F. Lignet (1,3), A. Emde (2), Y. Yarden(2), R.M.H. Merks(3,4,5), B. Ribba (1)
(1)Inria Rhône-Alpes, project-team NUMED, Montbonnot; Ecole Normale Supérieure de Lyon, France; (2) Department of Biological Regulation, Weizmann Institute of Science, Rehovot 76100, ISRAEL; (3) Centrum Wiskunde & Informatica, 1098 XG Amsterdam, The Netherlands; (4) Netherlands Institute for Systems Biology/Netherlands Consortium for Systems Biology; (5) Mathematical Institute, Leiden University, Leiden. The Netherlands.
Objectives: The receptor HER2 is over-expressed in 20 to 30% of breast cancers and is associated with invasive phenotypes and poor survival prognosis [1]. Treatments directed against this receptor or the downstream pathways exist but show low efficacy or induce resistance [2]. The objective of this study was to develop a computational model that mimics breast cancer cells spheroid formation in order to better understand and optimize treatment action.
Methods: Our model was developed based on data for wild and HER2+ mammary epithelial cells (MCF10A) grown in 3D Matrigel cultures. It was observed that normal cells form regular duct-like spheroids while HER2 over-expressing cells form irregular, filled structures comparable to ductal carcinoma in-situ, a pre-malignant lesion [3]. Measures of volume, lumen size and compactness of the spheroids were assessed using an image analysis software [4].
We used the Cellular Potts Model [5], a cell-based model, in which virtual cells are sets of points of the computational domain and possess user-defined individual behaviour rules. Simulations were performed in the modelling environment CompuCell3D [6], starting from a single cell that undergoes mitosis. Cells evolve depending on the contact with other cells, extracellular matrix (ECM) and lumen. We determined measures of volume, lumen size and compactness of the simulated spheroid to compare with the data.
Results: Our model includes the following rules: depending on their position in the cluster, cells may proliferate, polarize orthogonally to the lumen or enter apoptosis, and their adhesion properties vary. We performed a sensitivity analysis of the model parameters, estimating the parameter values resulting in normal and mutated structures. We showed that the length of the contact to ECM triggering mitosis is of primary importance in normal spheroids. In addition, passage from a normal phenotype to a tumorigenic phenotype includes a decrease of the apoptosis rate, a higher probability of proliferation, and a loss of polarization.
Conclusions: We developed a computational model that permitted us to highlight the cellular processes involved in the formation of HER2+ structures compared to the wild phenotype of human mammary spheroids. The next step of this project is to integrate the molecular pathways of HER2 signalling and to simulate their blockade by targeted treatments. Comparison with experimental data may give an insight on links between molecular signals and cellular processes [7]
References:
[1] Salmon et al. Science (1987) 9:177-182
[2] Morrow et al, Breast Cancer Research (2009) 11:207
[3] Debnath et al, Methods (2003) 30:256-68
[4] Emde et al, Oncogene (2011) 30:1631-1642
[5] Glazier and Graner. Phys. Rev. E (1993) vol. 47(3):2128-2154
[6] Izaguirre et al, Bioinformatics (2004) 20:1129-1137
[7] Del Bene et al, Cancer Chemotherapy Pharmacology (2008) 63:827-836.
Reference: PAGE 21 () Abstr 2565 [www.page-meeting.org/?abstract=2565]
Poster: Other Modelling Applications