a) Viscoelastic properties and activity are measured within the cytoplasm of cells using a custom-built optical tweezers setup. b) Sinusoidal forces (blue) at varying frequencies are applied to a probe particle while the particle displacement (red) in response to this force is monitored. c) Repeating this procedure at different frequencies allows to calculate viscoelastic material properties in terms of the complex shear modulus G*(f) which consists of the storage modulus (blue) and a loss modulus (red). Both storage and loss modulus can collectively be fitted using a generalized fractional Kelvin-Voigt (fGKV) model (inset). d) This fGKV model consists of two power-laws, which in a simplified way can be pictured as a more solid-like material class resembling polymeric filaments and a more fluid-like material class (crowded molecules). e) Using an additional passive measurement, violation of the fluctuation-dissipation theorem can be directly visualized. (red area) f) The effective energy quantifies intracellular activity and can be fitted with a two-parameter power-law. g) Complex and frequency-dependent intracellular active mechanical properties can be reduced to a fingerprint of just 6 parameters, which describe the intracellular mechanical state.

Effect of cytoskeletal perturbations on the intracellular mechanical fingerprint of HeLa cells.

a) Composite images of HeLa WT cells treated with Cytochalasin B (CytoB), Latrunculin A (LatA), Nocodazole (Noc), Cytochalasin B + Nocodazole (CytoB+Noc), and Latrunculin A + Nocodazole (LatA+Noc). DNA stained with Hoechst in blue, microtubules stained with α-tubulin antibody in red, actin stained with phalloidin in green, and probe particles in white. Scale bar: 20 µm b-j) Changes in the fingerprint parameters and corresponding rheological quantities following cytoskeletal perturbation. The dashed lines in plots for G′, G″ and EEff show the models depicted in Figure 1c,f using the mean fit parameters A, B, E, α, β and γ.

a) The mechanical fingerprint of 7 different cell types was determined by measuring their complex shear modulus and effective energy. b) Comparing HeLa cells to C2C12 muscle cells shows that muscle cells are overall stiffer (higher A). c) The comparison between HeLa cells and macrophages shows that macrophages have a higher intracellular activity (increased E) but are also more liquid-like and softer (increased α, β and decreased B) d-l). Using the fingerprint, active mechanical differences between cell types can be compared according to changes in the more solid-like material properties, the more liquid-like material properties, and intracellular activity.

a) Number of significantly different fingerprint parameters in pairwise cell type comparisons. b) Frequency of parameters showing the highest z-score. c) Distribution of z-score for all parameters for pairwise cell comparison. d) Correlation analysis between fingerprint parameters shows that not all parameters are varied independently. e) Explained variance ratio of the different principal components (PC) of the principal component analysis (PCA). f) Relative contribution of the fingerprint parameter to the first three principal components. PC1 mainly consists of α, β and log B, PC2 mainly consists of log E and γ and PC3 to 50% of log A. g) Plotting all cell types according to the first two principal components PC1 and PC2 shows that two parameters are already sufficient to distinguish between most cell types. h,i,j,k) Qualitative phase diagram of the active mechanical space. Activity, mainly captured by parameter E, resistance, dominated by A, and solid-liquid switching can be described as fluidity that is determined by principal component 1 (PC1) are varied among different cell types. Using this three-dimensional space allows identification of physical differences among different cell types.

Results of principal component analysis.

The explained variance ratio quantifies how much infor-mation is captured by the corresponding component. The composition of each component is explained by its axis in parameter space. Here, the values are shown to the third decimal digit.

Statistics on rheology experiments

Parameter E: different cell types

Parameter − γ: different cell types

Parameter A: different cell types

Parameter α: different cell types

Parameter B: different cell types

Parameter β: different cell types

Parameter E: drug conditions vs HeLa

Parameter γ: drug conditions vs HeLa

Parameter A: drug conditions vs HeLa

Parameter α: drug conditions vs HeLa

Parameter B: drug conditions vs HeLa

Parameter β: drug conditions vs HeLa

Mean R2 values for fit quality: bootstrapped data

Mean R2 values for fit quality: raw data

Biological characteristics of the investigated cell lines

Schematic of bootstrapping significance test.

a) Histogram for the estimates of two different parameters. b) Bootstrapping procedure was performed on both parameters to get an estimate for the distribution of the mean. c) Difference of both distribution is calculated. d) Depending on which percentile of the distribution is below 0, the significance score is determined.

Representative images of immunostainings on HeLa WT cells treated with 10 µg/mL Cytochalasin B (CytoB), 1µM Latrunculin A (LatA), 10 µg/mL Nocodazole (Noc) as well as the combinations CytoB+Noc and LatA+Noc.

The composite images show the overlay of DNA labeled using Hoechst (blue), microtubules stained with anti-α-tubulin primary antibody with Alexa Fluor 568 (red), actin stained with Phalloidin 647 (green), and 1 µm carboxylated beads 480 nm (white). The wide emission spectrum of the beads causes them to appear in the microtubule channel. Scale bar: 20 µm.

Distribution of the phase-space parameters resistance (log A), fluidity (PC1), and activity (log E) after grouping the investigated cell types according to a-c) species, d-f) cell type, g-i) cancer status, j-l) metastatic potential, and m-o) migratory potential.