Alveolar Rhabdomyosarcoma 3D model development to mimic physiological cell-ECM interactions with focus on integrins

Introduction: Rhabdomyosarcoma (RMS) is the most common Soft Tissue Sarcoma in childhood, the two main subtypes are embryonal RMS (ERMS), associated with a better prognosis, and alveolar RMS (ARMS), more aggressive and highly metastatic. If the knowledge of RMS genomic alterations is well established, its microenvironmental characterization is still poorly defined. So far, in vitro 2D models are used to recapitulate the interactions between cancer cells and stromal cells. However, these models are not representative of the complex biological processes that happen in vivo, such as cell migration. This, in particular, depends on 3D interactions between cells and ECM via adhesion molecules, i.e. integrins. In this context, the cell-ECM and cell-cell interactions are better studied with 3D models that offer a platform where culture conditions approximate better the physiological conditions. Aim: This work aims at the development of a 3D model able to recreate the 3D complex cells-ECM interactions, with particular attention on integrins, and to represent the cell migration process taking place in physiological conditions. Material and Methods: Bioinformatic analysis was used to determine differential expression of ECM genes in ARMS end ERMS patients. Decellularization of ARMS xenogenic tumor masses employed cycles of detergents and enzymatic treatments (DET). Three different recellularization strategies were adopted: superficial seeding, microinjection and a perfusion bioreactor. Mass spectrometry (MS) analysis of ARMS tissue was performed to determine ECM protein composition. Two 3D models: 1- Ultrafoam collagen I sponge, 2- hyaluronic acid/PEG hydrogel (HA/PEG) were developed. ITGA5 role in cell motility was investigated in vitro upon siRNA transfection evaluating. Tumor growth and metastatic migration was tested in vivo. Results: 15 ECM genes were shown to be differentially expressed between ARMS from ERMS patients. Xenogenic ARMS were successfully decellularized but the three recellularization techniques tested were not optimal in terms of viability and cell distribution. MS revealed major presence of collagens (type I and type III), fibrillin, fibronectin and periostin in ARMS ECM building. With Ultrafoam collagen I sponge we obtained a tissue-like structure in 7 days of culture, higher proliferation rates (41% vs 24%), enhanced secretion of MMP-2 and overexpression of ITGA5 and CXCR4 mRNAs compared to 2D controls. HA/PEG hydrogel formed a 3D support where cultured spheroids showed no invasion. In vitro migration assay showed reduction of migrating cells upon ITGA5 siRNA silencing (24.3% vs 43.9% in the control). In the invasion assay, cells were unable to invade the Matrigel, however we reported differential cell clustering with larger multicellular strands in control cells and smaller spherical aggregates in ITGA5 silenced cells. In vivo tumor growth showed no dependence on ITGA5; conversely, extravasation rate was higher in presence of ITGA5 (30.6% vs 8.5%) in the zebrafish model. Discussion: This study highlighted the first preliminary results on ARMS cell-ECM interaction. Among the tested 3Dmodels, the direct use of the ARMS ECM evidenced reduced porosity, impacting on superficial cell seeding and prevalence of cell-cell interactions rather than cell-ECM adhesions. The use of commercially available scaffold composed of Collagen I (Ultrafoam) gave the best results in terms of interaction with the microenvironment; however, the bioreactor is inaccessible for fluorescence live imaging. In contrast, hydrogels are optically transparent and easier to enrich with other ARMS ECM specific proteins. In HA/PEG hydrogel, concentration of fibronectin has to be optimized together with the addition of other ECM proteins. In vitro results on ITGA5 expression by RH30 cells suggest that other fibronectin-binding integrins can cooperate for cell migration. Differences in cell clustering suggested an interplay between ITGA5 and cell-cell adhesion proteins. In vivo experiments imply that ITGA5 is not required for tumor growth and appeared to be functionally relevant for the extravasation process. Conclusions: This work developed three different 3D models of ARMS, each one with specific advantages and disadvantages that have to be considered depending on the investigated biological process. In the future, we foresee that deeper investigation on ARMS microenvironment could develop new prognostic or therapeutic markers to ameliorate the overall survival of the young patients.