^{*}

Edited by: Sudha Kumari, Massachusetts Institute of Technology, United States

Reviewed by: Michael Loran Dustin, University of Oxford, United Kingdom; Dylan Myers Owen, University of Birmingham, United Kingdom; Noa B. Martin-Cofreces, Princess University Hospital, Spain

This article was submitted to Cell Growth and Division, a section of the journal Frontiers in Cell and Developmental Biology

This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

Mechanical vibrations affect multiple cell properties, including its diffusivity, entropy, internal content organization, and thus—function. Here, we used Differential Interference Contrast (DIC), confocal, and Total Internal Reflection Fluorescence (TIRF) microscopies to study mechanical vibrations in live (Jurkat) T cells. Vibrations were measured via the motion of intracellular particles and plasma membrane. These vibrations depend on adenosine triphosphate (ATP) consumption and on Myosin II activity. We then used spectral analysis of these vibrations to distinguish the effects of thermal agitation, ATP-dependent mechanical work and cytoskeletal visco-elasticity. Parameters of spectral analyses could be related to mean square displacement (MSD) analyses with specific advantages in characterizing intracellular mechanical work. We identified two spectral ranges where mechanical work dominated vibrations of intracellular components: 0–3 Hz for intracellular particles and the plasma-membrane, and 100–150 Hz for the plasma-membrane. The 0–3 Hz vibrations of the cell membrane that we measured in an experimental model of immune synapse (IS) are expected to affect the IS formation and function in effector cells. It may also facilitate immunological escape of extensively vibrating malignant cells.

Mechanical work inside living cells plays a significant role in cell physiology (

Notably, mechanical work generates forces that are non-thermal and depend on ATP consumption. As a result of these forces, the cytoskeleton transfers mechanical work and augments the diffusion of intracellular particles (

The plasma membrane (PM) is physically connected to the cortical actin (

Here, we aimed to study intracellular diffusion and intracellular mechanical work, as they occur in T cells. Specifically, the impact of non-equilibrium forces on intracellular particles motion enables the investigation of those forces by analyzing the dynamics of these particles. Intracellular diffusion motion is usually characterized as anomalous diffusion, for which the mean square displacement (MSD) is not linearly correlated to the time-lag of measurements (^{2}⟩=_{α}^{α}, where _{α} is the diffusion coefficient and α is the diffusion power. Finding these specific anomalous diffusion parameters does not usually facilitate the identification of the main cellular mechanisms that explains those results. The reason is that different underlying mechanisms may lead to similar anomalous diffusion _{α} and α results (e.g.,

Three main mathematical models have been defined in relation to different cellular mechanisms that may govern the intracellular anomalous diffusion, including visco-elasticity, diffusion and percolation in a crowded environment and medium with traps or energetic disorder (

The diffusion motion patterns of a particle that is stuck in a trap and will randomly gain enough energy to jump to another location may be similar to the diffusion motion patterns of a free particle that randomly gains a large amount of mechanical energy that will cause it to jump to a relatively remote location. Both situations are described by the CTRW model. Accordingly, the impact of mechanical energy on intracellular diffusion is likely to cause the anomalous diffusion patterns to better match that model. A distinctive difference between the CTRW model and the other FBM or RWF models is that the CTRW model describes a non-ergodic process, while FBM and RWF describe ergodic processes (

Following that, an analytic framework of intracellular diffusion motion that combines the effects of spatial fluctuations with ergodicity breaking should clearly capture the impact of intracellular mechanical work on the anomalous diffusion. Such an analysis should be able to discern underlying mechanisms of intracellular diffusion motion that may yield similar anomalous diffusion parameters (_{α} and α) but differ in their ergodicity.

The power spectral density (PSD) analysis of a wide range of time-dependent parameters was studied in many fields of science, including physics, biophysics, geology, weather science, etc. (_{s}(f,∞)=A/f^{β} (

Spectral analysis of the dynamics live cells constituents may have the advantage of providing better insight into the biophysical mechanisms behind their anomalous diffusion and ergodicity breaking, especially in regard to intracellular mechanical work.

