TOM1L1 drives membrane delivery of MT1-MMP to promote ERBB2-induced breast cancer cell invasion - PubMed (original) (raw)

doi: 10.1038/ncomms10765.

Guillaume Collin 1, Simon Descamps 1, Heiani Touaitahuata 1, Valérie Simon 1, Nicolas Reymond 1, Laurent Fernandez 2, Pierre-Emmanuel Milhiet 2, Virginie Georget 3, Serge Urbach 4, Laurence Lasorsa 5, Béatrice Orsetti 5, Florence Boissière-Michot 6, Evelyne Lopez-Crapez 6, Charles Theillet 5, Serge Roche 1, Christine Benistant 1 2

Affiliations

TOM1L1 drives membrane delivery of MT1-MMP to promote ERBB2-induced breast cancer cell invasion

Clément Chevalier et al. Nat Commun. 2016.

Abstract

ERBB2 overexpression in human breast cancer leads to invasive carcinoma but the mechanism is not clearly understood. Here we report that TOM1L1 is co-amplified with ERBB2 and defines a subgroup of HER2(+)/ER(+) tumours with early metastatic relapse. TOM1L1 encodes a GAT domain-containing trafficking protein and is a SRC substrate that negatively regulates tyrosine kinase signalling. We demonstrate that TOM1L1 upregulation enhances the invasiveness of ERBB2-transformed cells. This pro-tumoural function does not involve SRC, but implicates membrane-bound membrane-type 1 MMP (MT1-MMP)-dependent activation of invadopodia, membrane protrusions specialized in extracellular matrix degradation. Mechanistically, ERBB2 elicits the indirect phosphorylation of TOM1L1 on Ser321. The phosphorylation event promotes GAT-dependent association of TOM1L1 with the sorting protein TOLLIP and trafficking of the metalloprotease MT1-MMP from endocytic compartments to invadopodia for tumour cell invasion. Collectively, these results show that TOM1L1 is an important element of an ERBB2-driven proteolytic invasive programme and that TOM1L1 amplification potentially enhances the metastatic progression of ERBB2-positive breast cancers.

PubMed Disclaimer

Figures

Figure 1

Figure 1. TOM1L1 is co-expressed with ERBB2 in breast cancer.

(a) The ERBB2 and TOM1L1 loci are frequently co-amplified in breast cancer. Upper part: cumulative array-CGH copy number frequency plots of chromosome 17 in breast tumours with ERBB2 amplification. Copy number gains are in blue, losses in red. Tumours with a log2 ratio higher than the threshold of 0.25 were considered as amplified. Lower part: copy number profiles at chromosome 17 in individual tumours (each line is one tumour). The colour code is the same as in the upper part. About 52.5% of ERBB2-positive tumours show TOM1L1 co-amplification. (b) Representative images showing TOM1L1 protein expression in breast cancer samples with absence (A), weak (B), moderate (C) and strong (D) immunoreactivity. Scale bar, 10 μm. (c) TOM1L1 expression was significantly associated with ER and ERBB2 expression. Histograms shows the results of screening by IHC of TOM1L1 expression in 108 breast tumours encompassing the four major subtypes (ER−, ER+, ERBB2− and ERBB2+). Statistical analysis was done with Student's _t_-test on raw data. **P<0.01 and ***P<0.001. (d) TOM1L1 is overexpressed in ERBB2+ cell lines. Western blot analyses of TOM1L1, ERBB2 and tubulin (used as loading control) expression in 12 breast cancer cell lines. (e) Kaplan–Meier survival analysis of patients with ERBB2+/ER+ breast cancer from the database of ref. . Patients were stratified based on TOM1L1 amplification (red curve) and were compared with patients with no amplification (green curve). The log-rank test was used to determine the statistical significance (P value≤0.05).

Figure 2

Figure 2. TOM1L1 regulates ERBB2-driven cell invasion.

