Assistant Professor University of Pittsburgh, Pennsylvania, United States
Introduction:: In ovarian cancer, the presence of macrophages in the tumor microenvironment (TME) has been associated with poor clinical outcomes. Understanding mechanisms of macrophage recruitment and signaling with tumor cells in the TME may uncover novel therapeutic targets to block the pro-tumor effects of macrophages and improve patient outcomes. Several tumor secreted factors, such as CSF1, CCL2, IL8, and periostin have been implicated in the recruitment of macrophages to the tumor site. We used a 3D physiologically relevant microfluidic platform that can accurately control cell-cell interactions.
Materials and Methods:: PDMS-based microfluidic devices were used and using live cell analysis we monitored the migration of macrophages (RAW264.7, primary bone-marrow derived murine macrophages and primary peripheral blood mononuclear cell derived human) under control and coculture conditions (Fig 1A). Tumor cells were seeded at a density of 2 million cells/ml in a 3D collagen type I (2.5 mg/ml) matrix. Macrophages were tracked for 96hours (dt=15min) and cell tracking trajectories were analyzed using Nikon Elements.
Results, Conclusions, and Discussions:: We first studied the abilities of a panel of high grade serous ovarian cancer models comprised of established cancer cell lines (murine ID8 and human CaOV3, OVSAHO, OVCAR8) and patient-derived high-grade serous ovarian cancer xenografts (4 PDX models) to recruit macrophages via microfluidic (Fig 1A) and collagen-droplet based 3D assays. The presence of tumor cells enhanced macrophage infiltration into collagen 2-fold after 48 hours (Fig 1B). We found that ovarian cancer models that efficiently recruited macrophages expressed higher levels of CSF, CCL2, and IL8 cytokines. Furthermore, tumor cell secreted factors more potently increased macrophage recruitment compared to chemotactic gradients of CCL2 and CSF1, indicating a complex signaling network between tumor cells and macrophages.
Treatment with BLZ945, a potent CSF1R inhibitor that is evaluated in clinical trials, effectively blocked the recruitment of the RAW264.7 macrophages towards ID8 ovarian cancer cells (Fig 1C). We subsequently characterized the attachment and invasion profiles of the human and murine ovarian cancer models in syngeneic and immunocompromised xenografts in vivo. We found that tumor cells attached and invaded the ovaries, abdominal organs, and fatty tissue. In a preliminary analysis, one of the PDX models that efficiently recruited macrophages also formed invasive implants, while another PDX model with poor macrophage infiltration did not exhibit invasive implant formation. Additionally, we studied the effect of cisplatin treatment on macrophage infiltration, revealing that cisplatin-treated ovarian cancer cells more potently recruit macrophages (Fig 1D).
This work demonstrates the capability of a microfluidic platform to model macrophage infiltration in ovarian cancer. In ongoing studies using this microfluidic platform we are investigating the effects of chemotherapy-induced changes in the tumor cell secretome on macrophage infiltration and the resulting effects on tumor cell chemotherapy sensitivity.