Here, we utilized a relatively simple spectral analysis framework for the exploration of intracellular diffusion and intracellular mechanical work. This framework is based on Discrete Fourier Transform (DFT) of temporal position changes of intracellular constituents. This framework then serves to analyze the intracellular diffusion of intracellular particles (e.g., vesicles or other small organelles) and fluctuations of cell diameter in live Jurkat cells, before and after ATP depletion. From the PSD results of the motion of cell constituents, we define parameters that reflect intracellular mechanical work. We show that cells under normal (physiological) conditions are active and produce significant extent of mechanical work. This work is diminished in the same cells that become non-active after ATP depletion. Next, we explore intracellular mechanical work over a wide spectrum of time-scales and frequencies. We identified two spectral ranges where mechanical work dominated vibrations of intracellular components: 0–3 Hz for intracellular particles and the plasma-membrane, and 100–150 Hz for the plasma-membrane. Such vibrations of the cell membrane are expected to affect the formation and function of the immune synapse by effector cells. Thus, we studied the membrane vibrations of Jurkat cells in an experimental model of the immune synapse using total internal reflection fluorescence (TIRF) microscopy. Indeed, we identified ATP-dependent membrane fluctuations at the model synapse, esp. below 3 Hz. These mechanical fluctuations of the cell membrane may also affect T cell recognition of extensively vibrating malignant cells. We expect that spectral analysis of intracellular vibrations and motion will become a useful tool for characterizing cell condition and activity in health and disease.

The cytoskeleton is an elastic polymeric mesh that spans the intracellular volume with a mesh size of around 50 nm (_{mechanical} = 0.5^{2} + _{loss}, where _{loss} is the (relatively small) dissipated energy, ^{2}. The integral of the spectrum of vibrations represents the approximated total mechanical energy of the measured part of the cytoskeleton mesh in the specified spectral range. Monitoring movements of multiple intracellular particles and averaging the spectral analysis results of these movements enable insight into the mechanical energy and work of the entire cytoskeleton and cellular system.

Thermal agitation forces and incoherent intracellular mechanical forces (which are a by-product of directed forces that are utilized for cell physiology), both act on the cytoskeleton. Together, they contribute to the cytoskeletal modes of vibrations. These vibration modes can then be revealed by monitoring embedded particles inside the mesh for their diffusion motion. Spectral analysis of the diffusion motion of these particles can be related to the modes of vibration and mechanical energy of the adjacent cytoskeleton.

In order to explore these relations we consider the diffusion motion of an intracellular particle embedded in the elastic cytoskeleton mesh as illustrated in

Analysis of intracellular particle motion embedded insidethe cytoskeletal mesh.

To further investigate these spectral amplitudes of spatial fluctuations, we consider the three following aspects:

The summation (integral) of all powers (the squared amplitudes), which relates to the total mechanical energy of the particle-elastic mesh system in the specified spectral range. Each power represents a specific amplitude of vibration (_{mechanical} = 0.5^{2}).

The vibration spectrum is fitted with a power equation for each set of spectral results for each cell and condition:

The relation of

The extent to which the actual spectrum of amplitudes fits to the model equation of estimated power. In other words, we quantify the magnitude of the sum of squared errors (SSE) that relates the spectrum to its power fit equation. Then, the correlations of SSE values to the intracellular diffusion and mechanical work characteristic are investigated.

