Fig 2: Rescue of 3D invasive growth by gremlin-1 overexpression(A) Early and late passage JP4 mesothelioma cells were embedded into 3D Matrigel and followed for 72 hours. (B) Expression of GREM1 and SNAI2 in early and late passage JP4 and JP5 mesothelioma cells. The results are expressed relative to the expression level in early passage JP4 cells, which was set to 1. A representative experiment is shown. (C) Late passage JP4 mesothelioma cells transduced to express gremlin-1 (JP4/grem1) or control cells (JP4/ctrl) were embedded into 3D Matrigel and followed for 72 hours. (D–E) Early passage JP3 and JP9 primary mesothelioma cells and JP5 mesothelioma cells transduced to express gremlin-1 (JP5/grem1) or control cells (JP5/ctrl) were embedded into 3D collagen 1. Images were taken at 24 or 72 hours. Graphs shows quantification as relative spheroid surface area. The error bars represent SD (n = 3). *p < 0.05. (F) Expression of SNAI2 and CDH1/E-cadherin in JP5/ctrl and JP5/grem1 cells. The results are expressed relative to the expression level in JP5/ctrl cells, which was set to 1. The error bars represent SD (n = 3). *p < 0.05. (G) Cell conditioned media (normalized according to cell number) or cell lysates were analyzed by Western blotting using antibodies specific to gremlin-1, E-cadherin or ß-tubulin (loading control). The molecular weight markers (kDa) are shown on the left. (H) JP5/ctrl and JP5/grem1 cells were stained for F-actin (red) and analyzed using immunofluorescence microscopy. Cell nuclei were counterstained with DAPI (blue).

Fig 3: Inhibition of BMP receptor pathway alone does not induce invasive growth(A) Expression of BMP-pathway genes in gremlin-1 silenced (grem1_siRNA) H2052 or JL-1 cells and in JP5/grem1 cells. The results are expressed relative to each control (control siRNA transfected or JP5/ctrl), which was set to 1. The error bars represent SD (n = 3). *p < 0.05. (B) BMP [(Bre)2-luc] pathway activity was analyzed in H2052 cells using promoter reporter assays. Promoter activities in BMP-2 (25 ng/ml) and/or LDN193189 (100 nM) treated (24 hours) cells are shown relative to the level in control treated cells, which was set to 1. The error bars represent SD (n = 3). *p < 0.05. (C) BMP [(Bre)2-luc] and TGF-ß [(CAGA)12-luc] pathway activities were analyzed using promoter reporter assays. Promoter activities in LDN193189 (100 nM) treated cells (3 days) are shown relative to each control, which was set to 1. The error bars represent SD (n = 3). *p < 0.05. (D) Expression of ID1, SNAI2, and MMP2 in LDN193189 (100 nM) treated mesothelioma cells (H2052, JL-1 and JP5/ctrl). The results are expressed relative to each control, which was set to 1. The error bars represent SD (n = 3). *p < 0.05. (E) Control or LDN193189 (100 nM) treated JP5/ctrl cells were embedded into 3D Matrigel and followed for 72 hours.

Fig 4: Mesothelioma cell adhesion and migration is regulated by gremlin-1(A) Extracellular matrix and adhesion molecule PCR array was used to analyze gene expression in control siRNA (ctrl_siRNA) or gremlin-1 siRNA (grem1_siRNA3) transfected H2052 cells 3 days after transfection. The results are expressed relative to the expression level of ITGA1 in H2052 control siRNA transfected cells, which was set to 1. The error bars represent SD (n = 2). (B) Expression of integrin genes in gremlin-1 silenced (grem1_siRNA) H2052 or JL-1 cells and in JP5/grem1 cells. The results are expressed relative to each control (control siRNA transfected or JP5/ctrl), which was set to 1. The error bars represent SD (n = 3). *p < 0.05. (C) Control siRNA (ctrl_siRNA) or gremlin-1 siRNA (grem1_siRNA3) transfected H2052 cells were allowed to adhere to fibronectin or collagen 1 coated wells for 1 hour. (D) JP5/ctrl or JP5/grem1 cells were allowed to adhere to collagen 1 for 30 min. Quantification of cells with a spread/elongated morphology is shown. The error bars represent SD (n = 3). *p < 0.05. (E) Cell migration was analyzed using a scratch-wound assay and monitored by 24-h live cell imaging. The results are presented as relative wound density. The error bars represent SD (n = 2). (F) Migration (1% serum as chemoattractant) or invasive migration (collagen 1 coated Transwell inserts) of JP5/ctrl, JP5/grem1 and JL-1 cells transfected with control or gremlin-1 siRNA is shown. Migrated cells were fixed, stained and imaged 16 hours after seeding. Graphs represent quantification of relative migration. The error bars represent SD (n = 3). *p < 0.05.

Fig 5: MMP activity is crucial for gremlin-1 induced 3D collagen invasion(A) Extracellular matrix and adhesion molecule PCR array was used to analyze gene expression in control siRNA (ctrl_siRNA) or gremlin-1 siRNA (grem1_siRNA3) transfected H2052 cells 3 days after transfection. The results are expressed relative to the expression level of MMP9 in H2052 control siRNA transfected cells, which was set to 1. The error bars represent SD (n = 2). (B) Expression of MMP genes in gremlin-1 silenced (grem1_siRNA) H2052 or JL-1 cells and in JP5/grem1 cells. The results are expressed relative to each control (control siRNA transfected or JP5/ctrl), which was set to 1. The error bars represent SD (n = 3). *p < 0.05. (C) Cell conditioned media (normalized to cell number) were analyzed by gelatin zymography. MMP-9 and MMP-2 (inactive, intermediate and active forms) in control sample (PMA treated HT1080 lysate) are shown on the left, molecular weight markers (kDa) on the right. Quantification of MMP-2 level is indicated from a representative experiment. (D) Invasive growth of H2052 and JP5/grem1 cells was analyzed in 3D collagen 1 matrix in the presence of DMSO or MMP inhibitors (GM6001 or BB2516 at 10 µM concentration). Graphs show quantification as relative spheroid surface area. The error bars represent SD (n = 3). *p < 0.05. (E) Images from H2052 cells treated with DMSO or MMP inhibitors were taken at 72 hours.