Interfering with multimerization of netrin-1 receptors triggers tumor cell death

Abstract

Netrin-1 was recently proposed to control tumorigenesis by inhibiting apoptosis induced by the dependence receptors DCC (Deleted in colorectal cancer) and UNC5H. Although the loss of these dependence receptors' expression has been described as a selective advantage for tumor growth and progression in numerous cancers, recent observations have shown that some tumors may use an alternative strategy to block dependence receptor-induced programmed cell death: the autocrine expression of netrin-1. This alternative strategy has been observed in a large fraction of aggressive breast cancers, neuroblastoma, pancreatic adenocarcinoma, and lung cancer. This observation is of potential interest regarding future targeted therapy, as in such cases interfering with the ability of netrin-1 to inhibit DCC or UNC5H-induced cell death is associated with apoptosis of netrin-1-expressing tumor cells in vitro, and with inhibition of tumor growth or metastasis in different animal tumor models. The understanding of the mechanism by which netrin-1 inhibits cell death is therefore of interest. Here, we show that netrin-1 triggers the multimerization of both DCC and UNC5H2 receptors, and that multimerization of the intracellular domain of DCC and UNC5H2 is the critical step to inhibit the proapoptotic effects of both of these receptors. Taking advantage of this property, we utilized a recombinant specific domain of DCC that (i) interacts with netrin-1 and (ii) inhibits netrin-1-induced multimerization, to trigger apoptosis in netrin-dependent tumor cells.

Figure 1. Netrin-1 mediates DCC and UNC5H2 multimerization

(A) DCC multimerization in the presence of netrin-1 in HEK293T cells. Lysates of HEK293T cells transiently transfected with HA-DCC and/or Myc-DCC expressing constructs together plus recombinant netrin-1 (300 ng/ml) were subjected to Myc pull-down (IP αMyc). DCC-HA presence was revealed with an anti-HA antibody. As control DCC surface expression in the presence or absence of netrin-1 was analyzed by FACs analysis and showed no significant change (not shown). (B) UNC5H2 dimerization in the presence of netrin-1 in HEK293T cells. Cell transfection and cell lysate preparation were done as in (A) but with HA-UNC5H2 and/or FlagM2-UNC5H2 expressing constructs. Cell lysates were subjected to HA pull-down (IP αHA). FlagM2-UNC5H2 presence was revealed with an anti-HA antibody. Total: Western blot on lysate before pull-down.

Figure 2. Forced DCC dimerization blocks DCC pro-apoptotic activity

(A) Schematic representation of Fv2e-UNC5H2 fusion constructions showing the two constructs (one tagged HA, the other one tagged Myc) used to validate the artificial dimerization system. (B) Lysates of HEK293T cells transiently transfected with Fv2E-UNC5H2 tagged HA or Myc with or without the dimerization drug (AP20187, 10 nM) were subjected to Myc pull-down (IP αMyc). Total: Western blot on lysate before pull-down. HA-Fv2E-UNC5H2 presence was revealed with anti-HA antibody. (C) DCC- induced cell death is inhibited by dimerization induced by AP20187, as measured by trypan blue exclusion. HEK293T cells were transfected with mock plasmid (Cont.), Fv2E (Fv), Fv2E-DCC-IC (Fv-DCC) with or without 10 nM AP20187 (AP). In all conditions, cells were also transfected with the surface marker pK^k. Transfected cells expressing the marker were magnetically labeled with MACSelect Microbeads and separated using a MACS Separator and Separation Columns. Trypan blue exclusion was assayed on these purified cells. A western blot was done using anti HA antibody. (D) UNC5H2- induced cell death is inhibited by dimerization induced by AP20187, as measured by trypan blue exclusion as in (C). Cells were transfected with pMACSKk and Fv2E (Fv), Fv2E-UNC5H2-IC (Fv-UNC5H2) with or without AP20187 (AP). A western blot was done using anti HA antibody. (E) UNC5H2-induced caspase activation is inhibited by dimerization induced by AP20187, as measured by relative caspase-3 activity. HEK293T cells were transfected with mock vector pCMV (Cont.), Fv2E (Fv), Fv2E-UNC5H2- IC (Fv-UNC5H2) with or without 10 nM AP20187 (AP). Index of relative caspase activity is presented as the ratio between the caspase activity of the sample and that measured in HEK293T cells transfected with pCMV. Standard deviations are indicated (n=3).