A general diffusion process can be characterized by its typical probability density function (PDF) of translocations for each corresponding time-lag, and following that, by the statistics of MSDs. First we analyze the condition of Brownian diffusion, for which α=1. The PDF of translocation for each time-lag in this case is Gaussian. We simulated the movement of particles undergoing normal diffusion with different diffusion coefficients (represented by the different standard deviations of the PDF of translocation Gaussians; _{estimated} = ^{b}] is correlated to the standard deviation (SD) of PDF of translocations (

Simulations results. ^{b}

The characteristic PDF of a diffusing particle (along with its parameters _{α} and α) can be related to the spectral analysis results of its position changes in time. For an ergodic diffusion process, the PDF of translocations in a time-lag that corresponds to the interval between measurements reveals the statistic of the sequential translocation steps of that diffusive object. Knowing this statistic of translocations does not allow us to reconstruct the exact trajectory of that particle during measurements. Still, one can anticipate the different amplitudes of spatial fluctuations and their correspondent statistical frequencies from the given PDF of translocations. The knowledge of the different amplitudes of spatial fluctuation with their corresponding frequencies of occurrence is generally equivalent to the results of the DFT analysis of the position fluctuations of that same particle over time. In this analysis, the different frequency-dependent amplitudes of the DFT represent the different amplitudes of the particle’s spatial fluctuations (i.e., translocations). Similarly, the corresponding frequencies of these amplitudes (in the DFT) represent the equivalent probability of occurrence of these translocations (in the PDF). Thus, it is expected that the DFT results of this particle diffusion movements will be related to its PDF of translocations and accordingly, to its diffusion parameters.

Following that, the possible analytical relations between the DFT power fit parameters _{α} and α in conditions that included anomalous or non-ergodic diffusion were explored and described in _{α} (this compatible with the simulation experiment results presented in

In living cells: the applied forces that relate to intracellular mechanical work may be directional in contrary to the perfect random and symmetric thermal forces or the elastic forces of the relatively symmetric cytoskeleton mesh. Those forces due to intracellular mechanical work will break the ergodicity and the symmetry of the affected diffusion motion in cells. Following that, the high

To summarize the expected influence on the

The _{α} and diffusivity associated with active cells.

Values of

According to these arguments, the product value of

In an ergodic diffusion process, SSE values of the power fit formula that fit the spectral amplitudes of position temporal fluctuations are correlated to diffusivity, as demonstrated in

In this section, we present the results of the spectral analysis of position fluctuations in time of intracellular particles in Jurkat T lymphocytes, each cell before and after ATP depletion. Specifically, we explore the relations of these results to diffusion parameters (_{α} and ^{b}

From the position results of the detected intracellular particles, we also calculated the MSD values for time-lags form 0.3 to 3 s (with a 0.3 s gradual increase). The average MSD values were determined for that series of time-lags for each cell before and after ATP depletion. These values were fitted to a model of power series to determine the corresponding _{α} and

As expected, the _{α} and _{α}: 0.023 μm^{2}/s vs. 0.013 μm^{2}/s with ^{–4}. ^{–3}).

To estimate ergodicity in the cells, we employed a basic principle which implies that in an ergodic system the distribution of translocations of a specific trajectory is not dependent on its spatial location. Following that, the distributions of translocations of all trajectories in a perfectly ergodic system are similar. In our experiment, the distribution of translocations of each trajectory or a particle is reflected by this particle’s MSD results. Evaluating the heterogeneity of all particles MSD results in a cell (SD of MSD results in that cell) will produce an estimation of the ergodic level in that cell system (^{2} in normal cells vs. SD of MSD’s = 0.018 μm^{2} in ATP-depleted cells,

We summarize in

Differential Interference Contrast (DIC) microscopy results of intracellular particles motion in live cells before and after adenosine triphosphate (ATP) depletion analyzed by amplitude spectral density (ASD) and power spectral density (PSD).

Next, we evaluated the ability of these parameters, derived from spectral analysis of position fluctuations, to capture different aspects of particle motion, esp. in comparison to the prevalent anomalous diffusion parameters _{α} and

Correlation of amplitude spectral density (ASD)/ power spectral density (PSD) parameters and mean square displacement (MSD) parameters in live cells before and after adenosine triphosphate (ATP) depletion.

