A composite hydrogel platform for the dissection of tumor cell migration at tissue interfaces

Abstract

Glioblastoma multiforme (GBM), the most prevalent primary brain cancer, is characterized by diffuse infiltration of tumor cells into brain tissue, which severely complicates surgical resection and contributes to tumor recurrence. The most rapid mode of tissue infiltration occurs along blood vessels or white matter tracts, which represent topological interfaces thought to serve as “tracks” that speed cell migration. Despite this observation, the field lacks experimental paradigms that capture key features of these tissue interfaces and allow reductionist dissection of mechanisms of this interfacial motility. To address this need, we developed a culture system in which tumor cells are sandwiched between a fibronectin-coated ventral surface representing vascular basement membrane and a dorsal hyaluronic acid (HA) surface representing brain parenchyma. We find that inclusion of the dorsal HA surface induces formation of adhesive complexes and significantly slows cell migration relative to a free fibronectin-coated surface. This retardation is amplified by inclusion of integrin binding peptides in the dorsal layer and expression of CD44, suggesting that the dorsal surface slows migration through biochemically specific mechanisms rather than simple steric hindrance. Moreover, both the reduction in migration speed and assembly of dorsal adhesions depend on myosin activation and the stiffness of the ventral layer, implying that mechanochemical feedback directed by the ventral layer can influence adhesive signaling at the dorsal surface.

Introduction

Cell migration and the mechanisms that underlie specific migratory phenotypes are increasingly recognized to depend on extracellular context, especially the structure and mechanics of the extracellular matrix (ECM) [1], [2], [3]. On planar two-dimensional substrates, migration is typically described as being driven by a balance between actin polymerization at the cell front and actomyosin contraction at the cell rear that is transmitted to the ECM via adhesions [4]. In three-dimensional ECMs, migration can take various forms including mesenchymal migration (perhaps most analogous to classical two-dimensional migration) to amoeboid migration, which is less adhesion-dependent and leverages intracellular hydrostatic pressure generated by actomyosin contractility to extrude the cell body through matrix pores [5]. Importantly, the molecular mechanisms that control these migration modes are as diverse as the number of migratory phenotypes. In fact, many cells dynamically switch from one mode to another as they encounter and navigate different microenvironments, highlighting the importance of studying cell migration in culture systems that capture defining architectural features of tissue [6], [7], [8].

Cell migration is often guided by heterogeneous structures within the ECM; for example, a diverse variety of invasive solid tumors proceed along pre-existing anatomical structures [9], [10], [11], [12]. Metastastatic tumor cells have been clinically observed to preferentially migrate in bone cavities or between adipocytes, suggesting that the topographies of these structures may facilitate tissue dissemination [10]. Migration in this context may be regarded as being “interfacial” in nature, in that cells translocate along a ventral two-dimensional surface while surrounded on their dorsolateral surface by an amorphous ECM of a different composition. Other examples of interfacial migration are tumor cells that migrate between bundles of myelinated axons and connective brain tissue [10], [13].

A particularly important example of interfacial migration is the invasion of glioblastoma multiforme (GBM), the most common and deadly primary brain tumor. The extreme lethality of this malignancy is attributed in part to its diffuse and unrelenting infiltration of brain tissue, effectively precluding complete surgical resection [14]. GBM invasion patterns are unlike most other aggressive malignancies, in that GBM cells rarely intravasate and metastasize to distant tissues, instead remaining within the brain [14], [15]. The pre-existing structures that guide GBM, collectively known as the secondary structures of Scherer, include the subpial space, white matter tracts, and vascular beds [16]. While these structures are widely acknowledged to facilitate invasive migration, relatively little is known about the biophysical and molecular mechanisms through which they do so. For example, cells migrating along vascular beds simultaneously experience strong integrin-based inputs via fibronectin and laminin in the vascular basement membrane [15] while also receiving adhesive inputs from hyaluronic acid (HA) in the brain parenchyma, which can be mediated by HA receptors such as CD44 and RHAMM [17], [18]. There are also substantial biophysical asymmetries within this adhesive microenvironment, as vascular beds tend to be orders of magnitude stiffer than the surrounding parenchyma [19], [20], [21]. How these asymmetric signals are integrated to regulate migration in GBM remains unknown.

Despite the acknowledged importance of migration along asymmetric tissue interfaces in many tumors, comparatively little is known about the molecular mechanisms that underlie this process. The fact that migration mechanisms depend strongly on context has created an unmet need for experimental paradigms that recapitulate key aspects of these interfaces. To address this need, we developed a simple experimental system that features asymmetric ECM signals representative of the brain parenchyma–vascular interface, and used it to investigate molecular mechanisms of adhesion and motility.