Matrix invasion of leader cells and cancer-associated fibroblasts regulated by distinct functions of cellular protrusions and collagen remodeling

Introduction: Breast tumor microenvironment is composed of heterogeneous cell subpopulations, including healthy epithelial cells, aggressive leader cells, and cancer associated fibroblasts (CAFs). These cell types are assumed to play different roles in tumor invasion, such as generating protrusive forces and performing collagen remodeling across the tumor-stromal boundary. Among these, leader cells and CAFs are both known to invade aggressively and generate forces. Yet, their relative functions in these processes remain unclear. As leader cells and CAFs are distinct cell types with different modes of migration, understanding of their relative contributions could play a crucial role in reducing the metastatic potential of tumor microenvironment. For instance, if leader cells are more important than CAFs in collagen remodeling, or vice versa, it would be important to target the correct cell type in modulating TME.

Materials and

Methods: Mouse breast cancer cell line 4T1, representing leader-like cells, and mouse tumor-derives CAF immortalized cells are made into spheroids using ultra-low binding u-shaped-bottom microplates. Spheroids are inserted between the collagen layers to ensure full 3D environment. Fluorescence beads are embedded in collagen to image and analyze collagen deformation. For the magnetically aligned collagen conditions, we designed and fabricated 3D-printed scaffolds (using Form-3B printer) with N52 magnets around wells for cell culture (Figure 1K). For the collagen pre-alignment condition, magnetic beads are used in a place of fluorescence beads. Two-photon confocal microscopy was performed to image collagen fibers around cells. Live-cell tracking is done to observe cell’s invasion and bead movements, which are analyzed by particle velocimetry from MATLAB. Collagen alignment is analyzed by CT-FIRE and R-studio code. CK-666 and Blebbistatin treatments are used to pharmacologically inhibit ARP2/3 and myosin-II, respectively.

Results, Conclusions, and Discussions: Our results with two different cell types interrogate the conventional proportional relationship between cellular forces and migration (Fig. A). Here, we found that the leader-like cells (4T1s) invade in a protrusion-dominant fashion without requiring significant collagen remodeling, while CAFs are more contractile and remodel collagen fibers yet undergo lower invasion (Fig. B-D). Thus, inherent protrusive forces can navigate existing pores to invade, without requiring force-based collagen remodeling. However, in denser collagen, the importance of force-based collagen remodeling for effective invasion arises, as 4T1s that invade faster must generate higher collagen remodeling compared to CAFs (Fig.E). After Arp2/3 inhibition, 4T1s continue to invade using filopodial protrusions while CAFs turn to blebbing to explore and invade through existing pores in collagen. While conventionally fiber alignment is thought to aid persistent invasion, here we found that more random collagen fibers allow greater collective invasion across cell types and collagen densities (Fig. F-G). Our co-culture experiments show that moderate (10%) addition of CAFs to 4T1s is advantageous for net invasion (Fig. H). In this condition, the largely 4T1s population keeps the fiber organization random due to their highly protrusive mode of migration, and yet the small population of contractile CAFs generates ample collagen deformation, both of which together aid net invasion (Fig.I-J). Consistent with this notion that random collagen fibers aid invasion, in the magnetically pre-aligned collagen fibers, cell invasion was insensitive to collagen density (Fig.K-M).
Tumor microenvironment is composed of heterogenous cell types and dynamically evolving matrix structure, which complicate understanding and controlling tumor invasion and metastasis. In a step towards decoupling these heterogeneities, our study reveals distinct functions of protrusive leader cells and contractile CAFs in collective cell invasion, which benefits from random collagen fiber organization and does not always require contractility-dependent collagen deformation. Our findings suggest that therapies individually targeting cellular protrusions or contractility may not work in all collagen densities, which vary significantly with breast cancer progression, and better understanding of functional importance of cell subtypes in breast cancer are needed for advanced cancer therapies.

Acknowledgements (Optional):