Metastatic growth from dormant cells induced by a col-I-enriched fibrotic environment

Abstract

Breast cancer that recurs as metastatic disease many years after primary tumor resection and adjuvant therapy seems to arise from tumor cells that disseminated early in the course of disease but did not develop into clinically apparent lesions. These long-term surviving, disseminated tumor cells maintain a state of dormancy, but may be triggered to proliferate through largely unknown factors. We now show that the induction of fibrosis, associated with deposition of type I collagen (Col-I) in the in vivo metastatic microenvironment, induces dormant D2.0R cells to form proliferative metastatic lesions through beta1-integrin signaling. In vitro studies using a three-dimensional culture system modeling dormancy showed that Col-I induces quiescent D2.0R cells to proliferate through beta1-integrin activation of SRC and focal adhesion kinase, leading to extracellular signal-regulated kinase (ERK)-dependent myosin light chain phosphorylation by myosin light chain kinase and actin stress fiber formation. Blocking beta1-integrin, Src, ERK, or myosin light chain kinase by short hairpin RNA or pharmacologic approaches inhibited Col-I-induced activation of this signaling cascade, cytoskeletal reorganization, and proliferation. These findings show that fibrosis with Col-I enrichment at the metastatic site may be a critical determinant of cytoskeletal reorganization in dormant tumor cells, leading to their transition from dormancy to metastatic growth. Thus, inhibiting Col-I production, its interaction with beta1-integrin, and downstream signaling of beta1-integrin may be important strategies for preventing or treating recurrent metastatic disease.

Figure 1. Fibrosis induces the dormant-to-proliferative switch of D2.0R cells in the lung

A) Left: Lung sections of Ad-empty (no fibrosis) and Ad-TGFβ1 treated (fibrosis) mice stained for Col-I (red). Right: Metastatic outgrowth of D2.0R-GFP cells (green) in fibrotic lungs stained for Col-I (red) (immunofluorescence, confocal microscopy, X63). B) Total D2.0R-GFP tumor cell burden/lung in fibrotic lungs (n=5) compared to non-fibrotic lungs (n=5; ***p ≤0.001). C) Average size of metastatic lesions/lung in fibrotic lungs (n=5) compared to non-fibrotic lungs (n=5; **p ≤0.01). D) Percentage of single cells vs. multicellular proliferative metastatic lesions in non-fibrotic (n=5) and fibrotic lungs (n=5). Insert: SCOM images of D2.0R-GFP lung lesions in mice, x100.

Figure 2. Col-I induces the transition from quiescence to proliferation of D2.0R and D2A1 cells through Intβ1

A) Proliferation of D2.0R and D2A1 cells in BME+Col-I, compared to cells grown in BME (mean ± SE, n=8, p ≤0.0001). B) Western blot for Intβ1 expression. Left panel: D2.0R cell lines. Right panel: D2A1 cell lines. C–D) Proliferation of D2.0R (C) and D2A1 cells (D) in BME+ Col-I (mean ± SE; n=8). (*p ≤0.05; **p ≤0.01; ***p ≤0.0001). Representative results of three independent experiments for all data shown.

Figure 3. Loss of Intβ1 expression inhibits the fibrosis-induced transition from dormancy to metastatic outgrowth in mice

A) Fold difference in total tumor burden per lung for each cell line in response to fibrosis (total pixels in fibrotic lungs/total pixels in non-fibrotic lungs). Fibrosis significantly increases total tumor burden/lung of D2.0R-sh–NT cells compared to non-fibrotic lungs (*p=0.018 unadjusted, p=0.05 with Bonferroni correction [BFC]), whereas inhibition of Intβ1(sh-Intβ1) abolishes the response to fibrosis (no significant difference, p=0.89 unadjusted, p=1 with BFC). [sh-NT - sh-Intβ1] represents the comparison between the sh-NT class analysis vs sh- Intβ1 analysis (ns). B) Average size of the metastatic lesions in fibrotic lungs (Ad-TGFβ) from D2.0R-GFP- sh-NT cells and D2.0R-GFP-sh-Intβ1 cells (*p=0.00014 uncorrected; p=0.00042641 with BFC) as well as for D2.0R-GFP-sh-non-target (sh-NT) cells in non-fibrotic lungs (Ad-empty) (p=0.013 unadjusted; p=0.039 with BFC). C) Percentage of single cells vs. proliferative metastatic lesions in non fibrotic vs. fibrotic lungs of D2.0R-GFP-sh-NT or D2.0R-GFP sh-Intβ1 cells. D2.0R-GFP- sh-NT cells had a significantly higher percentage of multi-cellular, proliferative lesions in fibrotic lungs compared to D2.0R-GFP-sh-Intβ1 cells (p=0.001 uncorrected; p= 0.003 with BFC) as well as for D2.0R-GFP- sh-NT cells in non-fibrotic lungs (p=1.75E-05 uncorrected; p= 5.25E-05 with BFC).

Figure 4. Src activation is required for Col-I-induced FAK^Y397 phosphorylation and the transition from quiescence to proliferation of D2.0R cells

A, C) Co-immunofluorescence staining for Src^Y416, FAK^Y397 and nuclear staining with DAPI. A) Cells cultured on BME and BME+Col-I for 24h. C) D2.0R cells cultured on BME+Col-I for 24h with nonspecific IgG (100μg/ml) or D2.0R non-target shRNA cells or D2.0R cells treated with anti-Intβ1 antibody (100μg/ml) or pooled D2.0R cells and D2.0R-sh-Intβ1 cells B) Co-immunofluorescence staining for Intβ1, FAK^Y397 and nuclear staining with DAPI of D2.0R cells cultured on BME+Col-I for 24h treated either with PP1 (10 μM) or sh-Src. D) Proliferation of D2.0R cells in BME or BME+Col-I with or without 10 μM PP1, non-target shRNA, or sh-Src (n=8; mean ±SE; P ≤ 0.0001). Confocal microscopy, x63, white bar = 20um. Representative of three independent experiments.

Figure 5. CoI-I induces actin stress fiber formation and activation of ERK through Intβ1 and Src

A–D) Co-immunofluorescence staining of D2.0R cells for pERK, F-actin with phalloidin and nuclear staining with DAPI after A) 24h culture on BME or BME +Col-I; B) D2.0R cells on BME+Col-I treated with non-specific IgG (100μg/ml) or anti-Intβ1 antibody (100μg/ml); C) sh-non target and sh-Intβ1; D) 10 μM PP1, or sh-Src. Confocal microscopy x63, white bar = 20 microns. Representative of three independent experiments.

Figure 6. Col-I mediated signaling through Intβ1 induces phosphorylation of MLC by MLCK leading to proliferation of quiescent cells

A, C–D) Co-immunofluorescence staining for phospho-MLC, F-actin and nuclear localization with DAPI. of D2.0R cells cultured for 24h on BME or BME+Col-I A) either untreated or treated with ML-7 (5 μM), U-0126 (10μM) or W-13 (10μM). B) Proliferative response to Col-I with or without inhibitors (n=8, mean ±SE; P ≤0.0001). C) D2.0R cells on BME+Col-I with non-specific IgG (100μg/ml), anti-Intβ1 antibody (100μg/ml) and D2.0R cells stably expressing non-target-shRNA and D2.0R sh-Intβ1cells (pooled or clone #47-6). D) D2.0R cells cultured on BME+Col-I (Control) or treated with 10 μM PP1, or sh-Src. Confocal microscopy x63, white bar = 20 microns. Representative results of three independent experiments.