Strikingly, the two groups of parameters were significantly correlated in ATP-depleted cells (_{α}, the

On the other hand, for cells before ATP depletion, there was no correlation between

Moreover, in active cells the correlation between

The

Next, we further wanted to test if these new parameters, which relate to the amplitudes or powers of temporal position fluctuations, have a better discriminative ability to differentiate between active working cells and non-active ATP-depleted cells in comparison to the classical diffusion parameters of _{α} and α

Comparing the discrimination power of power spectral density (PSD)/amplitude spectral density (ASD) vs. mean square displacement (MSD) analyses for intracellular mechanical work. T-statistics values for the difference between cells before ATP depletion and the same cells after ATP depletion according to the two groups of parameters: ASD/PSD analysis based parameters in compare to MSD analysis based parameters. Corresponding parameters from the two groups are compared: the

We conclude that the new DFT-derived parameters may detect better the increase in mechanical energy that characterizes active and physiologically normal cells (in contrast to non-active ATP-depleted ones), in comparison to the classical diffusion parameters of _{α} and α

In active cells the mechanical work of the elastic cytoskeleton augments the motion of intracellular particles (

Live cells Measurements of cell diameter fluctuations.

In order to test this hypothesis we first highlighted the cell membrane via fluorescent staining of CD45, an abundant surface glycoprotein in T cells (

Amplitude spectral density (ASD) and power spectral density (PSD) analysis results of live cells diameter fluctuations before and after adenosine triphosphate (ATP) depletion.

Our fast confocal imaging of the cell diameter and its fluctuations enabled us to examine also the modes of vibration of this diameter at relatively high frequencies. Specifically, we studied the amplitudes of fluctuations of cell diameter at 50–200 Hz in cells before and after ATP depletion and also in fixed cells (

Amplitude spectral density (ASD) results in high frequencies (>50 Hz).

In an ideally viscous medium in equilibrium, the spectrum of thermal forces on a particle is equivalent to white noise and independent of frequency. Thus, the power spectrum of spatial fluctuations of such a particle should be completely random and non-correlated. If correlations in the particle motion appear due to extra thermal forces, as elastic forces in the medium or forces due to mechanical work, then the amplitude spectrum of that particle motion is expected to be less random. Autocorrelation analysis of the amplitude spectra will lead to higher values and a decrease in decay with increased frequency lags. Therefore, autocorrelation analysis may differentiate an amplitude-spectrum that is more typical to ideal Brownian process or to noise from an amplitude-spectrum that is more typical to elastic forces or mechanical work.

Following this concept, three ranges of frequencies of the amplitude-spectra of the cells were analyzed for autocorrelation: 50–100 Hz, 100–150 Hz, and 150–200 Hz. In each range of frequencies, the average autocorrelation results for each lag, for each cell and for each cellular condition are presented in

Average autocorrelation results of Amplitude spectral density (ASD) analysis.

We assume that in ATP-depleted cells no significant mechanical work is produced. Following that assumption, it seems that the difference in the decay between normal and ATP-depleted cells in 100–150 Hz may be due to a larger extent of mechanical work in normal active cells that reduces the randomness of their amplitudes of vibrations. If elasticity was the main contributing factor in this frequency range, the decay of autocorrelation in fixed cells and ATP-depleted cells are not expected to be similar. This is since the mechanical characteristics of the intracellular medium are very different under these two conditions. Last, at the frequency ranges of 50–100 Hz and 150–200 Hz, autocorrelation decay functions under all conditions are pronounced and similar. This indicates that the measured powers in this frequency range may represent thermal agitation or noise of the measurement system.

Differences in the shape of the amplitude-spectra (

So far, we have described intracellular fluctuations as captured by the motion of intracellular particles or by the cell boundaries. The fluctuations of the cell membrane may have an impact on the formation and function of the immune synapse that forms between T cells and antigen presenting cells (APCs) (

Fluctuation of Jurkat cell membrane in a model of the immune synapse imaged by TIRF microscopy.

