Transient Plasmonic Photothermal Imaging Deciphers Dynamic Thermal Transfer during the Cell Cycle

Abstract

A stable thermodynamic environment is a key prerequisite for the orderly conduct of metabolic activities within cells. For cell division, a common but important life-sustaining process the balance between heat generation and dissipation is especially important. Here, we measure the thermal conductivity within single cells and reveal dynamic thermal regulation during the entire cell cycle by using transient plasmonic photothermal microscopy with the help of fluorescence-image-based cell cycle analysis. We find that the thermal conductivities within cells are highly heterogeneously distributed, and some sites in the perinuclear region have relatively low values (as low as 0.035 W·m^–1·K^–1). During different stages of the cell cycle, the thermal conductivity of the cell is dynamically adjusted (with a fluctuation of 0.063 W·m^–1·K^–1) with the level of energy metabolism to maintain thermodynamic robustness. We further perform metabolomic analysis during the cell cycle using single-cell mass spectrometry and extract relevant metabolite pathways. These results provide important insights into how cells adapt and regulate their internal heat balance to ensure the robustness of thermal adaptability.

1. Introduction

A robust thermodynamic adaptability is crucial to ensure highly ordered metabolic processes in biological systems. Heat is an important byproduct of intracellular biochemical reactions, which, in turn, warms the organism and maintains body temperature by balancing energy metabolism and heat dissipation. As a basic thermodynamic parameter, the temperature within cells is closely related to cellular metabolic activity and even physiological changes. Particularly, numerous studies have shown that tumors exhibit heat-seeking behavior due to the higher metabolic activity in cancer cells, , e.g., the metabolic rate increases by 10% when the temperature rises by 1 K. , Excessive intracellular temperature, such as during a fever, can lead to many serious pathologies. An in-depth understanding of the intracellular thermodynamics and regulation is crucial to decipher the mechanisms behind diseases such as fever and even actively regulate our body temperature.

Tremendous progress has been made in recent years to study the temperature within cells. Several nanothermometers have been established based on thermosensitive materials, including luminescent molecules, fluorescent proteins, inorganic nanoparticles, , and microscale thermocouples. The measurement accuracy could be up to 1.8 mK, and the temporal resolution is as fast as 4 ms. Many results show that different regions within cells exhibit significant temperature differences. , More recently, Jin and co-workers detected a huge temperature variation of 2–7 K in lysosomes and mitochondria inside HeLa cells under Ca²⁺ ion shock or treatment with an oxidative phosphorylation inhibitor. Chrétien et al. even found a greater temperature increase of more than 10 K in mitochondria of human embryonic kidney 293 cells and primary skin fibroblasts when the respiratory chain was fully activated. It follows that the intracellular temperature is not constant but changes dynamically with cellular metabolism, energy uptake, stress response, and cell cycles. , However, it is theoretically difficult to understand such significant temperature fluctuations inside of cells. Based on existing thermodynamic models, the temperature gradient inside the cell has been estimated to hardly exceed 1 K. , An important reason for the above controversy is the insufficient understanding of local heat production and heat dissipation within cells. Since temperature is the apparent result of heat production and dissipation, it is hard to get a full picture of intracellular thermodynamics by measuring temperature alone. ,

