Biophysical Properties of Hyaluronic Acid Differentially Regulate Breast Cancer Invasion In Single Cells vs. Spheroids

Introduction:: One in eight women will be diagnosed with breast cancer over the course of their lifetime. Survival rates are high when the tumor is confined to the breast; however, they are drastically reduced when the tumor spreads regionally or metastasizes to a distal site. Thus, understanding invasion, the first step in cancer dissemination, is critical. Aberrant accumulation of hyaluronic acid (HA) within the breast microenvironment is associated with worse patient prognosis. However, therapeutics designed to degrade HA have failed in clinical trials. Therefore, the mechanisms connecting HA to cancer invasion require further elucidation. HA is a linear hydrophilic glycosaminoglycan regulating key processes (e.g. during development) via its osmotic pressure. HA is synthesized by Hyaluronic Acid Synthases (HAS), and overexpression of HAS2 leads to increased invasion of single cells. However, in primary tumors cells initially engage in 3D cell-cell interactions. To examine how HAS2 regulates invasion from a microtumor and which role local accumulation of HA plays in this process, we have developed spheroids of breast epithelial cells overexpressing HAS2. These spheroids were embedded into a model of the breast stroma consisting of fibrillar type I collagen hydrogels and adipose stromal cells. Combining our models with advanced imaging techniques, inhibition studies, and biological assays allowed previously undiscovered insights into how HA regulates invasion during initial stages of tumorigenesis.


Materials and Methods:: HAS2-overexpressing MCF10A cells were generated via lentiviral transduction and characterized for HA production by HA ELISA and immunofluorescence staining. Phenotypic differences induced by HAS2 overexpression were analyzed through RNA sequencing and single cell migration and invasion assays. To study 3D invasion in the context of a hyperplastic breast lesion, monoculture spheroids of parental or HAS2 MCF10A cells were embedded into fibrillar collagen hydrogels designed to recapitulate the architecture and mechanical properties of the interstitial breast stroma. Brightfield imaging was used to determine size changes of the spheroids, while confocal imaging of fixed samples was used to analyze spheroid morphology and collagen remodeling. Optical Coherence Tomography was used to visualize the structure of spheroid lumens and surrounding collagen matrix at greater depths. Blebbistatin and Batimastat were used to inhibit cell contractility and MMP activity respectively. Spatially resolved multiphoton laser ablation of points in the outer cell layer of live spheroids combined with time lapse imaging of labeled cells allows for an analysis of intra-spheroidal pressure and movement of cells from the spheroid into the matrix on a second to minute level time scale. Co-culture studies with adipose stromal cells (ADSC; Lonza) were performed to analyze invasion behavior of HAS overexpressing epithelial cells in the presence of matrix remodeling stromal cells.


Results, Conclusions, and Discussions:: Contrary to invasion studies performed with single tumor cells, HAS2 overexpression did not promote invasion from spheroids. Instead HAS2 overexpression promoted proliferation-independent spheroid expansion when compared to parental controls (A). OCT imaging showed lumen formation in the HAS2 spheroids (B) and ELISA analysis of the internal fluid of lysed spheroids indicated high levels of HA within the lumens (D) suggesting HA accumulation drives lumen formation and spheroid expansion. Confocal reflectance imaging of the spheroid-collagen interface revealed a dense band of collagen with respectively aligned cells (C) surrounding HAS2 but not parental spheroids. We hypothesized that HA accumulation inside the spheroid drives spheroid expansion, elevates intra-spheroidal pressure, and compresses the collagen matrix surrounding the spheroid to form a physical barrier to invasion. Laser ablating a microtrack into the collagen hydrogel and puncturing the outer cell layer in HAS2 spheroids caused immediate expulsion of single cells and fluid from the spheroid into the track (E), a phenomenon not observed with parental spheroids. This supports our hypothesis that HA accumulation caused elevated pressure within the spheroid. Breaching the collagen ring with ablation allowed for invasion into the microtracks after 24 hours (E) confirming the collagen barrier was preventing invasion. Parental cells invaded in a dense mass while HAS2 cells formed hollow structures, similar to their characteristic lumens (E). Co-culture with matrix remodeling stromal cells promoted invasion of HAS2 overexpressing MCF10As (F). Treating co-cultures with Batimastat prevented invasion, suggesting ADSCs promote invasion by enzymatically degrading the collagen. Indeed, collagen reflectance imaging showed the collagen barrier is not formed in untreated co-culture spheroids (F). In summary, these results demonstrate the biophysical properties of HA regulate tumor invasion in a microenvironment-dependent context. While HAS overexpression promotes single cell invasion, our data suggests that HA limits invasion of 3D aggregates of tumor cells through promoting spheroid expansion and collagen compaction in the surrounding matrix, a finding which may illuminate the clinical failure of HA degrading therapeutics. Our finding that stromal cells perturb collagen barrier formation through enzymatically degrading the matrix to permit tumor cell invasion, suggests that future therapeutic development should consider combined stroma and HA-targeted approaches.

Acknowledgements (Optional): :

References (Optional): :