Figure 1.
Schematic and diagrams illustrating device features.
The oxygen insert device is fabricated by conventional photolithography (microfluidic network), replica molding (microfluidic network and insert scaffold), and defined spinning of PDMS (gas-permeable membrane). A) The oxygen device nested into a 6-well plate. B) Examples of 24 and 96-well pillar arrays. C) A cross-sectional schematic of a pillar. Oxygen flows into the device through the inlet and travels across a microfluidic network at the bottom of the pillar. Oxygen can freely diffuse across the gas-permeable PDMS membrane at the bottom of the pillar and dissolve into the culture media. D) A macroscope image showing the various features of a single-channel pillar from above, with bonded glass posts for the equilibration studies.
Figure 2.
Deflection of PDMS membranes of varying thicknesses.
Deflection distance was inversely proportional to membrane thickness, with the 100 µm membrane deflecting the most and the 230 µm membrane deflecting the least. The data indicates that deflection of the 100 µm is negligible for operational flow rates (∼25 mL/min) relative to the overall gap distance between the membrane and cells.
Figure 3.
Validation of the device with oxygen sensors.
Oxygen tension within each well was characterized using a planar ruthenium oxygen sensor. All oxygen mixtures contained balanced nitrogen and 5% CO2 for media buffering. A) Plot illustrating the effect of post height, and thus oxygen diffusion distance between the membrane and cells, on the equilibration time and effectiveness. Heights were established by cut-glass posts bound to the bottom of the device. All three post sizes yield equilibration times much improved over the hypoxic chamber. Note that time is on a log scale. B) Plot depicting the rapid oxygen equilibration response time of the 0.2 mm gap device. C) Multi-position linescans were also taken across the well under the microchannel to ensure homogeneity of the oxygen concentration introduced by the device. Graph depicts the oxygen concentration measured after infusing 0%, 10%, and 21% oxygen for 10 min. D) Device effectively maintains 10% oxygen over 5 days.
Figure 4.
Experimentation with more complex oxygen microchannel designs.
A) Dual-condition microchannel setup yields a stable 0% and 21% oxygen profile over 14 days. B) An interdigitated and winding pattern of 500 µm width microchannels extending across the pillar results in a cyclic oxygen profile. Note that the data only depicts one representative trial as microchannel alignment was difficult.
Figure 5.
Common sterilization methods do not alter device's ability to deliver oxygen.
An important aspect of any cell culture tool is the ability to sterilize it without the loss of function. Four separate devices were exposed to each of the commonly available sterilization methods over 32 cycles, with an untreated control. The oxygen concentration at the bottom of the well was measured with the fluorescent oxygen probe with infused 10% O2 on cycle 1, 2, 8, 16, 24, and 32. No sterilization method had a significant effect on device function compared to control.
Figure 6.
Validation of the device with cellular models.
A) Human adult dermal fibroblast cells were used as a cellular model to study proliferation in response to oxygen concentrations established by our device. Cells were seeded in the wells of a 6-well plate and cultured to confluency. The device was removed from the plate at 24, 48, 72, and 96 h to briefly exchange the culture media and to take images of the cells for later cell counting. The plot depicts the proliferation of cells. Data is presented as cell number normalized to baseline (day 0) over the course of 4 days for cells continually exposed to 0%, 10%, and 21% O2. Proliferation was highest in 10% O2 (as expected). No cells survived after 4 days in 0% infused oxygen. B) Human osteosarcoma cells (U2OS) were seeded in a 6-well plate to ∼85% confluency and exposed to hypoxia and normoxia for 24 h using the hypoxic insert device. Control cell media was supplemented with 10 µM CoCl2, which mimics the hypoxic state and results in HIF-1α upregulation [27]. Cells cultured in atmospheric oxygen without exposure to the device served as the device control (DC). Western analysis was used to qualify regulation of HIF-1α expression. The figure depicts upregulation of HIF-1α in cells exposed to hypoxia using the device, relative to the no device control and 21% oxygen. The device 0% oxygen band intensity is comparable to the CoCl2 control.