Development of a Microfluidic Array to Study Drug Response in Breast Cancer

Abstract

Luminal geometries are common structures in biology, which are challenging to mimic using conventional in vitro techniques based on the use of Petri dishes. In this context, microfluidic systems can mimic the lumen geometry, enabling a large variety of studies. However, most microfluidic models still rely on polydimethylsiloxane (PDMS), a material that is not amenable for high-throughput fabrication and presents some limitations compared with other materials such as polystyrene. Thus, we have developed a microfluidic device array to generate multiple bio-relevant luminal structures utilizing polystyrene and micro-milling. This platform offers a scalable alternative to conventional microfluidic devices designed in PDMS. Additionally, the use of polystyrene has well described advantages, such as lower permeability to hydrophobic molecules compared with PDMS, while maintaining excellent viability and optical properties. Breast cancer cells cultured in the devices exhibited high cell viability similar to PDMS-based microdevices. Further, co-culture experiments with different breast cell types showed the potential of the model to study breast cancer invasion. Finally, we demonstrated the potential of the microfluidic array for drug screening, testing chemotherapy drugs and photodynamic therapy agents for breast cancer.

Keywords: lumen; microfluidics; polystyrene.

Figure 1

Design and operation of the microfluidic array. (A) Top-down schematic view of the microfluidic array. (B) Top-down view of one of the devices. (C) Inset and magnification of one microdevice in the microfluidic array with marked cross-sections in (D) (pink cross-section) and (E) (blue cross-section). (F) The microfluidic array is milled on both sides of the polystyrene (PS) piece, and a glass coverslip is used to seal the lumen side with cell culture-grade double sided Advanced Research tape. (G) Each line of devices is sealed from adjacent lines, allowing for testing of multiple doses in quadruplet of target compound in one microdevice array. (H) Schematic of collagen filling and cell seeding in each of the devices in the microfluidic array.

Figure 2

Diffusion profile. (A) A Rhodamine B solution (in red) was injected into polydimethylsiloxane (PDMS)-based microdevices, and Rhodamine B diffusion into the material was monitored with real-time microscopy. Images showed Rhodamine B penetrated into the PDMS bulk material. (B) Graph shows the Rhodamine B diffusion profile. Rhodamine B fluorescence was analyzed across the delimited region (yellow rectangle). The location of the microdevice wall is denoted with dashed line. (C) Rhodamine diffusion experiment in the PS-based microfluidic array. No Rhodamine B diffused or leaked into the material. (D) Rhodamine B diffusion profile in the microfluidic array.

Figure 3

Cell migration. (A–C) confocal images showing MCD10A cells labelled with vibrant Dil (in red) and MDA-MB-231 GFP cells at 3:1 ratio after 24 h in culture. The confocal images showed no observable cell migration. (D–F) Confocal images after 48 h showing MCF10A and MDA-MB-231 invading the hydrogel. (G,H) Migration analysis. The confocal image was vertically divided in 8 columns and the area occupied by MCF10A (in red) or MDA-MB-231 GFP (in green) was analyzed, normalized, and plotted.

Figure 4

Effect of Doxorubicin on cell viability. (A) GFP⁺ MDA-MB-231 cells were cultured in the lumen. After 72 h in culture, a solution containing propidium iodide was perfused through the lumen to stain dead cells in red and cell viability was observed by confocal microscopy. GFP⁺ MDA-MB-231 cells showed high viability in the absence of doxorubicin. (B) GFP⁺ MDA-MB-231 cells were cultured for 3 days in 100 µM doxorubicin. Cell viability images showed a significant decrease in cell viability and most dead cells were removed from the lumen during the media changes and the staining process. *** p < 0.001.

Figure 5

Photodynamic therapy in the microfluidic array. (A) MDA-MB-231 green fluorescent protein (GFP) cells were injected in the lumen and incubated with 500 ng/mL verteporfin for 24 h. The right end of the lumen was exposed to 485/35 nm light for 45 s to photoactivate verteporfin. After another 24 h, cell viability was evaluated, showing a gradient of viability across the lumen. (B) Images showing the left, center, and right section of the lumen. (C) Graphs showing the normalized area under the curve of the luminescence plot for live cell (green) and dead cell (fluorescence) in the left, central, and right region of the lumen. *** p ≤ 0.001.