The amplitude spectral density (ASD) results for frequencies >3 Hz (and up to 100 Hz) were similar for normal cells and cells after fixation (

At frequencies <3 Hz, membrane fluctuations were significantly higher in normal active cells than in fixed cells and in ATP-depleted cells (

Here we studied active modes of vibrations of the cytoskeleton and plasma membrane in lymphocytes. Specifically, we analyzed spatial fluctuations of intracellular particles and of the cell diameter utilizing DFT analysis. The cytoskeletal motion was studied by monitoring the motion of intracellular large particles, larger than the cytoskeleton mesh size of around 50 nm (

The extent of mechanical vibrations in lymphocytes, especially the mechanical vibrations of the plasma membrane, could significantly impact the immune synapse (

There are two main approaches to characterize the complex intracellular medium. If viewed as a highly complex solution, then the motion of intracellular particles could be naturally analyzed in terms of time-dependent translocations and diffusivity. Still the intracellular content can also be viewed as a two component elastic gel; i.e., an elastic polymeric mesh made of the actin cytoskeleton, immersed in crowded viscous gel. In this case, the medium would be more intuitively analyzed in terms of the spectrum of its vibrations.

The random motion of particles has been theoretically and experimentally investigated through their modes of vibration using PSD analysis. The PSD analysis is classically calculated by first performing a Fourier transform of each individual trajectory x(t) [or y(t)] over the finite observation time T and then averaging the spectral results for a statistical ensemble of all possible trajectories (

In the case of Brownian motion the relation between powers of fluctuations and the related frequencies could be described by a power-law equation in the form:

where μ_{s} stands for the power, _{α} stands for the diffusion coefficient and

In the case of FBM sub-diffusion, when α < 1, the power of the frequency (

Our experimental results in living cells follow these theoretical equations. Indeed, we find that the PSD (or ASD) results can be accurately fitted with a power-law equation (Eq. 1). The parameter _{α} results (

Aside from the fit parameters of

We found that the ASD- and PSD-related parameters could better distinguish mechanically active cells from non-active ATP-depleted cells, as compared to the regular MSD-based anomalous diffusion parameters (

Incoherent forces due to intracellular mechanical work may be applied at different locations on the cytoskeleton, each with its own frequency. Such forces are expected to make the ASD results of cytoskeletal fluctuations more complex, lowering the quality of a fit to a relatively simple power-law model. In this situation the system could be characterized as having relatively high energetic disorder, which directly relates to lower ergodicity. In contrast, ASD results of the same network experiencing only thermal forces (i.e., “white noise” forces that don’t have any frequency preference) will more accurately follow a fit of a power-law model. Accordingly,

The calculation procedure of the ASD or PSD fluctuation parameters seems to be simpler and more automatic in comparison to the calculation of MSD parameters. MSD analyses require taking statistical measurements of multiple displacement results that relate to different time-lags for each particle.

The cytoskeletal vibrations could be monitored by motion of particles that are embedded within it but also by the motion of it borders- namely, the cortical actin and the adjacent cell membrane. Analyzing fluctuations in cell diameter by ASD and PSD calculations that were conducted in the same cells before and after ATP-depletion revealed compatible results with the ASD and PSD analysis of intracellular particles motion. The power fit parameters of cell diameter fluctuations

Our confocal microscope line-scan imaging of fluctuations of the cell diameter enabled to conduct ASD analyses over a wide range of frequencies. Thus, we could define several specific ranges of frequencies, each with a dominant underlying mechanism (

We note that the T cell surface is covered with mobile microvilli. Microvilli mobility has been shown to depend on actin remodeling, and occurs over time scales of seconds to tens-of-seconds (

TCR activation has been shown to be a dynamic process, in which the TCR-pMHC bond is repeatedly ruptured and reconnected by perpendicular forces to the immune synapse plane. Rupture forces acting on the TCR-CD3-pMHC bond promote conformational changes of the TCR chains that promote TCR activation (

Here, we studied intracellular and membrane vibrations in Jurkat cells that originated from human leukemic T cells. Malignant cells typically have increased active mechanical fluctuations at relatively low frequencies (below 3 Hz) (

Possible effects of cell membrane fluctuations on T cell function.

