[Paper Review] Estimate of entropy generation rate can spatiotemporally resolve the active nature of cell flickering

Sreekanth K Manikandan, Tanmoy Kumar Ghosh+4 more|arXiv (Cornell University)|May 25, 2022

Neural dynamics and brain function62 references4 citations

TL;DR

This study introduces a model-independent method to estimate the spatiotemporal entropy generation rate from interference reflection microscopy flickering data of HeLa cells using a short-time inference scheme based on stochastic thermodynamics. The approach reveals active membrane fluctuations at ~1 μm resolution, distinguishing ATP-active membranes from ATP-depleted ones and mapping heterogeneous activity across the cell membrane.

ABSTRACT

We use the short-time inference scheme (Manikandan, Gupta and Krishnamurthy, Phys. Rev. Lett. 124, 120603, 2020), obtained within the framework of stochastic thermodynamics, to infer a lower-bound to entropy generation rate from flickering data generated by Interference Reflection Microscopy of HeLA cells. We can clearly distinguish active cell membranes from their ATP depleted selves and even spatio-temporally resolve activity down to the scale of about one $μ$m. Our estimate of activity is model--independent.

Motivation & Objective

• To develop a model-independent method for estimating entropy generation rate from experimental flickering data of living cells.
• To resolve the active, non-equilibrium nature of cell membrane fluctuations at high spatial and temporal resolution.
• To distinguish active membranes from ATP-depleted membranes using entropy generation as a quantitative, physical measure of activity.
• To map spatial heterogeneity of activity across the cell membrane, revealing non-uniform energy consumption.
• To provide a generalizable framework for quantifying cellular activity without relying on equilibrium assumptions or detailed mechanistic models.

Proposed method

• Application of the short-time inference scheme (based on the thermodynamic uncertainty relation) to infer a lower bound on entropy generation rate from flickering trajectories.
• Use of principal component analysis (PCA) to reduce dimensionality of pixel-level flickering data to a few dominant modes.
• Numerical optimization of the entropy production rate estimate from the reduced stochastic trajectory of the first two principal components.
• Validation of results using a simple active membrane model and comparison with equilibrium spectra under ATP depletion.
• Spatial mapping of entropy generation rate across the membrane by analyzing patches of ~1.44 μm².
• Use of the inequality σ ≥ 2kB⟨J⟩² / (t Var(J)) to bound entropy production, with saturation in the short-time limit.

Experimental results

Research questions

• RQ1 Can entropy generation rate be reliably estimated from short-time flickering data of living cells without assuming a specific model?
• RQ2 To what extent can the spatiotemporal resolution of active processes in cell membranes be improved using this method?
• RQ3 How does the entropy generation rate differ between ATP-active and ATP-depleted cell membranes?
• RQ4 Does the method reveal spatial heterogeneity in membrane activity, and can it be linked to underlying cytoskeletal dynamics?
• RQ5 Can this approach be generalized to map activity in conjunction with fluorescence imaging of specific proteins or structures?

Key findings

• The entropy generation rate was estimated at ~10⁻³ kB s⁻¹ over a 1.44 × 1.44 μm² patch, with a total of ~0.4 kB s⁻¹ across all patches, exceeding previous single-trajectory measurements.
• Active membranes showed significantly higher entropy generation rates than ATP-depleted membranes, confirming the method’s ability to distinguish non-equilibrium states.
• Spatial maps revealed localized high-activity regions (e.g., arrows in Fig. 2), which disappeared upon ATP depletion, indicating heterogeneous and dynamic activity.
• The method achieved a spatial resolution of approximately 1 μm, enabling the first direct spatiotemporal mapping of entropy generation at the cellular membrane level.
• The entropy generation rate decreased monotonically with ATP depletion time, consistent with the system approaching equilibrium.
• The results qualitatively matched predictions from a simple active membrane model, validating the approach’s consistency with theoretical expectations.