Drug sensitivity of single cancer cells is predicted by changes in mass accumulation rate

A RT I C L E S

Drug sensitivity of single cancer cells is predicted by changes in mass accumulation rate

Mark M Stevens1,2,14, Cecile L Maire3,14, Nigel Chou1,4,14, Mark A Murakami3,14, David S Knoff3, Yuki Kikuchi1,5, Robert J Kimmerling1,4, Huiyun Liu3, Samer Haidar3, Nicholas L Calistri1, Nathan Cermak6, Selim Olcum1, Nicolas A Cordero3, Ahmed Idbaih7,8, Patrick Y Wen9, David M Weinstock3,10, Keith L Ligon3,1012 &

Scott R Manalis1,4,13

Assays that can determine the response of tumor cells to cancer therapeutics could greatly aid the selection of drug regimens for individual patients. However, the utility of current functional assays is limited, and predictive genetic biomarkers are available for only a small fraction of cancer therapies. We found that the single-cell mass accumulation rate (MAR), profiled over many hours with a suspended microchannel resonator, accurately defined the drug sensitivity or resistance of glioblastoma and B-cell acute lymphocytic leukemia cells. MAR revealed heterogeneity in drug sensitivity not only between different tumors, but also within individual tumors and tumor-derived cell lines. MAR measurement predicted drug response using samples as small as 25 ml of peripheral blood while maintaining cell viability and compatibility with downstream characterization. MAR measurement is a promising approach for directly assaying single-cell therapeutic responses and for identifying cellular subpopulations with phenotypic resistance in heterogeneous tumors.

2016 Nature America, Inc., part of Springer Nature. All rights reserved.

The choice of drug regimens for individual cancer patients has historically been based on treatment responses observed in large studies across heterogeneous populations. The shortcomings of this approach have motivated a broad effort to personalize treatment decisions for each patient based on the presence or absence of genetic, epigenetic or other biomarkers in an individual tumor1,2.

Although population-based studies have been successful in some instances (e.g., in lung cancers with mutations of EGFR or rearrangements involving ALK), the vast majority of cancer therapeutics have no known markers for susceptibility or resistance3. Even when marker-based predictions can be made, they do not guarantee patient response, as many are the result of correlations from population-based studies4,5.

The second major shortcoming of nearly all available biomarkers is that they are derived from analyses of bulk tumor populations, and therefore do not predict the emergence of resistant subpopulations. More sensitive approaches of genetic...