Fig 2: Proposed mechanisms for the ligand-induced activation of ErbB1, -2, and -3 and a model for the inhibitory mechanism of Necl-4 for the HRG-induced activation of the ErbB2-ErbB3 signalling. (a) Proposed mechanism for the EGF-induced formation of the ErbB1-ErbB1 homo-dimer. (b) Proposed mechanism for the HRG-induced formation of the ErbB2-ErbB3 hetero-dimer. (c) A model for the inhibitory mechanism of Necl-4 for the HRG-induced formation of the ErbB2-ErbB3 hetero-dimer. In the absence of HRG, the third Ig-like domain of Necl-4 interacts with the domain 3 of the ‘tethered’ form of ErbB3. This interaction inhibits the HRG-induced formation of the ErbB2-ErbB3 hetero-dimer. (d) A model for the activation of the ErbB2-ErbB3 signalling in cancer cells. In cancer cells, Necl-4 is down-regulated or dys-functioned and the inhibitory role of Necl-4 in the HRG-induced formation of the ErbB2-ErbB3 hetero-dimer could be lost and induce cell movement and survival.

Fig 3: Inhibition of the HRG-induced activation of ErbB3 and dimerization of ErbB3 with ErbB2 by the third Ig-like domain of Necl-4. (a) Inhibition of the HRG-induced activation of ErbB3 by reFLAG-Necl-4-Ig3. MCF7 cells were starved of serum and stimulated by HRG for 4 min in the presence of 20 μg/ml of the indicated recombinant Necl-4. The samples were subjected to Western blotting using the indicated Abs. Arrowheads, phospho-ErbB3 on Tyr-1289. The band intensities of the phospho-ErbB3 on Tyr-1289 were normalized to that of the total ErbB3 protein. The normalized value of the MCF7 cells treated with HRG alone was set as 1.0 (n = 3). n.s., not significant; **P < 0.01. (b) Inhibition of the HRG-induced dimerization of ErbB3 with ErbB2 by reFLAG-Necl-4-Ig3. HEK293E cells were co-transfected with ErbB3 and ErbB2-HA and treated with 200 μg/ml of reFLAG-Necl-4-Ig3 and stimulated by HRG for 10 min. ErbB3 was immunoprecipitated using the anti-ErbB3 mAb. The samples were subjected to Western blotting using the indicated Abs. Arrowheads, co-immunoprecipitated ErbB2-HA. The band intensities of the co-immunoprecipitated ErbB2-HA with ErbB3 were quantified and the value of the precipitated ErbB2-HA in the lysates treated with HRG alone was set as 1.0 (n = 4). *P < 0.05; **P < 0.01. (c) Inhibition of the HRG-induced dimerization of endogenous ErbB3 with ErbB2 by reFLAG-Necl-4-Ig3. MCF7 cells were treated with or without 200 μg/ml of reFLAG-Necl-4-Ig3 and stimulated by HRG for 10 min. Endogenous ErbB3 was immunoprecipitated using the anti-ErbB3 mAb. The samples were subjected to Western blotting using the indicated Abs. Arrowheads, co-immunoprecipitated ErbB2. The band intensities of the co-immunoprecipitated ErbB2 with ErbB3 were quantified and the value of the precipitated ErbB2 in the lysates treated with HRG alone was set as 1.0 (n = 3). *P < 0.05; **P < 0.01. (d) Interaction of the third Ig-like domain of Necl-4 with ErbB3. HEK293E cells were co-transfected with various combinations of the indicated plasmids. FLAG-Necl-4-Ig3-TM was immunoprecipitated using the anti-FLAG mAb. The samples were subjected to Western blotting using the indicated Abs. Arrowheads, ErbB3-GFP; square brackets, FLAG-Necl-4-Ig3-TM. IB, immunoblotting; IP, immunoprecipitation.

