Distinct roles of protrusions and collagen deformation in collective invasion of cancer cell types.

Abstract

The breast tumor microenvironment is composed of heterogeneous cell populations, including normal epithelial cells, cancer-associated fibroblasts, and tumor cells that lead collective cell invasion. Both leader tumor cells and CAFs are known to play important roles in tumor invasion across the collagen-rich stromal boundary. However, their individual abilities to utilize their cell-intrinsic protrusions and perform force-based collagen remodeling to collectively invade remain unclear. To compare collective invasion phenotypes of leader-like tumor cells and CAFs, we embedded spheroids composed of 4T1 tumor cells or mouse tumor-derived CAF cell lines within 3D collagen gels and analyzed their invasion and collagen deformation. We found that 4T1s undergo greater invasion while generating lower collagen deformation compared to CAFs. Although force-driven collagen deformations are conventionally associated with higher cellular forces and invasion, here 4T1s specifically rely on actin-based protrusions, while CAFs rely on myosin-based contractility for collective invasion. In denser collagen, both cell types slowed their invasion, and selective pharmacological inhibitions show that Arp2/3 is required but myosin-II is dispensable for 4T1 invasion. Furthermore, depletion of CDH3 from 4T1s and DDR2 from CAFs reduces their ability to distinguish between collagen densities. For effective invasion, both cell types reorient and redistribute magnetically pre-aligned collagen fibers. With heterogenous cell populations of co-cultured CAFs and 4T1s, higher percentage of CAFs impeded invasion while increasing collagen fiber alignment. Overall, our findings demonstrate distinctive mechanisms of collective invasion adopted by 4T1 tumor cells and CAFs, one relying more on protrusions and the other on force-based collagen deformation. These results suggest that individually targeting cellular protrusions or contractility may not be universally applicable for all cell types or collagen densities, and a better cell type-dependent approach could enhance effectiveness of cancer therapies.