Arresten Inhibits HSC-3 Carcinoma Cell Invasion in the 3D Organotypic Model
To further explore the invasive properties of the Arr-HSC cells and to gain insight into the mechanisms of action of arresten, we performed three-dimensional (3D) organotypic assays in which HSC-3 carcinoma cells were allowed to invade into a collagen matrix supplemented with human gingival fibroblasts. After a 2weeks culture period, the organotypic sections were immunostained with E-cadherin and pancytokeratin AE1/AE3 antibodies, and the maximal invasion depth and area, and the thickness of the top cell layer were determined. As expected, the Ctrl-HSC cells invaded deep into the collagen matrix and E-cadherin staining clearly decreased in the matrix-invaded cells indicating loosening of the cell-cell contacts during the invasion (Figure 3A). Arresten overexpression almost completely blocked HSC-3 cell invasion, the maximal invasion depth and the area of invading cells being significantly smaller than those of the control cells.
Figure 1. Arresten inhibits migration of HSC-3 cells. A. 30 000 Ctrl-HSC and Arr-HSC cells were allowed to migrate through Transwell inserts and the number of migrated cells was counted under a microscope at 506magnification. Mann-Whitney U-test, ***p,0.001, (n = total number of fields analyzed, 2? fields per Transwell insert). B. 30 000 HSC-3 cells were allowed to migrate through Transwell inserts in the presence of human recombinant purified arresten (5 and 20 mg/ml) and the number of migrated cells was counted as described above. Mann-Whitney U-test, **p,0.01, (n = total number of fields analyzed, 3? fields per Transwell insert). C. Scratch wound healing assay with Ctr-HSC and Arr-HSC clones in which the closure of the wound was measured at 0, 16 and 48 h. Scale bar 50 mm. E. Quantification of scratch wound healing in the Ctrl-HSC and Arr-HSC clones. Mann-Whitney U-test, ***p,0.001, (n = 70 fields at 0, 16 and 48 h per clone). a very thin top cell layer, with prominent membranous E-cadherin staining (Figure 3A).mesenchymal marker vimentin were observed in some individual Ctrl-HSC and Arr-HSC cells, but evident differences in these signals could not be detected between the cell lines (Figure S7).
Arresten Overexpression Promotes an Epithelial Morphology and E-cadherin in Cell-cell Contacts
Besides the non-migratory and less invasive phenotype of ArrHSC cells observed in the previous assays, we noticed a prominent change in their cell morphology. Compared to the control HSC-3 cells, the Arr-HSC clones displayed a flatter, less spindle-shaped phenotype and they grew in aggregated cobblestone-like clusters (Figure 4A). Similar morphological changes were observed in MDA-MB-435 breast carcinoma cells in the presence of excess arresten (Figure S2E), These findings led us to hypothesize that arresten may affect the epithelial plasticity of the HSC-3 cells, and induce a switch from the mesenchymal carcinoma cell phenotype to a one resembling normal epithelial cells. The carcinoma cells undergo EMT-like events during cancer progression, and a reversed process MET is suggested to occur, endowing a less motile phenotype [22,31]. Accordingly, we further investigated whether arresten overexpression could restore the epithelial characteristics of the tumor cells. The Arr-HSC cells growing in tightly packed clusters expressed more epithelial marker E-cadherin on their cell surfaces than the Ctrl-HSC cells (Figure 4B), which is likely to contribute to their epithelial-like morphology and reduced motility. Besides the recruitment of E-cadherin to the Arr-HSC cell membrane, its expression in these cells was increased when compared to the Ctrl-HSC cells (Table S1, Figure 4C). The amount of E-cadherin mRNA in the Arr-HSC cells was 1.9-fold 60.06 (p,0.001) (Table S1, Figure 4E), and that of protein 1.6fold 60.12 (p = 0.019), both significantly higher than in control cells (Figure 4C).
Arresten Affects Cell Proliferation and Apoptosis of HSC-3 Cells in vitro
We next wished to determine the reason underlying the thin top cell layer formed by the Arr-HSC cells in the organotypic model, and set out to study tumor cell proliferation and apoptosis. The number of proliferating Ki-67-positive tumor cells was smaller, but not statistically significant, in the Arr-HSC than in the Ctrl-HSC 3D cultures (Figure 5A), which is in agreement with our observation on reduced tumor cell proliferation in Arr-HSC xenografts (Figure 2D). The TUNEL assay showed that the Arr-HSC cells underwent apoptosis more often than the control cells in the 3D model (Figure 5C). Since the TUNEL assay also detects other types of cell death in addition to apoptosis, we wanted to confirm our finding by caspase-3 staining. We observed a similar and significant (p = 0.030) trend on increased apoptosis in Arr-HSC cells (Figure 5E) although the increase was milder than the one in the TUNEL assay. In HSC-3 xenografts, however, only few TUNEL-positive cells were detected mainly in the keratinized central tumor areas (Figure S8). We have previously shown that recombinant arresten affects mitochondrial apoptosisrelated Bcl-family signaling molecules in microvascular endothelial cells [18].
Figure 2. Effects of arresten on HSC-3 xenografts. A. One million Ctrl-HSC and Arr-HSC cells were injected subcutaneously into the flanks of nude mice and tumor growth was monitored over 16 days. Students t-test, *p,0.05, (n = 10 mice per group). B. Local invasiveness of the tumors. C. Representative hematoxylin-eosin stainings of HSC-3 xenografts. Scale bar 100 mm. D. HSC-3 xenografts were stained for the proliferation marker Ki-67 (brown) and the cell proliferation was defined as a percentage of Ki-67-positive cells among the total number of carcinoma cells per microscopic field (4006magnification; n = number of fields analyzed, 3? fields per xenograft). Scale bar 50 mm. F. The tumor blood vessels were stained with a CD31 antibody and counted under a microscope (2006magnification; n = number of fields analyzed, 3? fields per xenograft). Mann-Whitney Utest, ***p,0.001. Scale bar 100 mm.(p = 0.12), thus shifting the balance towards a situation favoring apoptosis (Figure 5G).Electric Cell-substrate Impedance Sensing Reveals Alterations in Arr-HSC Cell Spreading and Cell-cell Contacts
To pursue the mechanisms underlying the altered behavior and morphology of Arr-HSC cells we performed measurements using electric cell-substrate impedance sensing (ECIS), a method that provides quantitative data on cell attachment, spreading and the strength of cell-cell contacts by monitoring changes in the system impedance [35]. The Arr-HSC cells showed markedly higher impedance at a low frequency than the control cells (Figure 6A). Also the HSC-3 cells treated with ArrCM showed higher
impedance than those treated with CtrlCM (Figure S9A). The change in the impedance can be related either to cell inherent dielectric properties, formation of cell-cell junctions or cellsubstrate interactions, and a mathematical ECISTM Model can be applied to distinguish these parameters from each other [36]. Thus, a cell membrane capacitance (Cm) reflects the structure and folding of cell membrane, a barrier resistance (Rb) refers to establishment of cell-to-cell junctions, and a cell-substrate interaction parameter a is linearly related to the cell surface area and, inversely, to the distance between cell and substrate [36?8]. This modeling supported our observations on altered cell morphology and E-cadherin of the Arr-HSC cells. First, significantly increased Rb of the Arr-HSC cells relative to the Ctrl-HSC implied tightening of intercellular junctions (Figure S9C).
