S11 - Engineering an omentum-on-a-chip for studying adiposopathy in ovarian cancer metastasis

Introduction: : Considered a silent killer in women, ovarian cancer (OC) is characterized by its nonspecific symptoms that present late in the disease course, which results in the disease most often diagnosed after metastasis has already occurred. OC preferentially metastasizes to the omentum, and patient 5-year survival drops 50% once OC has spread to this organ. The omentum has a complex tumor microenvironment (TME), predominantly made up of highly vascularized visceral white adipose tissue overlaid with a layer of mesothelial cells on an extracellular matrix made predominantly of collagen 1 (Fig A). Currently, limited physiologically-relevant models exist to study the diverse TME of the omentum in OC metastasis. It is also thought that adiposopathy, or the presence of metabolically dysfunctional adipose tissue found in cases of clinical obesity, may act as a catalyst for metastasis to the omentum in OC. Accurate modeling of the omentum can help us understand if and how adiposopathy contributes to increased metastatic tumor burden in OC patients, which would have significant consequences for patient survival. Therefore, our overall objective is to accurately recapitulate the TME of OC metastasis by creating an omentum-on-a-chip model that includes relevant biological components under conditions of inflammation stemming from metabolically dysfunctional fat. In the absence of these models, elucidating the mechanisms behind OC metastasis - especially as related to adipose cell dysfunction - and improving patient prognosis at advanced stage will continue to be challenging.

Materials and

Methods: : A novel tri-compartmentalized omentum-on-a-chip was designed using K-Layout to mimic the layers of the omentum (Fig B). The device was manufactured using SU-8 photolithography and polydimethylsiloxane (PDMS). Microfluidic chips were pre-coated with poly-L-lysine to allow for increased cellular adhesion. The central chamber of the device was filled with a mixture of adipocytes and human umbilical vein endothelial cells (HUVECs) in a patient-derived physiologically-relevant 3D matrix based on the crosslinking of human plasma fibrinogen in combination with collagen. Adipocytes were obtained by differentiating a novel immortalized patient-derived preadipocyte cell line (DPAD) in vitro, whereby cells were given adipogenic induction media for 14 days in our 3D matrix and then extracted with collagenase. HUVECs were pre-exposed to Vasculife media in 2D cultures for 7 days in advance of being seeded into the microfluidic device to induce blood vessel-like behavior. Both mature adipocytes and HUVECs were injected into the microfluidic chip using the same 3D matrix to support their co-culture and with a specialized maintenance media tailored to maintain vascularized adipose tissue health. Met5a mesothelial cells were then seeded as 2D layer into the top chamber. Adiposopathic adipocytes were engineered by exposing mature adipocytes to the inflammatory cytokine Tumor Necrosis Factor (TNF)-ɑ for 24 hours following 14 days of differentiation (Fig C). qPCR was conducted to ensure genetic reprogramming of adisopathic traits. Confocal imaging of microfluidic chips was used to analyze differences between adiposopathic vs. healthy adipocytes in the multi-culture model with BODIPY505 staining used to mark neutral lipids.

Results, Conclusions, and Discussions:: Adipocytes, HUVECs, and mesothelial cells were successfully cultured in our microfluidic chip to create an omentum-on-a-chip model representative of the TME in OC metastasis. An epithelial-like layer of Met5a cells was formed in the microfluidic chip in the top chamber while close-up imaging of the central chamber demonstrates vessel-like structures formed by the HUVECs intermingled with mature adipocytes marked by the presence of lipid droplets (stained by BODIPY505) in a 3D matrix (Fig D). qPCR of adiposopathic adipocytes revealed significantly decreased expression of SLC2A (GLUT4) and adiponectin genes amongst exposed cultures when compared to control, aligning with traits of metabolic dysfunction in patients classified as being clinically obese (Fig E). When the multi-culture omentum-on-a-chip was analyzed with healthy adipocytes vs reprogrammed adipocytes, we saw significant phenotypic changes, particularly with regards to lipid droplet formation and accumulation. Notably, BODIPY505 staining showed a significant upregulation in the integrated density when normalized to cell count, indicating an increased amount of lipids in the adiposopathic omental model when compared to control (Fig F). This seems to be due to an increased secretion of lipids by adiposopathic fat cells within the tri-culture model. Critically, these results demonstrate that adiposopathy may play a role in promoting OC metastasis, potentially by causing changes in lipid secretion, which other studies have linked to being a source of fuel for OC cells. Future studies will study OC infiltration and metastatic potential using this model. In conclusion, we have recapitulated the TME of the omentum with three cell types in a compartmentalized, multi-niche microfluidic chip, creating a patient-derived physiologically-relevant-model for studying OC metastasis. This model will have a significant positive impact by allowing for further study of the cellular and extracellular processes that characterize OC metastasis, and particularly the effects of adiposopathy in the disease process, with great relevance for patients classified as having clinical obesity. By studying the complex nature of the omental TME, we hope to identify therapeutic targets unique to metastatic disease that will allow for improved treatment options for women diagnosed with OC, especially for the large majority who are diagnosed late in the disease process.

Acknowledgements and/or References (Optional):: This project is supported by Sanford Research and the University Of South Dakota Sanford School of Medicine’s Medical Student Research Program grant.