Figure 3. The recombinant soluble fifth fibronectin domain of DCC (DCC-5Fbn) inhibits netrin-1-induced DCC-multimerization

(A) Scheme showing DCC-5Fbn as one of the six fibronectin domain (Fn) of DCC. Affinity curve of netrin-1 on DCC-5Fbn measured by ELISA test shows that DCC-5Fbn is able to bind netrin-1. DCC-5Fbn (100 ng) or IL3-R (600 ng) was coated and increasing doses of netrin-1 were added (0 to 800 ng). The IL3 values were substracted from the DCC-5Fbn values. The approximate Kd of DCC-5Fbn/netrin-1 was estimated at 5 nM. (B) Competition assay. As in (A) but the complete extracellular domain of DCC (DCC-EC, 150 ng) was coated instead of DCC-5Fbn and netrin-1 (net) was added (50 ng) in the presence of increasing concentrations of either DCC-5Fbn or the complete DCC-EC. Note that DCC-5Fbn failed to compete with DCC/netrin-1 interaction. I+II (primary and secondary antibody), DCC-5Fbn 2 mol, DCC-EC 2 mol: controls performed in the absence of netrin-1. (C) Netrin-1-induced DCC multimerization is inhibited by DCC-5Fbn. Lysates of HEK293T cells transiently transfected with HA-DCC and/or Myc- DCC and/or netrin-1 (1/3 of total DNA) expressing constructs with or without DCC-5Fbn (900 ng/ml) were subjected to HA pull-down (IP αHA). Myc-DCC presence was revealed with anti-Myc antibody. Anti-HA immunoblot was also performed to detect HA-DCC precipitate. Total: Western blot on lysate before pull-down.

Figure 4. DCC-5Fbn triggers cell death via disruption of netrin-1-mediated inhibition of DCC/UNC5H2 pro-apoptotic activity

(A) HEK293T cells were transiently transfected with a mock (Cont.) or a full-length DCC construct (DCC) and incubated with or without netrin-1 (net) (150 ng/ml) and/or DCC- 5Fbn (800 ng/ml) or the complete extracellular domain of DCC (DCC-EC) (200 ng/ml). Cell death was assessed by trypan blue staining. Inset: DCC immunoblot showing DCC expression level in the different tested conditions. (B) HEK293T cells were transiently transfected with a mock (Cont.) or a full length UNC5H2 construct (UNC5H2) and incubated with or without netrin-1 (net) (500 ng/ml) and/or DCC-5Fbn (1000 ng/ml) or the complete extracellular domain of DCC (DCC-EC) (200 ng/ml). Cell death was measured by fluorescent TUNEL staining. Hoechst staining shows cell nuclei. For each condition, at least 5 randomly chosen fields were analyzed and photographed under epifluorescence microscopy. A representative field is presented from 3 independent experiments. Superimposed photographs (Merge) are shown. Scale bar represents 100 μM.

Figure 5. DCC-5Fbn triggers HCT116 cell death in vitro and inhibits HCT116 cell metastasis in a chicken model

(A) Quantification of netrin-1 and netrin-1 receptors. Ratio of netrin-1 and netrin-1 receptors expression to HPRT housekeeping gene is presented. (B) Representative netrin-1 immunohistochemistry on HCT116 tumor cell line. Inset: Control without primary antibody is presented. (C) Quantitative analysis of cell death in HCT116 cell line treated with DCC-5Fbn, with or without addition of netrin-1 in excess to reverse the effect of DCC-5Fbn. Cell death was quantified by trypan blue exclusion assay (below panel) while apoptosis was monitored by measuring relative caspase-3 activity (upper panel). (D) HCT116 cells were grafted in chick chorioallantoic membrane at day 10 and DCC- 5Fbn or PBS was injected on day 10 and day 13. Tumours and lungs were harvested on day 17. Percentage of embryos with lungs invaded by human HCT116 cells after two injections (day 10 and day 13) of either DCC-5Fbn or PBS. * indicates a p<0,05 calculated using a χ-squared test.