Optimization and Validation of a Fat-On-A-Chip Model for Non-invasive Therapeutic Drug Discovery

Introduction: Obesity is a significant public health concern that is closely associated with various comorbidities such as heart disease, stroke, type II diabetes, and certain cancers. It is defined as abnormal fat accumulation that is a risk to the patient's health and is a chronic disease that accounts for approximately 2.8 million deaths annually. Due to the central role of adipose tissue in many disease etiologies and the pervasive nature in the body, engineered adipose tissue models are essential for drug discovery, disease progression, and personalized medicine. While adipose tissue models exist ex vivo, many models fail to recapitulate human physiology due to the use of 2D environments and stem cell lines. Unlike 2D environments, 3D environments mimic the structure of the extracellular matrix (ECM) and promote normal physiological function. Additionally, use of stem cell lines allows for repeatable and reproducible methods, but often does not display the same phenotype and function as mature adipocytes.

Our fat-on-a-chip (FOAC) model uses mature human adipocytes in a Micronit perfusion device and introduces a novel approach for collecting continuous data for FOAC models by using two non-invasive readout techniques, resazurin and glucose uptake. Using primary cells is essential as patient demographics such as age, gender, ethnicity, and medical history are known to affect the phenotype and behavior of adipose tissue. The Micronit system creates a consistent FOAC model that uses a hyaluronic acid (HA) hydrogel to drive adipocyte size.

Materials and

Methods: The Micronit device creates a 3D microenvironment that has two distinct compartments, one containing human umbilical vein endothelial cells (HUVECs) in-line with the fluid flow and the other containing adipocytes suspended in a hydrogel protected from shear stresses (Figure 1). A porous membrane separates the two compartments but allows for signaling between the two cell types. Mature adipocytes were isolated from primary human adipose tissue obtained from panniculectomies, suspended in a hyaluronic acid hydrogel, and seeded onto one side of the chip. The FOAC platform was perfused with adipocyte maintenance media (DMEM, 10% FBS, and 1% Pen/Strep) with a flow rate of 20 μl/hr for adipocytes and 100 μl/hr for HUVECs.

Viability was assessed with a LIVE/DEAD™ Viability/Cytotoxicity Kit resazurin assays. In DMEM, 0.1 mM resazurin was perfused through the Micronit chips and collected every 24 hours and a plate reader was used to measure the absorbance at 560/590 nm. Metabolic activity was measured every day of culture until trial completion. For glucose uptake experiments, effluent media was collected every 24 hours and glucose concentrations were determined using the GlucCellTM Glucose Monitoring System. Samples were perfused with either adipocyte maintenance media or adipocyte maintenance media supplemented with 100 nM insulin. Triglyceride and DNA Content were quantified using the EnzyChrom Triglyceride Assay Kit and Quant-iT Picogreen DNA Assay Kit, respectively. Lastly, the morphology of the cells were assessed using ImageJ through fixing and staining with BODIPY, DAPI, and phalloidin 555.

Results, Conclusions, and Discussions: Results and Discussion
First, we sought to evaluate the consistency of adipocyte sizes in our platform. Three patients with different demographics were assessed (Figure 2A) where excised tissue was imaged and quantified (Figure 2B-E). Interestingly, once the adipocytes were seeded in a homogeneous HA hydrogel (Figure 2F-I) their unilocular lipid diameter was consistent across patients, despite their initial starting variability. Adipocyte cell size is known to vary with ECM composition where enhanced collagen deposition restricts adipocyte size. This supports the concept that adipocyte size is highly dependent on ECM stiffness, as cells adapted rapidly to their new hydrogel environment. This result also supports the reproducibility of our platform for culturing human primary adipocytes.

To validate that resazurin can be used on a perfusion device, the number of cells was incrementally increased to evaluate the sensitivity of the signal in the effluent media. As the volume of adipocytes increases, the relative fluorescence signal proportionally increased (Figure 3A). This shows resazurin can be used as a live cell assay indicator in FOAC systems. Further, the viability of cells was quantified with LIVE/DEAD (Figure 3B). This resulted in ~95% adipocyte viability on the final day of culture.

Lastly, the functionality of the adipose platform was evaluated through measuring glucose uptake in response to insulin (Figure 4). While both groups display a decrease in the amount of glucose consumed throughout the 10-day trial, the insulin-perfused adipocytes consistently exhibited a higher average glucose uptake compared to the non-insulin control group at each timepoint. These findings suggest that our microfluidic system successfully recapitulated in vivo functionality by demonstrating an increase in glucose uptake in response to insulin perfusion.

Conclusion
This work validates the use of two non-invasive readout techniques, resazurin and glucose uptake, to obtain continuous data for adipose tissue FOAC models. The Micronit platform proved to be a reproducible model that effectively controls adipocyte patient variability while measuring metabolic activity and functionality. These non-invasive readouts and the ability to mimic in vivo responses will contribute to advancing the understanding of dynamic tissue functions and can provide valuable tools for drug delivery and personalized medicine.

Acknowledgements (Optional):