We conclude that spectral analysis (either ASD or PSD) may provide a simple and effective technique to study active cellular vibrations and the overall mechanical activity of cells. Active vibrations of the cell membrane may influence lymphocyte ability to respond to immunological cues and may further enable malignant cells to escape immunological surveillance.

Complete Medium (medium): RPMI-1640, DMEM medium, heat-inactivated fetal calf serum (FCS), penicillin, streptomycin, glutamine, sodium pyruvate, and HEPES obtained from Biological Industries (Kibbutz Beit Haemek, Israel). Rotenone and 2-deoxy-d-glucose from Sigma-Aldrich (St. Louis, MO, United States). CD45 proteins were purchased from BioLegend. Anti-human CD3 from eBioscience Inc. (Thermo Fisher Scientific). Blebbistatin was purchased from Sigma-Aldrich (St. Louis, MO, United States).

Jurkat (human leukemic) E6.1 (CD4^{+}) T cells were a kind gift from the Samelson lab at the NIH. Jurkat cells were maintained in RPMI-1640 medium supplemented with 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, 2% glutamine, 2% sodium pyruvate and 2% HEPES. Cells were maintained in completely humidified air with 5% CO2 at 37°C.

CD45 proteins were labeled using mouse anti-human primary antibodies conjugated to Alexa647 fluorophore (BioLegend, 304056). Labeling procedure followed the manufacturers’ protocols. Briefly, 0.5 μg of mouse anti human anti-CD45 monoclonal antibody conjugated to Alexa647 was added to 500 × 10^{3} cells suspended in FACS buffer for 45 min on ice. Cells were then washed in phosphate buffered saline (PBS) for three times and suspended in imaging buffer (RPMI without phenol red, 10% FBS, 25 mM HEPES).

Coverslip preparation was as follows: coverslips (#1.5 glass chambers, iBidi) were washed with acidic ethanol at room temperature (RT) for 10 min and dried at 37°C for 1 h. Coverslips were than incubated at RT for 15 min with 0.01% poly-L-lysine (Sigma) diluted in water. This was followed by washing and drying of the coverslips at 37°C for 1 h. For the immune synapse model experiment the poly-L-lysine covered coverslips were incubated for 2 h at 37°C with 10 μg/ml anti CD3 antibodies diluted in PBS. Than the chambers were washed three times with PBS and left with PBS till the application of cells. Finally, cells were suspended in imaging buffer at a concentration of 1 million and 100,000–500,000 cells and were applied onto coverslips.

Paraformaldehyde (PFA) 4% was added to the cells medium while on the coverslips in a ratio of 3/2 for 45 min incubation afterword all liquid were gently aspirated and replaced with imaging buffer (RPMI without phenol red, 10% FBS, 25 mM HEPES).

Upon completion of measurements in all the cells and after recording the location of each cell, blebbistatin 10 μM or Rotenone 0.2 μM and 10 mM 2-deoxy-d-glucose were added to the cells medium. The samples were than incubated for 30 min on the microscope stage. At the end of incubation, each cell was measured again according to its recorded location. We excluded from analysis moving cells that may have changed their location during measurements.

Differential interference contrast (DIC) image stacks were taken with FV-1200 confocal microscope (Olympus, Japan) equipped with an environmental incubator (temperature and CO2) using a 60×/1.42 oil objective.

