Poster S7 - Profiling Adhesion and Gene Expression of Patient Circulating Tumor Cells on Decellularized Bone, Liver, and Lung Matrix

Introduction: Circulating tumor cells (CTCs) isolated from the bloodstream represent a “missing link” between primary tumors and secondary metastases, and can exhibit “tropism” towards preferred tissue sites. For example, some hormone receptor positive breast cancers metastasize preferentially to bone, while others metastasize to liver and lung, which represent profoundly different metastatic niches with varied extracellular matrix (ECM) composition. Here, we use live cell imaging and RNA-seq to profile how patient CTCs adhere to decellularized bone, liver, and lung ex vivo, for comparison with patient disease progression and xenograft models.

Materials and

Methods: Patient-derived breast cancer CTC lines were kindly provided by S. Maheswaran (Massachusetts General Hospital), isolated using microfluidic devices from blood draws of metastatic breast cancer patients with informed consent and IRB approval. CTCs were maintained in serum-free media supplemented with epidermal growth factor (EGF) and basic fibroblast growth factor (FGF) under hypoxic conditions (5% O2). CTC lines were cultured in suspension on Ultra Low Adhesion (ULA) plates or on lung, bone, or liver decellularized extracellular matrix (Xylyx), with regular live cell imaging using confocal microscopy. Cell shape was segmented and classified using custom MATLAB code. Further, CTCs were cultured on polyacrylamide hydrogel substrates of varying stiffness with embedded fluorescent microparticles for traction force microscopy, then analyzed using custom MATLAB code. After four days, CTCs were isolated using collagenase, then sent for RNA-seq at Genewiz / Azenta. Gene expression profiling was performed using over representation analysis (ORA) and gene set enrichment analysis (GSEA). The ORA provided stringent and relaxed gene lists based on a significance threshold of FDR < 0.1 and p < 0.05, respectively.

Results, Conclusions, and Discussions: CTCs typically remain rounded in suspension as weakly adherent multicellular “grape-like” aggregates (Figure 1A), but cannot be subcultured on tissue culture plastic. In comparison, CTCs exhibited striking differences on decellularized matrices, showing more dispersion with spread morphologies and protrusions (Figure 1B). We observed increased proliferation rates on matrix conditions relative to suspension. To further elucidate these differences, CTCs were collected for RNA-seq to compare suspension culture with lung, liver, and bone. Indeed, hierarchical clustering revealed CTCs from suspension were well separated from all matrix conditions. Moreover, CTC gene expression on liver and bone were more similar to each other than lung. Notably, gene set enrichment analysis revealed significant alterations in “collagen-containing extracellular matrix,” “leukocyte migration,” “epithelial cell proliferation,” “cell-cell adhesion” as well as “epithelial differentiation.” Some of these genes are consistent metastatic signatures previously established in xenograft models and human patients. Further, we observed alterations in metabolic signaling pathways for matrix vs suspension culture. We then examined whether CTCs are responsive to substrate stiffness using traction force microscopy.
We have profiled how CTC gene expression in suspension relative to decellularized matrix, showing that adhesion can drive profound differences in functional phenotype. We find that matrix conditions alone are sufficient to recapitulate some signaling pathways associated with metastasis in vivo, despite the absence of stromal and immune cell components. We will further explore CTC drug response on decellularized matrix for comparison with patient response and previous drug screens, as a proof of concept towards precision medicine.

Acknowledgements (Optional): We thank S. Maheswaran (Massachusetts General Hospital) for the gift of CTC lines (BRx-68) and funding from NIGMS under P20GM109035 and R01GM140108.