Several years ago, I created a new cell culture reagent designed to more faithfully reflect nutrient levels in human blood: Human Plasma-Like Medium (HPLM), thus pioneering the systematic development and use of synthetic physiologic media. Our work to date demonstrates that culture in HPLM versus traditional media has profound impacts on cell metabolism, genetic dependencies, drug sensitivity, and immune cell biology. My group has also developed a bioreactor platform (chemostat) that enables the continuous-flow culture of blood cells at metabolic steady state in tightly controlled conditions. Our systems are integrated with software to monitor and adjust pump rates for the continuous flow of fresh media into – and conditioned media (containing cells) out of – the reactors, and further, to maintain set points for temperature (T), pH, and dissolved oxygen (DO) through a set of sensors and programmable control loops. My group has established proof-of-concept data for maintaining CRISPR screens, competition assays on mixed pools of DNA-barcoded cell lines, and primary T cell cultures under steady state conditions in this system. Overall, by integrating HPLM and our chemostat system, we have established a new platform to examine genetic and environmental contributions to cell fitness in controlled conditions that more faithfully reflect those that may be faced in the human body.
For example, we recently sought to test the hypothesis that CRISPR screens carried out in our platform would reveal fitness genes that are masked in a traditional batch format. To test this idea, we used leukemia cells infected with a genome-wide single guide (sg)RNA library to seed three parallel screens: flasks containing (1) conventional media or (2) HPLM in a standard incubator that exposes cells to atmospheric oxygen (~20%); or (3) a chemostat set to simulate conditions in blood circulation: (I) HPLM; (II) pH, 7.4; and (III) DO, 8%. By comparing dependency scores from the chemostat screen to the average of those carried out in flasks, we identified hundreds of selectively essential genes based on the culture format. Notably, the set of “chemostat-essential” hits is largely comprised of genes that are rarely defined as dependencies across the 1,100 cancer cell lines catalogued in the Broad Institute Dependency Map (DepMap) project. We are now poised to investigate the roles of these previously hidden cell fitness genes.
In addition, we have constructed a panel of over forty human blood cancer cell lines that samples various subtypes – likely making ours among the largest such panels in the academic community. By introducing DNA “barcodes” across this panel, we have shown that deep sequencing can be used to track individual cell lines within mixed pools. Beyond eliminating the need to passage cultured cells, chemostats enable the evaluation of cellular responses to single perturbations while holding all other conditions constant, thus minimizing unintended secondary effects. In turn, we reason that pooling DNA-barcoded cells for competition assays in our platform could enable more precise mapping of nutrient liabilities for cells grown in circulation-like conditions. Indeed, we have initiated efforts to systematically map heterogenous nutrient dependencies across our DNA-barcoded panel, with initial focus on metabolites uniquely defined in HPLM versus traditional media. Ultimately, we plan to use hypothesis-driven (e.g., expression analyses along relevant metabolic pathways) and unbiased approaches (e.g., metabolomics and CRISPR screening) to determine the molecular basis for each respective auxotrophy.
Finally, it has been long appreciated that immune cells must retain functionality in the face of dynamic and possibly hostile conditions as they traffic throughout the body and, in some cases, infiltrate tumors. However, existing knowledge of immune cell biology has also been largely derived from conventional models. Moreover, despite achieving durable responses in some patients, the broad clinical benefit of adoptive cell therapies that have revolutionized the landscape of cancer treatments over the last decade has been somewhat limited. Harnessing the promise of immunotherapy will demand progress in understanding determinants of immune cell fitness and function in conditions relevant to human health and disease. T cells are the main “soldiers” of adaptive immunity, helping to protect the body against infections and cancer, while Natural Killer (NK) cells are central effectors of innate immunity that have emerged as another promising source of cell-based cancer therapies. Leveraging our chemostat platform, my lab is now uniquely poised to examine cell-environment interactions in immunity under controlled conditions with greater relevance to those encountered in the body. Our initial focus will be directed toward decoding the behavior of T cells and NK cells upon precisely introduced perturbations that would be otherwise inaccessible in existing models.