(a) TOM1L1 expression was assessed by western blotting in SKBR3 cells infected with empty retroviruses (mock) or coding TOM1L1. (b) SKBR3 cells infected as in a were seeded in the upper compartment of a Boyden chamber containing 1 mg ml−1 matrigel. After 48 h, cells present in the lower chamber were visualized by Hoechst staining. (c) Quantification of b. The histogram shows the invasion ratio normalized to control conditions (_n_=3). (d) SKBR3 cells infected as indicated were embedded as spheroids in a collagen and matrigel matrix for 75 h and then imaged by phase contrast microscopy. Yellow dotted line shows the sphere size at _t_=0 and the red dotted line shows invasive cells or invasive fronts at the end of the experiment. Invasive fronts were twice larger in TOM1L1-expressing cells than in controls cells (_n_=4). Scale bar, 100 μm. (e) TOM1L1 expression was assessed by western blotting in BT-474 cells infected with viruses containing control shRNA (Ctrl) or two different TOM1L1 shRNAs (TOM1L1 #1 and #2). (f) Invasion of the BT-474 infected cells was assessed in Boyden chambers with matrigel as in b. Cells in the lower chamber were stained with Hoechst and counted as in c (_n_=3) (g,h) Infected BT-474 cells were embedded as spheroids in a collagen and matrigel matrix for 5 days and imaged by phase contrast microscopy. Yellow dotted line shows sphere size at _t_=0, green dotted line shows sphere size at _t_=120 h and red dotted line shows invasive cells or invasive fronts. The surface explored by the cells was decreased by 1.5-fold in TOM1L1 silenced cells compared with controls (_n_=3). Scale bars, 400 μm. (i) SKBR3 cells infected as indicated were seeded in Boyden chambers as in b,c and incubated with DMSO, 12.5 μM GM6001 (MMP inhibitor), 1 μM lapatinb (an EGFR/ERBB2 inhibitor) and 2 μM SU6656 (SRC inhibitor) for 48 h. The histogram shows the invasion ratio normalized to control conditions (_n_=10). All histograms in Fig. 2 show mean±s.e.m. *_P_≤0.05 (Student's _t_-test).

Figure 3

Figure 3. TOM1L1 pro-invasive activity requires ERBB2-induced interaction of its GAT domain with TOLLIP.

(a) Schematic showing TOM1L1 and the different mutants used in the study. (b) Upper panel: invasion of SKBR3 cells infected with the indicated viruses was assessed in Boyden chambers with matrigel. The histogram shows the invasion ratio normalized to the control value after 48 h of invasion (_n_=4). Mean±s.e.m *P<0.05, **P<0.01, NS, no significant (Student's _t_-test). Lower panel: western blots showing the expression of the various TOM1L1 mutants in SKBR3 cells. (c) Lysates from BT-474, 3T3-SRCY527F and SKBR3 cell lines expressing the indicated constructs were immunoprecipitated with an anti-TOM1L1 antibody and blotted as indicated. Note that only the 3T3-SRCY527F cell lysate shows a band corresponding to p-tyr above the Ig band. (d,e) In vivo metastasis assay: luciferase-expressing HCC-1954 cells infected as indicated were injected in mice hearts (left ventricle). At the indicated times, mice were injected intraperitoneally with luciferine and imaged. (d) Mice imaging show higher bioluminescence signal in the brain of mice injected with TOM1L1-expressing cells compared with mock or ΔGAT-expressing cells (day 21 post injection). Right panel shows representative images of mice brains in each condition. (e) Quantification in time of the number of mice exhibiting bioluminescence signals in brain or legs. Note that bioluminescence signals are detected earlier and in more mice (day 15) when injected with cells that express TOM1L1 than with control cells (mock) or the ΔGAT mutant. (f) Western blot analysis of TOLLIP expression in 12 breast cancer cell lines. Note that TOLLIP is preferentially expressed in ERBB2+ cell lines. Lysates used were the same as in Fig. 1d. (g) Upper panel: invasion of SKBR3 cells infected as indicated and transfected twice with control (−) or TOLLIP siRNAs (+) was assessed and quantified as in b (_n_=5). Mean±s.e.m *P<0.05 (Student's _t_-test). Lower panel: the efficiency of TOLLIP silencing was confirmed by western blot analysis of TOLLIP expression in the transfected cells. (h) Whole cell lysates of SKBR3 cells infected with indicated viruses were immunoprecipitated with an anti-TOLLIP antibody and immunoblotted as shown.

Figure 4

Figure 4. TOM1L1 regulates invadopodia and requires MT1-MMP.