Confocal microscopy: Jurkat cells were imaged using an Abberior Expertline confocal/STED microscope (Abberior Instruments, Göttingen, Germany), mounted on a TiE Nikon microscope and operated by the Imspector software (v0.13.11885; Abberior Instruments, Göttingen, Germany). The cells were excited using a 638 mn pulsed laser (90 ps) 2 mW/cm2 at 50% power for x-t live cell experiments. Samples were imaged with a (CFI-SR-HP) Apochromat TIRF X100 NA 1.49 oil immersion objective (Nikon Instruments). Image stacks were generated by taking 1,000 serial images with acquisition time of 2 ms for frames of unidirectional 300 pixels (50 nm pixel size, 5 μs pixel dwell time). The reflection light was detected using an APD with a band-pass filter of 650–720 nm and a pinhole setting of 1.1 Airy units. Each line was scanned once.

TIRF microscopy: Jurkat cells were imaged using a TiE Nikon microscope. The cells were excited using a 647 mn pulsed laser (90 ps) at 2 mW/cm^{2} (20% power). Samples were imaged using a (CFI-SR-HP) Apochromat TIRF X100, NA of 1.49, oil-immersion objective (Nikon Instruments). Image stacks were generated by taking 1,000 serial images with an acquisition time of 4.8 ms per individual frames of 128 × 128 pixels (160 nm pixel size). The reflection light was detected using an avalanche photodiode (APD) with a band-pass filter of 650–720 nm.

TIRF images analysis: In each cell, a squared ROI of 121 pixels was chosen at the cell interface with the coverslip. Fluorescence intensity of each pixel in each image was normalized by dividing its intensity with the average intensity of that time-dependent image. The temporal fluctuations of the normalized fluorescence intensities were analyzed by DFT for each pixel in a ROI. The amplitudes of the DFT analyzes were then averaged for each frequency for all the pixels of an ROI to obtain the averaged DFT results of each ROI (or cell) in each condition.

Jurkat cells were measured using a microscope in DIC mode, utilizing × 60 magnification and conditions that were described in detail in the previous sections. The measurements included repeated measurements every 0.3 s over a time window of 30 s. This measurement time allowed us to effectively avoid the constrains of the limited cell size (up to ∼10 μm) on diffusion. The cell image stacks were first converted to 8-bit images and thresholded (yielding binary images) to segment individual entities for tracking.

We defined thresholding levels according to the histogram of gray levels of the images. We noticed that a small range of thresholding values (in gray levels) were appropriate for segmentation, since too narrow threshold values caused fragmentation of the objects into isolated pixels, whereas threshold values that were too wide resulted in object contour thickening and unification. Analyzing the size distribution of the segmented objects, revealed that most of these objects were in the size range of intracellular vesicles or organelles (0.15 to ∼1.17 μm, average diameter 0.5 μm). Particles diameter were similar in cells before and after ATP depletion.

Further analyses of MSD statistics and fitting (“one term power series model fit”) were carried out using Matlab R2017b (MathWorks). Calculations of MSD values of intracellular objects were performed using the ImageJ plugin MultiTracker (The Kuhn lab; The University of Texas at Austin).

The acquired data was exported to Excel spreadsheets (Microsoft Office Professional plus 2010, Microsoft Inc., Redmond, Washington, United States) for graph and table presentation and for statistical analysis with Real Statistic Resource pack. Significance of differences between groups was calculated using Analysis of Variance (ANOVA) single factor function or

The experimental results are shown from same-day experiments. The results were verified to be similar to those of 1–2 additional independent experiments (which are not shown).

The original contributions presented in the study are included in the article/

ES supervised the research. ES and IW designed the research and wrote the manuscript. IW performed the research. Both authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

We thank Naomi Book (The Silberman Institute at HUJI) for her assistance with DIC microscopy.

The Supplementary Material for this article can be found online at:

Differential interference contrast

Total internal reflection fluorescence

Adenosine triphosphate

Mean square displacement

Immune synapse

Plasma membrane

T cell receptor

Major histocompatibility complex

Antigen presenting cell

Fractional Brownian motion

Random walk on fractal

Continues time random walk

Power spectral density

Brownian diffusion

Discrete Fourier transform

Sum of squared errors

Probability density function

standard deviation

Region of interest

Amplitude spectral density

Paraformaldehyde

Analysis of Variance.