How AI-Powered Cell Culture Could Revolutionize Drug Discovery
Scientists have not embraced 3D models of human organs. AI could change that.
Complete the form below to unlock access to ALL audio articles.
The following article is an opinion piece written by Shantanu Dhamija. The views and opinions expressed in this article are those of the author and do not necessarily reflect the official position of Technology Networks.
Organoids – living, 3D multicellular models of human organs – could revolutionize early research and discovery of new medicines. Because organoids mirror human biology more closely than animal models like mice and rats, they are ideal tools for scientists who want to quickly separate out the drug candidates most likely to fail in human clinical trials from those most likely to succeed.
However, scientists have not yet fully embraced organoids in drug discovery because, quite frankly, they’re hard to use. Large-scale cell cultures such as organoids have traditionally demanded 24/7 oversight, including feedings that must be done with precision and at prescribed times, sometimes on weekends, to avoid ruining experiments. Some cell cultures also have to be “passaged,” or broken apart into smaller pieces and then re-seeded into new plates so they have enough room to grow. These hands-on procedures are time-consuming, error-prone, expensive, mundane and repetitive, making organoid studies hard to scale up and reproduce. Though individual procedures can be automated, it has been difficult to automate end-to-end organoid workflows because of the subjective nature of what an acceptable organoid culture looks like. For example, scientists rely upon visual inspection to determine if an organoid is thriving, dying, sterile, contaminated, or mature. Differences of opinion have made standardization challenging.
Artificial intelligence (AI) is changing this. AI technology is advancing so rapidly that there are already tools on the market that offer end-to-end automation of cell culture experiments. AI-powered software with built-in protocols can generate both 2D and 3D cellular models, and then manage the entire cell journey with automated feeding and passaging.
Thanks to machine learning, these tools make protocol development much more flexible and powerful, with algorithms that can trigger decision-making by either being automated or managed by a live scientist. Researchers can then leverage these fully automated workflows to generate reproducible results across multiple experiments.
AI in action
Some early adopters are already capitalizing on AI for cell modeling, including scientists who study the heart. Cardiovascular disease is the leading cause of death worldwide – claiming one life every 34 seconds in the United States alone. At the same time, unexpected cardiotoxicity has been implicated in cases where drugs have been withdrawn from the market or clinical trials of experimental medicines had to be halted.
AI-powered experiments using models that accurately reflect the human heart could serve two purposes: They could help researchers discover novel drug targets to treat heart disease and they could be used in early discovery to determine which compounds are likely to cause toxic cardiac side effects, so they can scrap those compounds and redirect resources into more promising drug candidates.
Our partner, Austria-based Heartbeat.bio, is an early believer in the potential of AI to revolutionize drug research. They have set out to create off-the-shelf cardiac toxicity organoid screening kits that combine cell culture media with AI-enabled research protocols.
Quantifying the benefits of AI in drug discovery
To quantify the benefits AI-enabled organoid culture at-scale will bring to drug discovery, it helps to understand the scale of pharma R&D. The average cost of bringing a new drug to market is $1.1 billion to $1.2 billion and rising. The low adoption of models with high predictive power plays a role in these cost increases, because of high preclinical to clinical attrition. A recent study estimated that more than $24 billion of lost R&D productivity could be recaptured by adopting models that were more predictive of clinical safety alone. Better models of efficacy would be equally impactful. AI-enabled organoid culture can transform this.
The benefits extend beyond the transformative. Cell models are demanding and require hours and hours of hands-on attention. Cells frequently need daily media changes, which requires scientists to come in on the weekends This causes employee burnout and attrition, which in turn results in lost productivity and increased hiring and training costs. This work requires precision, and even highly trained operators can make errors, let alone new associates. The high cost of consumables (cell culture media and labware), the long duration of organoid culture – ranging from 5 days to 6 months – along with the disruption of project timelines make mistakes costly.
AI also makes protocol development more efficient. By prompting decision-making throughout a cell model’s life, AI powers the real-time customization of research protocols with full traceability of what occurred during the culture run. It also enables seamless, quality-controlled collaboration within research teams through the electronic transfer of protocols, allowing them to deliver reproducible results across multiple team sites, instilling confidence in testing outcomes and helping them reach milestones more rapidly.
In summary, AI takes the “art” out of cell culture by helping researchers apply precise protocols to their organoids. These emerging technologies help ensure that experiments will be done the same way regardless of the user and that cells will be tended to on time regardless of the day of the week. For drug developers, who want to find the best candidates and move them forward as quickly as possible, AI promises to be the ideal research partner.
About the author:
Shantanu Dhamija joined Molecular Devices, a Danaher company, in April 2021 as Vice President of Strategy and Innovation to lead strategic planning and inorganic growth initiatives. He previously worked at Danaher’s Leica Biosystems, where he served as Director of Strategy and Market Analytics and was responsible for strategic planning, mergers and acquisitions market work and market intelligence. Prior to Danaher, Shan was a part of Deloitte Consulting's Healthcare and Life Sciences Strategy and M&A practice. He earned his MBA and graduate degrees in Electrical and Biomedical Engineering from the University of Michigan, Ann Arbor, and his bachelor’s degree in electrical engineering from the University of Delhi, India.