In Vitro Angiogenesis: Co-Culture Model of the Retina in Microfluidic Devices

Abstract

Background Age-related macular disease (AMD) is a common eye condition and the most severe form of a disease that is the leading cause of vision loss among people age 50 and above. Simple tasks such as driving or reading become impossible without a functional central vision, which is impaired as the result of damages to the macula. Wet form AMD is responsible for 90% of the blindness out of all AMD patients; it is marked by the development of subretinal choroidal neovascularization and degeneration of retinal pigmental epithelial (RPE) cells. Wet macular degeneration occurs when abnormal blood vessels disrupt the retina and leak blood or fluid that blurs the central vision. Recently, MEMS technology which integrates multidisciplinary field of studies offers a new direction in medical investigation and diagnosis. Here we utilized microfabrication technology to reconstruct ocular fundus tissues (OF) in vitro and to simulate neovascularization in the hope of building a disease model for the retina. Method Human retinal pigment epithelial (ARPE-19) and human umbilical vein endothelial cells (HUVECs), representing the retinal cells and the choroid, were co-cultured in the microfluidic devices. Two different approaches are developed in fabricating microfluidic devices resulting in 2D and 3D models. 2D model:the microfluidic device was made up of 2 layers, on each a channel (width 500 µm; height 150 µm) was embedded. A porous membrane (thickness 6.5 µm, hole diameter 10 µm) was sandwiched between the layers. Microchannels and membranes were made of poly(dimethylsiloxane) (PDMS). Retinal pigment epithelial (RPE) cells were introduced into the upper channel of the device and umbilical vein endothelial cells (HUVECs) were loaded into the lower channel. 3D model: the wire with a 0.2 mm diameter was pretreated before embedded in collagen (6.5 mg/ml). After the collagen had polymerized, the needle was removed resulting in an empty channel where HUVECs were to be seeded. RPE cells were seeded on the top of the collagen gel to initiate co-culture experiments. Data analysis: RPE cells and HUVEC were dyed respectively with Cell TrackerTM red and green and cell growth was recorded at certain time intervals. RPE cells were under hypoxia and hypoglycemia 24hr prior the start of the co-culture experiment. Before the aforementioned co-culture model, RPE were cultured alone on the upper channel and the VEGF concentration was collected and measured by ELIZA after replacing DME medium High Glucose (20 mM) with DME medium Low Glucose (3 mM). HUVECs were loaded into the lower channel and cultured alone. The upper channel was perfused with 50 ng/ml of VEGF-A165and the number of HUVECs migrating to the upper channel was analyzed using the same monitoring system. Characteristics of cells in the 3D model were characterized following the same methodology. Results Validation of the devices: The devices were validated prior culturing cells inside the devices. Confocal images of the cross-sections of channel length and diameter were taken after perfusion of fluorescent microbeads. ARPE characterization: Results show that RPE cells responded to the lowered glucose microenvironment by increasing VEGF secretion (58.8 % compared to the control). In addition, RPE cells exhibited a polar secretion of VEGF which is consistent to the cell physiology. HUVEC characterization: The increased number of HUVEC migrating through the porous membrane in response to VEGF gradient signifies neovascularization. Comparing to the control, there was a 31.3 % increase of cell migration. Conclusions and Future work The physiology of RPE cells and HUVECs is reproduced in the microfluidic devices. The preliminary results suggest that the alteration of the microenvironment induces responses of RPE cells and the significance of VEGF in neovascularization. In the near future, we anticipate studying responses of the co-culture model to hypoxia to characterize the AMD pathology.