Drug-resistant diseases could cause 10 million deaths per year by 2050, according to a new report from the UN Interagency Coordination Group on Antimicrobial Resistance. As part of a broad set of recommendations for changing the course of this healthcare crisis, the report states, “Additional effort, investments and incentives are needed to spur innovation in antimicrobial medicines.” As such, there is an enormous need to re-imagine traditional research pipelines to deliver new treatments for antimicrobial-resistant diseases. My team at IBM Research is focused on developing new systems that accelerate materials discovery, including materials that can be used as therapeutics, by using a combination of computer and materials science.
Discovery of new therapeutic materials often requires tedious trial-and-error experimentation to determine the optimal candidate for different applications. While effective, this approach is limited by researcher intuition, time, and the number of test candidates that can be feasibly produce and, hence, results in long material development timelines. To facilitate faster identification and commercialization of promising material candidates, there is a huge need for new ways to greatly accelerate materials synthesis with precise control over their features and properties. This is especially true in our efforts to develop new therapeutic materials for treating drug-resistant antimicrobes1 or cancer,2,3 where minor changes in the material’s structural components has a dramatic influence over their resulting efficacy.
To address this challenge, we investigated the use of continuous manufacturing techniques for the rapid, high-throughput synthesis of materials. Our findings were published in the Journal of the American Chemistry Society. Long used by the petroleum and pharmaceutical industries for their large-scale manufacturing processes, continuous manufacturing (or continuous-flow synthesis) has only recently been utilized for laboratory-scale synthesis of high-performance and specialty materials. Given its safety, mass, and heat transfer advantages, continuous-flow synthesis has a lot to offer both in terms of rapidly advancing materials synthesis and potentially enabling programmatic design of structural features.
Figure 1: Strategy for developing automated continuous flow reactors for materials.
In collaboration with the Waymouth lab at Stanford University, we developed a new continuous-flow platform that could harness highly active polymerization catalysts to help deliver precise, well-defined materials on short timescales. In this regard, we were highly successful in designing systems that prepared materials in as quickly as 6 milliseconds or delivered multiple grams in under a minute, both important features for delivering sufficient quantities of materials for testing. However, this process was initially limited to materials consisting of a single type of repeat unit.
In order to more broadly impact materials design and development, we needed to build a system that could accommodate multiple types of repeat units without compromising the integrity of the material. To achieve this, we discovered it was necessary to perform an in-line catalyst switch in order prevent the onset of side reactions that could decompose the material. This step allowed us to prepare highly controlled and well-defined di- and tri-block copolymers within seconds, greatly expanding the potential application space for continuous-flow polymerizations.
Figure 2: Video showing controlled scale-up synthesis of polylactide.
Despite these significant advances, these reactor setups were limited to preparing a single type of material per run. A key advantage in continuous-flow synthesis is the ability to directly vary the reactor parameters via computer control, allowing for multiple materials to be made in a single run. By programming the reactors to adjust the flow rates at predefined intervals, we were able to synthesize libraries of 40 to 100 distinct materials in 4–9 minutes. This is a level of throughput that can be difficult to match via traditional materials development; manually producing an equivalent number of materials with similar levels of precision would require days to weeks. Additionally, using a computer to control the reactor allowed us to programmatically design specific structural features of the material and enable them to be accurately realized in the final product. Again, this represents a level of control that is difficult to achieve by standard synthetic methods.
Figure 3: Computer controlled synthesis of material libraries. A) Synthesis of polylactide homopolymer library in 4 minutes. B) Synthesis of polyvalerolactone-polylactide block copolymer library in 9 minutes.
Our first demonstration of programmable synthesis of materials points to a future where the structural features of materials could be systematically varied to produce highly tailored materials with unique properties. This is critical for our efforts in developing new materials to overcome drug-resistant infectious diseases and cancer. Having the ability to quickly generate targeted libraries of therapeutic materials will allow for a potentially better understanding of their biological activity as a function of structural compositions.
This in turn can facilitate more rapid identification and advancement of new therapeutic materials. The potential to accelerate the discovery of new treatments for drug-resistant infections is just one of many potential applications of automated continuous manufacturing platforms that could greatly accelerate development timeframes for new materials by helping to eliminate the tedious and error-prone lab-scale synthesis efforts, allowing researchers focus on the more critical aspects of design and analysis.
1) Chin, W.; Zhong, G.; Pu, Q.; Yang, C.; Lou, W.; Sessions, P. F. D.; Periaswamy, B.; Lee, A.; Liang, Z. C.; Ding, X.; et al. A Macromolecular Approach to Eradicate Multidrug Resistant Bacterial Infections While Mitigating Drug Resistance Onset. Nature Communications 2018, 9(1), 917. https://doi.org/10.1038/s41467-018-03325-6.
2) Park, N. H.; Cheng, W.; Lai, F.; Yang, C.; Florez de Sessions, P.; Periaswamy, B.; Wenhan Chu, C.; Bianco, S.; Liu, S.; Venkataraman, S.; et al. Addressing Drug Resistance in Cancer with Macromolecular Chemotherapeutic Agents. J. Am. Chem. Soc. 2018, 140(12), 4244–4252. https://doi.org/10.1021/jacs.7b11468.
3) Zhong, G.; Yang, C.; Liu, S.; Zheng, Y.; Lou, W.; Teo, J. Y.; Bao, C.; Cheng, W.; Tan, J. P. K.; Gao, S.; et al. Polymers with Distinctive Anticancer Mechanism That Kills MDR Cancer Cells and Inhibits Tumor Metastasis. Biomaterials 2019, 199, 76–87. https://doi.org/10.1016/j.biomaterials.2019.01.036.
At the 18th European Conference on Computational Biology and the 27th Conference on Intelligent Systems for Molecular Biology, IBM will present significant, novel research that led to the implementation of three machine learning solutions aimed at accelerating and guiding cancer research.
Our team of IBM researchers published research in Radiology around a new AI model that can predict the development of malignant breast cancer in patients within the year, at rates comparable to human radiologists.