Assistant Professor The University of Texas at Dallas, United States
Introduction:: Mechanical interactions between cells and the extracellular matrix (ECM) have been shown to regulate important processes such as cell adhesion, proliferation, and migration [1]. These interactions have a particular relevance in breast cancer, where rogue signaling from transformed epithelial cells leads to increased deposition, alignment, and stiffening of peritumoral collagen, which can ultimately trigger cancer cell invasion and lead to poor prognosis. Despite the importance of mechanical remodeling in disease initiation and progression, the underlying microscopic relationship between ECM mechanics and cancer cell-driven remodeling remains poorly understood. This is primarily due to the inherent complexities involved in measuring the micro-scale mechanics of ECM networks, which are known to be structurally and mechanically heterogeneous and anisotropic.
To enhance our understanding of the micro-mechanics of fibrous ECM scaffolds, we have developed a novel in-vitro approach that employs ferromagnetic beads embedded into collagen networks as localized micro-scale actuators. These beads are torqued about specific axes through the application of coordinated magnetic fields in orthogonal directions to generate deformations in the surrounding network (Figure 1a). Compared against traditional methods of characterizing matrix micro-mechanics, which usually operate in a 2D setting and are low throughput, our approach can measure the nonlinear and 3D aspects of collagen mechanics with a high throughput while exploring the associated spatio-temporal heterogeneities [2]. In this study, we demonstrate this capability by mapping spatial and temporal evolution in collagen structure and mechanics as it is remodeled by breast cancer cells.
Materials and Methods:: Custom iron-oxide beads (with an average diameter of 4.8µm and a density of 105 particles/mL) were distributed evenly in a collagen solution of fixed concentration (2 mg/mL), which was then polymerized using established protocols [3]. To establish covalent bonds between the beads and collagen, their surface was functionalized with NHS groups following the protocol described by Asgeirsson et al [4]. A Nikon AX confocal microscope, equipped with a 20x objective lens and operating in reflection mode, was used to assess the micro-structural organization of the collagen network. 3D twisting was achieved using a custom microscope insert unit with two orthogonal sets of coils to generate magnetic fields. A torque was applied by means of a strong magnetizing pulse followed by a weaker twisting field in a perpendicular direction (BM and BT in Figure 1a respectively). Confocal image stacks acquired before and after the magnetic stimulation were compared using a digital image correlation algorithm to measure the deformations induced in the collagen network in a plane perpendicular to the axis of the applied torque [5]. These deformations were compared against the deformations produced in an iterative finite element simulation (Figure 1b) to estimate the unknown local mechanical properties of the network. . For the sake of simplicity, in this study we modeled collagen networks as neo-Hookean materials described by W = mu/2*(IC-3) and the associated Young’s modulus E= 2mu(1+v) was used to report the local material properties.
Results, Conclusions, and Discussions:: We validated our novel method of measuring micro-scale mechanics using synthetic deformations derived from finite element simulations. We imaged acellular collagen gels and varied both fiber density (by changing concentration experimentally) and the magnitude of matrix displacements (by changing the Young’s moduli computationally). We found that our analysis successfully recovered the imposed Young’s moduli with a mean error of -4%. Next, we conducted experiments on acellular collagen and found that our method captures well changes in fiber density, cross-linking and alignment.
To investigate the mechanical signatures of breast cancer cell mediated ECM remodeling, we first dispersed either non-tumorigenic MCF-10A or post-metastatic MDA-MB-231 single cells in free-floating collagen gels. In contrast to acellular gels, which were mechanically stable over time, cell-laden collagen gels compacted by up to 50% of their original area and experienced a gradual increase in their material stiffness, with a near 2-fold stiffening observable by day 4 in the MDA-MB-231 laden gels respectively (Figure 1c). Conversely, MCF-10A laden gels displayed a 1.6-fold stiffening over the same time frame. Next, we generated tumor spheroids using the two cancer cell lines, embedded them in anchored collagen gels, and monitored their growth and invasion over the course of four days. The more epithelial MCF-10 spheroids invaded collectively while the more mesenchymal MDA-MB-231 spheroids invaded as single cells [6]. In both cases, heterogenous structural remodeling of the peritumoral collagen was accompanied by changes in material properties. Fiber densification, alignment, and stiffening were particularly marked around MDA-MB-231 spheroids (Figure 1d).
We have thus demonstrated that magnetic twisting is capable of monitoring cancer cell-driven evolution in material properties with both spatial and temporal resolution. While our work is currently limited by the adoption of a neo-Hookean material model, our approach, which combines confocal imaging and micro-mechanical measurements to include both structural and mechanical information, lends itself to the exploration of constitutive models that incorporate structural information on collagen remodeling. Future work will focus on evaluating such structure-based constitutive models of collagen network mechanics and on investigating the mechanisms of cancer cell driven ECM remodeling by selectively inhibiting processes such as contractility, proteolysis, and cross-linking.
Acknowledgements (Optional): :
References (Optional): : [1]. Doyle A. et al., Phys Biol, 19:021002, 2022
[2]. Ferruzzi J. et al., in Multi-scale Extracellular Matrix Mechanics and Mechanobiology, 343-387: 2020.
[3] Ferruzzi J. et al., Sci Rep, 9: 17151, 2019.
[4] Asgeirsson D.O. et al., Lab Chip, 21: 3850–3862, 2021.
[5] Yang J. and Bhattacharya K., Exp Mech, 4: 719-735, 2021
[6] Kang, W., et al. Iscience 24.11 (2021): 103252.
