Hydrodynamic stretching of single cells for large population mechanical phenotyping

Abstract

Cell state is often assayed through measurement of biochemical and biophysical markers. Although biochemical markers have been widely used, intrinsic biophysical markers, such as the ability to mechanically deform under a load, are advantageous in that they do not require costly labeling or sample preparation. However, current techniques that assay cell mechanical properties have had limited adoption in clinical and cell biology research applications. Here, we demonstrate an automated microfluidic technology capable of probing single-cell deformability at approximately 2,000 cells/s. The method uses inertial focusing to uniformly deliver cells to a stretching extensional flow where cells are deformed at high strain rates, imaged with a high-speed camera, and computationally analyzed to extract quantitative parameters. This approach allows us to analyze cells at throughputs orders of magnitude faster than previously reported biophysical flow cytometers and single-cell mechanics tools, while creating easily observable larger strains and limiting user time commitment and bias through automation. Using this approach we rapidly assay the deformability of native populations of leukocytes and malignant cells in pleural effusions and accurately predict disease state in patients with cancer and immune activation with a sensitivity of 91% and a specificity of 86%. As a tool for biological research, we show the deformability we measure is an early biomarker for pluripotent stem cell differentiation and is likely linked to nuclear structural changes. Microfluidic deformability cytometry brings the statistical accuracy of traditional flow cytometric techniques to label-free biophysical biomarkers, enabling applications in clinical diagnostics, stem cell characterization, and single-cell biophysics.

Conflict of interest statement

Conflict of interest statement: The authors have submitted patent applications related to the presented work.

Figures

Fig. 1.

Principles of deformability cytometry. (A) A photograph of the microscope-mounted and fluid-coupled microfluidic deformability cytometry device. Only a single inlet is required. (Scale bar: 25 mm.) (B) A schematic of the microfluidic device (channel height = 28 μm) that focuses cells to the channel centerline before delivering them to the stretching extensional flow is shown. Cells can enter the extensional flow from both directions. (C) A schematic of the deformation of a cell delivered to the center of an extensional flow by being previously aligned at an inertial focusing position, Xeq is shown. (D) High-speed microscopic images showing a focused cell entering the extensional flow region. Delivery and stretching occurs in less than 30 μs. (Scale bar: 40 μm.) (E) Definitions of the shape parameters extracted from images are shown. (F) Density scatter plot of 9,740 size and deformability measurements of single human embryonic stem cells.

Fig. 2.

Mechanical measurements help distinguish populations of cells within blood and pleural fluids. (A) Density scatter plots of the size and deformability of untreated PBMCs and PBMCs stimulated with 12F6 or PHA. (B) Density scatter plots of the size and deformability of untreated granulocytes and granulocytes stimulated with N-formyl-methionine-leucine-phenylalanine. (C) Locations of cell populations found in pleural fluids on a size-deformability map: (i) nonactivated leukocytes, (ii) nonactivated leukocytes, (iii) activated mononuclear cells, (iv) mesothelial cells, (v) suspicious cells. Density scatter plot of the size and deformability (Left) and typical cell blocks and smears (Right) of cells within pleural fluid of a patient diagnosed: (D) negative for carcinoma, (E) negative for carcinoma but with chronic inflammation, (F) negative for carcinoma but with acute inflammation, and (G) positive for carcinoma.

Fig. 3.

Increased deformability is correlated with increased pluripotency. (A) Density scatter plots of the size and deformability of undifferentiated mESCs, and mESCs differentiated by EB and adherent methods. (B) Cell extracts from undifferentiated day 0 and differentiated day 6 (EB) and day 7 (adherent) mESCs were analyzed for OCT4 protein expression. Decreased OCT4 protein in differentiated mESCs was confirmed by Western blot analysis with antiactin antibodies as a loading control. (C) Median and semi-interquartile deviation (SID) statistics of mESC measurements in part a; * Wilcoxon ranked sum, P < 0.001. (D) Density scatter plots of the size and deformability of undifferentiated hESCs, and hESCs differentiated adherently for 9 d and 14 d. (E) Differentiation of hESCs is accompanied by reduced expression of SSEA4 and TRA-1-60, although significant overlap is observed between populations. (F) Decreased OCT4 protein in differentiated hESCs was confirmed by Western blot analysis with anti-actin antibodies as a loading control. (G) Differentiation of hESCs was also confirmed by an absence of staining by AP. (H) Median and SID statistics of hESC measurements in part a; * Wilcoxon ranked sum, P < 0.001.

Fig. 4.

Effects of individual cytoskeletal components on whole-cell deformability. (A) Density scatter plots of the initial diameter and deformability of untreated NIH 3T3 cells (control), 3T3 cells treated with latrunculin A, nocodazole, sphingosylphosphorylcholine, and blebbistatin for 2 h. (B) Median and semi-interquartile deviation (SID) statistics of untreated 3T3 cells and 3T3 cells treated with different concentrations of one of four compounds for 2 h (error bars are SID). “Drug-induced substrate release” indicates that cells detached from the substrate into suspension before the 2-h treatment period elapsed.