(a) BT-474 cells infected with the indicated shRNAs were plated onto Oregon Green 488 gelatin-coated coverslips. After 6–24 h, cells were fixed and labelled with the relevant antibodies to visualize F-actin and cortactin by confocal microscopy. Arrowheads point to actin punctiform structure (white) and degradation of the gelatin matrix (black). Inset shows higher magnification of the boxed region. Scale bar, 10 μm. (b) Same experiment as in a but cells were imaged using a confocal orthogonal (x/z) slice view. Inset shows higher magnification of the arrowhead-pointed zone. Scale bar, 10 μm. (c) Quantification of Oregon Green 488 gelatin degradation areas in μm2 (_n_=3). Mean±s.e.m., ***P<0.001 (Student's _t_-test). (d) Western blot showing MT1-MMP expression in 12 breast cancer cell lines. Tubulin is used as loading control. Lysates used were the same as in Figs 1d and 3f. (e) Indicated proteins expression was assessed by immunoblot in BT-474 cells lysates expressing indicated shRNAs. (f) Upper panel: 48 h after transfection of control or _MT1-MMP_-specific siRNAs, invasion of BT-474 cells was assessed in Boyden chambers with matrigel (_n_=5). Mean±s.e.m., *_P_≤0.05 (Student's _t_-test). Lower panel: WCL of BT-474 cells transfected with control (−) or MT1-MMP (+) siRNAs were immunoblotted to check MT1-MMP expression. Numbers shown under MT1-MMP represents quantification of MT1-MMP silencing compared with control condition. (g) Invasion of BT-474 cells expressing indicated shRNAs was assessed in Boyden chamber with matrigel in the presence of 2μM Lapatinib (+) or DMSO (−). (h) Western blot showing expression of the indicated proteins in seven ERBB2+ breast cancer cell lines. Blots are the same as those presented in Figs 1d, 3f and . Lower panel indicates the presence of invadopodia in the different cell lines. N=not detected, Y=yes, ?=not tested. (i) Correlation between ERBB2/TOM1L1 and invadopodia. Histogram indicates the number of cell lines exhibiting invadopodia in function of ERBB2 and TOM1L1 expression. High ERBB2/Low TOM1L1 cell lines are SKBR3 and HCC-1954. High ERBB2/High TOM1L1 cell lines are BT-474 and UACC812 and Low ERBB2/High TOM1L1 cell line is MDA-MB 361.

Figure 5

Figure 5. TOM1L1 regulates MT1-MMP activity.

(a) BT-474 cells infected with shCtrl or shTOM1L1 encoding viruses were transfected with mCherry–MT1-MMP and GFP, TOM1L1–GFP or TOM1L1ΔGAT–GFP then plated on a gelatin/Oregon Green 488 gelatin mix and imaged by (xz) or (xy) confocal microscopy. Note the more basal localization of MT1-MMP when co-expressed with TOM1L1, but not with TOM1L1ΔGAT (arrowheads). Scale bar, 10 μm (See also Supplementary Movie 1). (b) Quantification of a. The fraction of mCherry–MT1-MMP in contact with the gelatin layer was evaluated as indicated in the upper panel. Lower panel shows fluorescence quantifications in indicated conditions (_n_=8). Mean±s.e.m., *_P_≤0.05; **_P_≤0.01 (Student's _t_-test). (c) 48 h after transfection with mCherry–MT1-MMP, BT-474 cells infected with the indicated viruses were seeded on plasma-cleaned glass coverslips coated with gelatin and imaged by epifluorescence or TIRF. Shown is a representative image out of 10. Scale bar, 10μm. (d) Confocal orthogonal imaging (x/z) of BT-474 shTOM1L1-infected cells were transfected with control (Ctrl) or TOLLIP siRNA, then transfected with mCherry–MT1-MMP and GFP–TOM1L1 and plated on gelatin-coated coverslips. Cells were treated with DMSO or 1 μM Lapatinib for 3 h and imaged by confocal orthogonal imaging (x/z). Note the strong co-localization (arrowheads) of GFP–TOM1L1 and mCherry–MT1-MMP at the basal plane (dotted lines) of cells and the loss of this localization after TOLLIP depletion or Lapatinib treatment. Scale bar, 10 μm. (e) Quantification of d. The fraction of mCherry–MT1-MMP in contact with the gelatin layer was evaluated as in b. Graph shows mean±s.e.m. of fluorescence quantifications in indicated conditions (_n_=4–10). **_P_≤0.01 (Student's _t_-test). (f) Western blot showing the efficiency of TOLLIP depletion 48 h after transfection of TOLLIP or control siRNAs in BT-474 cells.

