Microfluidics: Doing More with Less - Part 3

This is Part 3 of a three-part series on biological applications of microfluidic devices. Part 1 covered the history, physics, and popular fabrication methods of microfluidic devices, while Part 2 discussed the application of microfluidic devices in low resource and point-of-care applications. In this part we will focus on the role of microfluidic devices in cutting edge technologies.

During the preclinical drug discovery process therapies are tested at multiple different levels to ensure that a drug is both safe and effective. Initially molecular level screening is carried out to ensure that the candidate drugs can effectively interact with the target of interest. For example, if the overactivation of a specific receptor is known to cause a disease, a large number of small molecules may be tested against the isolated receptor to determine which of these molecules most effectively inhibit the receptor activation. This allows researchers to quickly eliminate compounds that do not interact with the target of interest and determine a pool of promising compounds to conduct further studies using in vitro and in vivo methods. In vitro studies use live cells maintained in conditions designed to replicate disease states or mimic the underlying mechanisms of the disease. Candidate compounds identified at the molecular level are tested in these in vitro models to characterize not only the potential efficacy, but also the safety of the compounds. While in vitro studies are well suited to study the effects of potential therapeutics on the primary cell types, they are not capable of capturing the complexity of biological systems. It may be that a candidate compound effectively treats the disease state in the primary cell type, but also significantly disrupts interactions between other cell types necessary for organ function. That is why following in vitro studies in vivo studies using animal models are necessary. These studies aim to identify any potential issues with safety, toxicity, and effectiveness at the organism level. However, the success of a drug candidate following in vivo animal models does not ensure it will remain safe and effective for humans as there are significant physiological differences between animals and humans. Therefore, at the end of the preclinical drug discovery process, successful drug candidates are then extensively characterized in human clinical trials before being approved for use. During each step of this process the cost of each study per drug candidate rises exponentially, with human clinical trials estimated to cost 10s of millions of dollars. Therefore, improving the success rate of candidate drugs identified during the drug discovery process is a critical factor to reduce drug costs (currently less than 10% of drugs that enter clinical trials make it to market). Microfluidic technologies have shown promise in improving both the compound screening process and improving in vitro models by increasing their physiological relevance. 

Droplet-Based Microfluidics for Drug Discovery

One of the major limiting factors in the initial candidate screening process is the limit of detection of macroscale detection methods. This is especially true for antibody treatments such as the ones targeting COVID-19. To produce the screening library of potential therapeutic antibodies, laboratory animals (typically mice) are injected with the target of interest. Antibody producing cells from the animal are then isolated and combined with another cell to produce an immortal cell line that continuously produces antibodies. Each of these cells secrete a specific antibody with different binding efficiencies to the target of interest. In the macroscale screening process individual cells are placed in a multi-well plate and cultured for several weeks to allow enough time for the cells to divide and produce enough antibodies for macroscale detection methods. This extended culture time limits the number of cells that can be screened (typically less than 1 in 1000), meaning that many potential antibodies are not included in the screening library.

Droplet-based microfluidic devices provide a promising alternative to macroscale screening processes that increases the screening throughput and eliminates the need to allow the antibody producing cells to expand. In these devices two immiscible fluids (fluids that do not mix such as oil and water) are used to generate nL to pL scale droplets. One fluid (oil) is flowed through the main channel while the other fluid (water) is flowed into the main channel to form droplets. By changing the flow rates and geometry of the channels droplets of different sizes can be formed. To screen for antibodies specific to the target of interest, the same method is used to produce the immobilized antibody producing cells. A very dilute suspension of these cells are then flowed into the droplet-based microfluidic device to produce pL volume droplets containing individual cells. Because the droplet has such a small volume, the antibodies produced by the individual cell are highly concentrated, eliminating the need to culture the cells. These droplets are then combined with another droplet containing the target of interest along with a fluorescent indicator that can quantify the binding affinity of the antibodies. This produces an extremely high throughput screening method without the need for the extended culture times.

Microphysiological Systems

While in vivo animal models are critical to understanding the efficacy and safety of preclinical drug candidates, the significant physiological differences between animals and humans leads to high clinical trial failure rates for even the most promising preclinical compounds. Additionally, with the increased emphasis in precision medicine and gene editing, animal models may no longer be capable of accurately predicting the safety and efficacy of candidate drugs. For example, we discovered that COVID-19 does not infect mice, the most common animal model). In order for SARS-CoV-2 to infect a cell it must first bind to a protein called ACE2, however the ACE2 protein natively expressed by mice is different than the ACE2 protein expressed by humans, and thus the SARS-CoV-2 could not bind to the mouse version of ACE2. Eventually scientists were able to create mouse models that expressed human ACE2 and were able to be infected by SARS-CoV-2, however this is not the case for every disease. This is especially true for neurodegenerative diseases, as even genetically modified animal models that produce proteins associated with the disease (such as amyloid-beta for Alzheimer’s disease) do not display the same neurodegeneration and observed disease progression seen in humans. 

One method to overcome this issue is the use of microfluidic based microphysiological systems (also known as organ-on-a-chip models). As the name suggests, these systems attempt to recreate the microscale architecture or organs or tissues, and allow cells cultured in these devices to recapitulate the 3D architecture and cell-cell interactions found in the native tissue. PDMS based microfluidics are well suited for this application as they are biocompatible, gas permeable, flexible, and can be fabricated with subcellular scale features. Furthermore these devices can be seeded with human cells that can improve the physiological relevance over animal models. One of the most famous examples of this is the “lung-on-a-chip” developed by a group at Harvard in 2010. This PDMS based device consisted of two main microchannels separated by a porous membrane that were flanked by two additional channels. On one side of the main channel the researchers cultured lung alveolar epithelium cells, while on the other side of the porous membrane they cultured blood vessel endothelial cells mimicking the structure of an alveoli and associated capillary. The two microchannels on the side were connected to vacuum lines that stretched the main microchannels imitating the expansion and contraction of alveoli during breathing. The researchers were able to use this device to observe the blood-borne immune response to the respiration of toxic nanoparticles. This design has been extremely successful and has been adapted to other organs with similar architecture to create “gut-on-a-chip,” “kidney-on-a-chip,” and “liver-on-a-chip” systems.

References

Dowden, Helen, and Jamie Munro. "Trends in clinical success rates and therapeutic focus." Nat. Rev. Drug Discov 18, no. 7 (2019): 495-6.

Huh, Dongeun, Benjamin D. Matthews, Akiko Mammoto, Martín Montoya-Zavala, Hong Yuan Hsin, and Donald E. Ingber. "Reconstituting organ-level lung functions on a chip." Science 328, no. 5986 (2010): 1662-1668.

Lu, Ruei-Min, Yu-Chyi Hwang, I-Ju Liu, Chi-Chiu Lee, Han-Zen Tsai, Hsin-Jung Li, and Han-Chung Wu. "Development of therapeutic antibodies for the treatment of diseases." Journal of biomedical science 27, no. 1 (2020): 1-30.

Paul, Steven M., Daniel S. Mytelka, Christopher T. Dunwiddie, Charles C. Persinger, Bernard H. Munos, Stacy R. Lindborg, and Aaron L. Schacht. "How to improve R&D productivity: the pharmaceutical industry's grand challenge." Nature reviews Drug discovery 9, no. 3 (2010): 203-214.

Shembekar, Nachiket, Chawaree Chaipan, Ramesh Utharala, and Christoph A. Merten. "Droplet-based microfluidics in drug discovery, transcriptomics and high-throughput molecular genetics." Lab on a Chip 16, no. 8 (2016): 1314-1331.