Abstract
Cell mechanical phenotype, or mechanotype, is emerging as a label-free biomarker for cancer and pluripotent stem cells. Here, we demonstrate a method for rapid, single cell, mechanotype measurements, termed quantitative deformability cytometry (qDC). We track changes cell shape as they deform into microfluidic constrictions, and observe that the time-dependent strain follows power law rheology; this enables us to obtain single cell measurements of apparent elastic modulus, Ea, and power law exponent, beta, in addition to standard DC measurements of cell size, maximum strain, and transit time. To validate our method, we measure Ea for HL-60 cells as 1.4 ± 0.6 kPa, and confirm qDC is sensitive to pharmacological perturbations of the cytoskeleton. We then characterize the mechanotype of MCF7 and MDA-MB-231 breast cancer cells (Ea = 8.1 ± 0.3 and 4.3 ± 0.7 kPa) and pancreatic adenocarcinoma cells (Ea = 3.1 ± 0.2 to 7.3 ± 0.5 kPa for HPDE, MiaPaCa-2, PANC1, AsPC-1, and Hs766T). Additionally, we observe Ea increases with cell size. By contrast, Ea does not scale with size for silicone oil droplets. Our results suggest that cells undergo strain stiffening when subjected to large deformations. Ea may also increase with larger deformation depths, as the nucleus is typically stiffer than the cytoplasmic region. To determine the value of qDC compared to standard DC outputs, we identify unique and redundant parameters using Pearson's correlation coefficients. We also employ a nearest neighbor machine learning algorithm to test the predictive power of the qDC parameters. Using our algorithm, we identify a minimal set of parameters containing cell size, transit time, and Ea, which confers highly accurate classification of our cancer cell lines. Taken together, the standardized measurements provided by qDC should advance applications of cell mechanotype in basic research and clinical settings.