Fig 4: Therapeutic targeting of METN375S tumors using inhibitors of HER2 signaling.a–c Cellular invasion on isogenic wild-type (METwt-tGFP) and N375S mutant (METN375S-tGFP) clones, treated with kinase inhibitors or siRNA silencing of MET or HER2, were being evaluated. a H2170 METN375S-tGFP cells were treated with crizotinib (10 µM), lapatinib (0.3 µM) or a crizotinib/lapatinib combination for 36 h. Representative images of cell invasion are shown. b H2170 METN375S-tGFP cells were co-incubated with 10 nM of the indicated siRNA overnight, and seeded in Matrigel invasion chambers for 36 h. Representative images are shown. Scale bar: 200 μm. c Percentage of invaded cells relative to vehicle control (0.1% DMSO) or Scr siRNA control were expressed as mean ± SD (n = 3). One-way ANOVA; *P < 0.05, **P < 0.01, ***P < 0.001. Left, immunoblots of siRNA-treated H2170 METN375S-tGFP cells. For immunoblots, cells were harvested 48 h after siRNA treatment, and β-actin was used as a loading control. d–f Calu-1 METwt-tGFP and METN375S-tGFP cells were co-incubated with 10 nM of the indicated siRNA overnight, and seeded in Matrigel invasion chambers for 24 h. Representative images of cell invasion for METwt-tGFP (d) and METN375S-tGFP cells (e) are shown. Scale bar: 200 μm. f Percentage of invaded cells relative to Scr siRNA control in Calu-1 METwt-tGFP and METN375S-tGFP cells were expressed as mean ± SD (n = 3). Left, immunoblots of siRNA-treated Calu-1 METN375S-tGFP cells. For immunoblots, cells were harvested 48 h after siRNA treatment, and β-actin was used as a loading control. One-way ANOVA; *P < 0.05, **P < 0.01, ***P < 0.001. g, h Anchorage-independent colony formation on isogenic H2170 METN375S-tGFP cells, treated with kinase inhibitors or siRNA silencing of MET or HER2, were being evaluated. g H2170 METN375S-tGFP cells were co-incubated with 10 nM of the indicated siRNA overnight, and seeded in soft agar for 4 weeks. Representative images are shown. h The number of colonies were quantified after treatment with crizotinib, lapatinib, trastuzumab, or in combination. Data were presented as percentage of colonies relative to vehicle control (0.1% DMSO) ± SD (n = 3). i–l Efficacies of HER2 inhibitors were evaluated in xenograft models. Tumor growth of METwt-tGFP (i) and METN375S-tGFP (j) xenografts after treatment with trastuzumab, pertuzumab, lapatinib, and ASLAN001 were expressed at mean ± SEM (n = 5). Two-way ANOVA; *P < 0.05, **P < 0.01, ***P < 0.001. k Immunohistochemistry staining showing the changes in p-MET after treatment with trastuzumab in METwt-tGFP and METN375S-tGFP tumors. Representative images are shown. Scale bar, 50 µm. l Expression of p-MET was quantified and expressed at mean of positive-staining/100 cells ± SD (n = 5). Two-tailed Student’s t test; *P < 0.05, **P < 0.01, ***P < 0.001. m HEK293 cells were transfected with 1 µg of either pCMV6-ERBB2-wt, ∆D1, ∆D2, ∆D3, ∆D4, or ∆TK plasmid, together with pCMV6-MET-N375S plasmid, for 24 h. MET/HER2 interaction in HEK293 cells was detected with immunoprecipitation and immunoblotting. Left, input controls and phosphorylated proteins. β-Actin was used as a loading control. n, o Isogenic Calu-1 METwt-tGFP (n) and METN375S-tGFP (o) cells were engrafted into SCID mice, and treated daily with vehicle (n = 6) or 15 mg/kg afatinib (n = 5) starting 14 days after inoculation. Arrows indicate treatment start date. Kaplan–Meier analyses of the mice are shown.

Fig 5: The Sema domain of MET is critical for dimerization with HER2.a, b Stable isotope labeling with amino acids in cell culture (SILAC) was performed to identify novel binding partners. METwt-tGFP and METN375S-tGFP cells were labeled with heavy (H) and light (L) amino acids. The cutoff values for SILAC ratios, after normalizing with MET expression, were set at (>2, <0.5) respectively. a Scatter plot of transformed METwt/METN375S ratios of membrane-bound proteins. Both axes represent METwt(H)/METN375S(L) and METwt(L)/METN375S(H) ratios, respectively. b List of various membranous proteins identified in SILAC analysis found to be associated with METN375S. c Interaction of ectopic MET and endogenous HER2 in H2170 METwt-tGFP and METN375S-tGFP cells was detected with immunoprecipitation and immunoblotting. Left, input controls. Total MET and HER2 band intensities, normalized to input controls and relative to METwt, are shown below (n = 5). d, e Detection of MET/HER2 co-localization (red) in EV, METwt-tGFP, and METN375S-tGFP cells with proximity ligation assay (PLA). Representative images are shown (d), with the PLA signals quantified (e) and expressed as the number of signals/cell ± SD (n > 3). Scale bar, 20 µm. f Representative confocal microscopy images of MET (Alexa-488; green) and HER2 (Alexa-594; red) in isogenic H2170 clones (n = 2). DAPI (blue) was used as nuclear counter stain. Co-localized proteins appeared in yellow. The smaller panels are detailed views of the outlined (white) squares in the respective images. Scale bar, 20 µm. The total MET fluorescence that co-localized with HER2 signal in each variant was tabulated on the right, data presented as mean ± SD (five fields in one representative experiment). g MET/HER2 interaction in H2170 METwt-tGFP and METN375S-tGFP tumors shown in Fig. 2g was detected with immunoprecipitation and immunoblotting. IgG was used as a loading control. h–k The role of Sema domain in METN375S-tGFP cells was examined with recombinant Sema proteins, wild-type, rSemawt; N375S mutant, rSemaN375S. Cell viability of METwt-tGFP (h) and METN375S-tGFP (i) cells after treatment with 1, 5, 10 µg/ml of rSemawt or rSemaN375S for 72 h, presented as mean ± SD (n = 3). Two-tailed Student’s t test; *P < 0.05, **P < 0.01, ***P < 0.001. j Immunoblots showing the total and phosphorylated expressions of MET, HER2, Src, Akt, and ERK1/2 in lysates of the indicated cell lines after treatment with 10 µg/ml rSema. β-Actin was used as a loading control. k MET/HER2 interaction in H2170 METN375S-tGFP cells was detected with immunoprecipitation and immunoblotting after treatment with 10 µg/ml rSema. Left, input controls. l MET/HER2 interaction in H2170 METN375S-tGFP cells was detected after serum starvation (0.1% FBS), or co-incubated with HGF (0.1% FBS + 10 ng/ml HGF). Left, input controls. m HEK293 cells were transfected with 1 µg of either pCMV6-EV-tGFP vector, MET-wt, N375S, ∆Sema, N375Q, M1268T, or Y1248H plasmid, together with pCMV6-ERBB2-DDK plasmid, for 24 h. MET/HER2 interaction in HEK293 cells was detected with immunoprecipitation and immunoblotting. Total MET and HER2 band intensities, relative to METwt, are shown below (n = 3). Left, input controls and phosphorylated proteins. β-Actin was used as a loading control.