) Effect of DCC-5Fbn on fresh [1yr⁺] stage 4 NB. Tumoral cells were directly resuspended from the surgical punction and were plated for 24 h in presence of treatment. In B and C, data are means of six independent experiments. In E, data are means of two independent experiments. Error bars indicate SEM. **, P < 0.01 calculated using a two-sided Mann-Whitney test compared with level of control.

Figure 5.

NB tumor cell death occurs via UNC5H/DAPK proapoptotic signaling. (A) Quantification of cell death in IMR32 cells transfected with either a dominant-negative mutant for DCC proapoptotic activity (DN-DCC) or a dominant-negative mutant for UNC5H proapoptotic activity (DN-UNC5H). Top: DN-DCC and DN-UNC5H proteins expression were analyzed by Western blot. Bottom: cell death was quantified by measuring relative caspase-3 activity after scramble or netrin-1 siRNA transfection. SEM are indicated. Data are means of three independent experiments. *, P < 0.05 calculated using a two-sided Mann-Whitney test compared with level of control. (B) Quantification of cell death in IMR32 cells transfected with either a scramble siRNA or a netrin-1 siRNA together or not with a DCC siRNA. Top: DCC siRNA efficiency was verified by Western blotting on HEK293T cells transfected with pCR-hDCC together with scramble or DCC siRNA. Bottom: cell death was quantified by measuring relative caspase-3 activity. Data are means of three independent experiments. SEM are indicated. *, P <0.05 calculated using a two-sided Mann-Whitney test compared with level of control. Similar results were obtained when neogenin or MYCN were down-regulated (Fig. S3 A). (C) Analysis of specificity and efficiency of each UNC5H siRNA by Western blot in HEK293T cells transfected with UNC5H1, H2, H3, or H4 encoding vector together with each UNC5H siRNA. (D and E) Quantification of cell death in IMR32 cells transfected with either a scramble siRNA or a netrin-1 siRNA together with various combinations of UNC5H siRNA, i.e, one UNC5H (D) or two or four UNC5H (E). Apoptosis was monitored by measuring relative caspase-3 activity. The use of combined UNC5H1, UNC5H2, UNC5H3, and UNC5H4 or UNC5H1, UNC5H3, and UNC5H4 siRNAs, leading to the absence of death induced by netrin-1 siRNA, are presented in gray. Data are means of three independent experiments. Error bars indicate SEM. *, P <0.05 calculated using a two-sided Mann-Whitney test compared with level of control. (F and G) Immunodetection of phosphorylated DAPK (P-DAPK) in IMR32 cells either treated with DCC-5Fbn (F) or transfected with netrin-1 siRNA alone or with UNC5H1, H3, and H4 siRNAs (G). In F and G, immunodetection was performed on three independent experiments.

Figure 6.

Disruption of netrin-1 autocrine loop inhibits NB progression and dissemination in a chick model. (A) Schematic representation of the experimental chick model. IMR32 or CLB-Ge2 cells were grafted in CAM at day 10 and DCC-5Fbn or PBS was injected on days 11 and 14. Tumors and lungs were harvested on day 17. (B–D) Effect of DCC-5Fbn on primary tumor growth and apoptosis. (B) Representative images of IMR32 (top) or CLB-Ge2 (bottom) primary tumors formed on CAM treated either with DCC-5Fbn (right) or PBS (left). Bars, 2 mm. (C) Quantitative analysis showing the mean primary tumor size. (D) Caspase-3 activity was determined in the primary tumor lysates from DCC-5Fbn/PBS-treated IMR32 or CLB-Ge2–grafted embryos. (E) Effect of DCC-5Fbn on lung metastasis. Percentage of embryos with lungs invaded by human IMR32 or CLB-Ge2 cells after two injections (days 11 and 14) of either DCC-5Fbn or PBS was performed as described in the Materials and methods. (F) Effect of DCC-5Fbn on lung metastasis regression. Quantification of lung metastasis in embryos CAM grafted with IMR32 cells and treated after metastasis formation (days 14 and 15) with DCC-5Fn or PBS. The number of embryos studied in each condition is indicated above the graphs and results are from three independent experiments. In C–F, errors bars indicate SEM. C–F: *, P < 0.05; **, P < 0.005 calculated using a two-sided Mann-Whitney test compared with level of control. E: **, P < 0.005 calculated using a Chi-squared test.

Figure 7.

Disruption of netrin-1 autocrine loop inhibits NB dissemination in a mouse model. (A) Analysis of netrin-1 expression using Q-RT-PCR in IGR-N-91 cell line and the IGR-N-91–derived cell lines PTX, BM, Blood, and Myoc. Note that although PTX cells fail to express netrin-1, netrin-1 mRNA is highly expressed in Myoc cells. (B) Immunostaining on IGR-N-91 cells and the different derived cell lines PTX, BM, Blood, or Myoc using netrin-1 antibody. Bar, 50 µm. (C) Cell death was quantified in IGR-N-91, PTX, BM, Blood, or Myoc cell lines treated or not with DCC-5Fbn, with or without addition of netrin-1 in excess to reverse the effect of DCC-5Fbn. A negative control was also performed by adding an unrelated IL3R peptide produced in the same condition as DCC-5Fbn. Data are means of three independent experiments. Errors bars indicate SEM. *, P < 0.05, calculated using a two-sided Mann-Whitney test compared with level of control. (D) Quantification of lung colonization in PTX or Myoc cells in i.v injected mice treated with PBS or DCC-5Fbn for 22 d. Quantification was performed as described in the Materials and methods. Large bars indicate the median values for both groups. P-value was calculated using a two-sided Mann-Whitney test compared with level of control.