Figure 6

Figure 6. TOM1L1 specifically drives long-range trafficking of RAB-7/MT1-MMP-positive late endosomes.

(a) Endosome tracking assay. About 48 h after mCherry–MT1-MMP transfection with Ctrl or mouse TOLLIP siRNA, 3T3 cells infected as indicated were plated on gelatin-coated glass bottom dishes. Time-lapse imaging was performed with one acquisition every 260 ms for 1 min to visualize endosome movements (see also Supplementary Movie 4). The track pattern of randomly selected endosomes is seen as coloured lines. Insets show the localization of the cell areas (boxed) showed at higher magnification. Scale bar, 10 μm. (b) Analysis of endosome tracking. Average speed (μm s−1), total distance (μm) and maximum distance from the point of origin (μm) were recorded. Histograms show mean±s.e.m. (_n_=101–202 tracked endosomes). **_P_≤0.01; ***_P_≤0.001 (Mann–Whitney test). (c) Co-localization of MT1-MMP with RAB-7. 48 h after GFP–RAB-7 and mCherry–MT1-MMP transfection, 3T3-neu cells infected with viruses expressing the indicated constructs were plated on gelatin-coated coverslips for 3 h to visualize RAB-7/MT1-MMP co-localization by confocal imaging. Boxed areas on the left panels are shown at higher magnification on the right panels. Scale bar, 20 μm (see also Supplementary Movie 5). (d) TOM1L1 does not affect lysosome trafficking. 3T3-neu cells infected as indicates were seeded on gelatin-coated glass bottom dishes and lysosomes were labelled with 50 nM Lysotracker-Red 30 min before imaging. Time-lapse imaging was performed with one acquisition every 260 ms for 1 min to visualize lysosome trafficking. The track pattern of randomly selected lysosomes is seen as coloured lines (see also Supplementary Movie 6). Scale bar, 20 μm. (e) Analysis of lysosome tracking. Average speed (μm s−1), total distance (μm) and maximum distance from the point of origin (μm) were recorded (_n_=160 lysosomes per condition). Histograms show mean±s.e.m/NS, no significant/(Mann–Whitney Test).

Figure 7

Figure 7. TOM1L1 is recruited by TOLLIP to RAB-7/MT1-MMP endosomes for MT1-MMP trafficking.

(a) Co-localization of TOLLIP and RAB-7. About 48 h after GFP–RAB-7 and HA–TOLLIP transfection, 3T3-neu cells infected with viruses expressing the indicated constructs were seeded on gelatin-coated coverslips for 3 h and immunolabelled using an anti-HA antibody. ‘Zoom' panels show higher magnification of the boxed areas. Note the relocalization of TOLLIP/RAB-7 co-localization at the cell periphery when TOM1L1 is expressed. Scale bar, 20 μm. (b) Endosomal TOM1L1 recruitment by TOLLIP. 3T3-neu cells were transfected with GFP–TOM1L1 or GFP–ΔGAT alone or with HA–TOLLIP. About 48 h after transfection cells were seeded on gelatin-coated glass bottom dishes to visualize TOM1L1 localization. ‘Zoom‘ panels show higher magnification of the boxed areas. Scale bar, 20 μm. (c) Endosomal co-localization of TOM1L1, TOLLIP and MT1-MMP. About 48 h after GFP–TOM1L1/ΔGAT, HA–TOLLIP and mCherry MT1-MMP transfection, 3T3-neu cells were plated on gelatin-coated coverslips for 3 h then immunolabelled with an anti-HA antibody to visualize co-localization. Insets show higher magnification of the boxed areas. Scale bar, 20 μm. (d) Co-localization of MT1-MMP, RAB-7 and TOLLIP in BT-474 cells. BT-474 cells were transfected with mCherry–MT1-MMP, GFP–Rab-7 and HA–TOLLIP. About 48 h after transfection, cells were plated on gelatin-coated coverslips and imaged by confocal orthogonal (x/z) imaging. Note the basal co-localization of mCherry–MT1-MMP, GFP–RAB-7 and HA–TOLLIP (arrowheads). Scale bar, 10 μm.

