Cancer associated fibroblast subtype heterogeneity in colorectal cancer influences primary tumor invasion

The influence of the tumor microenvironment (TME) on cancer progression has become a major area of study, driven in part by the continual advances in single-cell and spatial technologies that reveal the TME to be highly heterogeneous. Cancer associated fibroblasts (CAFs), a major component of the tumor stroma, have been classified into several subtypes, each with distinct functionality and influences on the tumor progression. Among these, inflammatory CAFs (iCAFs) and myofibroblastic CAFs (myCAFs) are typically defined by characteristic biomarker profiles and have been shown to have distinct spatial distributions within the tumor. However, the specific functions of these subtypes in contributing to cancer progression, particularly metastasis, is largely unknown. With metastatic disease having poor prognoses, it is vital to elucidate the roles these CAF subtypes play within the metastatic cascade. To address this gap, we have built a diverse biobank of CRC patient-derived tissue samples and matched CAF lines. Single cell RNA-sequencing of patient tissues identified distinct CAF subpopulations. These transcriptional signatures were then applied to bulk RNA-sequencing of the patient-derived CAF lines, which revealed expression patterns consistent with iCAF and myCAF profiles. Flow cytometry and secretome analysis further confirmed expression of subtype-specific markers, suggesting that each CAF line retains features of the originating subtypes.To investigate how these different CAFs influence tumor cell invasion, we developed a CRC organ-on-chip co-culture model. Specifically, this microfluidic device simulates gut-relevant biomechanical cues, including fluid flow and peristalsis-like motion, and is composed of two extracellular matrix-coated overlapping channels separated by a porous membrane, allowing biochemical and cellular exchanges between compartments. Patient-derived CAFs and GFP-labelled CRC tumor cells are cultured together in the top “epithelial” channel, while the endothelial cells line the bottom “vascular” channel. Using time-lapse confocal microscopy, GFP+ tumor cells intravasating into the vascular channel are tracked and quantified, providing real time invasion rates. We observed patient-specific differences in CAF-induced CRC invasion when exposing human CRC cells on-chip to CAF-conditioned media and in physical co-cultures. Interestingly, a myCAF-enriched line promoted invasion in only physical co-culture conditions, suggesting contact-dependent signaling. Additionally, each CAF line displayed distinct morphology and invasion patterns on-chip with the myCAF-enriched line invading before the CRC cells. Collectively, these results suggest that there may be specific CAF phenotypes that induce tumor invasion which could give insight into determining the invasive potential of a primary CRC tumor at diagnosis.