Figure 8

Figure 8. ERBB2 indirectly promotes TOM1L1-Ser321 phosphorylation for interaction with TOLLIP and MT1-MMP trafficking.

(a) Lysates from 3T3 cells infected as indicated were immunoprecipitated with an anti-TOM1L1 antibody and immunoblotted to visualize TOM1L1 phosphorylation. (b) SILAC mass spectrometry analysis. 3T3-neu cells transfected with GFP–hTOM1L1 were cultured for 2 weeks in medium containing light (12C614N4-Arg and 12C614N2-Lys) or heavy (13C615N4-Arg and 13C615N2-Lys) arginines and lysines and treated with 1 μM Lapatinib or not (DMSO) for 3 h before lysis. GFP–hTOM1L1 was immunoprecipitated using the GFP-Nanotrap technology then digested using trypsin. The hTOM1L1 peptides phosphorylation was then analysed by mass spectrometry (see Methods for details). Upper panel: fragmentation spectra of the single peptide find phosphorylated (localization on Ser321 with a probability >0.75 as calculated by MaxQuant). Lower panel: Heavy/Light SILAC ratio (H/L) for this peptide, traducing the phosphorylation ratio changes between the tested conditions. (c) 3T3-neu and 293 T cells infected as indicated were transfected with HA–TOLLIP. Lysates were then immunoprecipitated and immunoblotted as shown. (d) 3T3-neu cells infected as indicated were seeded in Boyden chambers with matrigel for 24 h and cells present in the lower chamber were counted. The histogram shows the invasion ratio normalized to control (_n_=3). **_P_≤0.01 (Student's _t_-test). (e) 3T3-neu cells infected as indicated were cultured on Oregon Green 488 gelatin for 24 h to visualize gelatin degradation areas. The quantification (mean±s.e.m.) of degradation areas per cell is shown (_n_=26-48). **_P_≤0.01; ***_P_≤0.001 (Student's _t_-test). (f) BT-474 cells infected with TOM1L1 shRNA and transfected with indicated constructs were imaged by confocal orthogonal imaging. Arrowheads show the change of MT1-MMP apico-basal polarity. Scale bar, 10 μm. (g) Quantification of f. The fraction of mCherry–MT1-MMP in contact with the gelatin layer was evaluated as in Fig. 5b (_n_=4). Mean±s.e.m. *_P_≤0.05 (Student's _t_-test). (h) 3T3-neu cells transfected with mCherry–MT1-MMP, HA–TOLLIP and GFP–TOM1L1 or the phosphomimetic mutant (S320E) were treated or not with 1 μM Lapatinib for 2 h. Localization of colocated spots was visualized by confocal imaging. Insets show higher magnification of the boxed areas. Scale bar, 10 μm.

Figure 9

Figure 9. A model for TOM1L1 invasive activity.

ERBB2 indirectly induces TOM1L1 phosphorylation at Ser321 to promote association with TOLLIP in RAB-7/MT1-MMP-positive late endosomes and MT1-MMP trafficking to plasma membrane for cell invasion.

References

    1. Harari D. & Yarden Y. Molecular mechanisms underlying ErbB2/HER2 action in breast cancer. Oncogene 19, 6102–6114 (2000). - PubMed
    1. Guo W. et al. Beta 4 integrin amplifies ErbB2 signalling to promote mammary tumourigenesis. Cell 126, 489–502 (2006). - PubMed
    1. Worzfeld T. et al. ErbB-2 signals through Plexin-B1 to promote breast cancer metastasis. J. Clin. Invest. 122, 1296–1305 (2012). - PMC - PubMed
    1. Johnson E. et al. HER2/ErbB2-induced breast cancer cell migration and invasion require p120 catenin activation of Rac1 and Cdc42. J. Biol. Chem. 285, 29491–29501 (2010). - PMC - PubMed
    1. Zaoui K., Honoré S., Isnardon D., Braguer D. & Badache A. Memo-RhoA-mDia1 signalling controls microtubules, the actin network, and adhesion site formation in migrating cells. J. Cell Biol. 183, 401–408 (2008). - PMC - PubMed

Publication types

MeSH terms

Substances

LinkOut - more resources