Measuring heat transfer properties within cells could help clarify the above mystery. Kim’s group measured the thermal conductivity of single cells via a 3ω-based thermal sensor, , and they found that the thermal conductivity of human cells (0.575 to 0.604 W·m^–1·K^–1) is close to that of pure water (0.607 W·m^–1·K^–1). Note that since the 3ω method does not have spatial resolution, it only shows the apparent averaged thermal conductivity of a whole cell. However, the thermal conductivity of local regions inside cells could be significantly different from that of water since it is well-known that the interior of cells is crowded and filled with various biological macromolecules, which will definitely affect heat transfer. Carlos’ group constructed an upconversion-based nanothermometry and measured the thermal conductivity of the lipid bilayer of the cell membrane, and they got a value of 0.20 ± 0.02 W·m^–1·K^–1, which is about 3 times lower than that of water. Suzuki et al. obtained a thermal conductivity of 0.11 ± 0.04 W·m^–1·K^–1 in cells most likely at lysosomes or endoplasmic reticulum at the perinuclear region. Our group has previously constructed a transient photothermal microscopy method and measured the subcellular thermal conductivity. We observed a significantly heterogeneous distribution of the thermal conductivity within the cytoplasm, ranging from 0.16 to 0.42 W·m^–1·K^–1. We also found a cellular-level regulation of the thermal conductivity in warm-blood animals according to ambient temperature. These results clearly imply that living cells cannot be viewed simply as a kind of biological material and evaluate their thermal conductivity in static. The thermal conductivity within cells may depend on different subcellular regions or be dynamically regulated by the cell activity and energy metabolism. In particular, the cells are not immutable and can divide into two daughter cells, which is fundamental to the continuation of life. And, cells must undergo the entire cell cycle (from G₁ to S phase, G₂, and M phases) to divide into two daughter cells. Numerous studies indicate that physical parameters, including mechanical elasticity, intracellular viscosity, and temperature, also undergo dynamic regulation as the cell cycle progresses. − These changes are driven by alterations in chromatin organization, cytoskeletal reorganization, biosynthetic activity, mitochondrial output, and membrane composition, , which may lead to variations in cellular heat transfer properties. Unfortunately, thus far, how heat transfer is regulated in response to cellular energy levels during different phases in the cell cycle phases still remains ambiguous.

Here, we have investigated the dynamic regulation of cellular thermal conductivity during different cell cycles using our transient photothermal imaging method. We found that the thermal conductivity within the cell is highly heterogeneous rather than uniformly distributed, and the values in different regions can differ several times. Membrane organelles such as endosomes and lysosomes have a lower thermal conductivity than the cytoplasm, with an average value of about 0.16–0.21 W·m^–1·K^–1, which is consistent with Carlos’ results. We obtained a value of local thermal conductivity as low as 0.035–0.088 W·m^–1·K^–1 in some regions near the cell nucleus, which is the lowest value reported so far. The thermal conductivity within cells is dynamically regulated with the cell cycle, cell energy level, cell metabolism, etc. These unique thermal regulations jointly maintain the intracellular thermodynamic environment and ensure the robustness of intracellular chemical reactions.

2. Results and Discussion

2.1. Principle of Mapping Cellular Thermal Conductivity

We utilized the previously reported transient photothermal microscope (Figure S1, Supporting Information) to reveal the thermal conductivity distribution in single cells during different cell cycles. Principally, the cooling lifetime (τ) of a hot body could be considered inversely proportional to the thermal conductivity (κ) of the surrounding medium: 1/τ∝κ. Then the thermal conductivity distribution inside a cell could be obtained by measuring the cooling lifetimes of the hot nanoprobes internalized in the cell. Here, a transient photothermal microscope system based on the “pump–probe” technique was established (Figure a), where a 532 nm nanosecond pulsed laser and a pulsed xenon lamp are used as the pump source and detection beam, respectively, and gold nanoparticles act as the nanoprobes (Figure S2, Supporting Information). Since the gold nanoprobes are heated, the “nanolens” , effect formed by a nonuniform thermal field in the surrounding medium will cause the enhancement of its scattering light signal. Yet, the scattering signal from nanoprobes gradually decreases during the cooling process, as the homogenization of the thermal field leads to a gradual dismissal of the scattering light (Figure b). The optical signal (Φ) of the nanoprobe changes during the cooling process (time = t) after heating (time = 0) can be simplified to an exponential decay: Φ­(t) = Φ(0) e^–t/τ (from eqs S1 and S2 in the Supporting Information). In order to simultaneously obtain information on intracellular nanoprobe position and their cooling time characteristics, a color charge-coupled device (CCD) camera and an intensified CCD camera were equipped for capturing the colorful dark-field image and the time-dependent scattering images of the nanoprobes during the cooling process after the pulse laser heating (Figure c,d). Then, by extracting the scattering light intensities of the nanoprobes and fitting them exponentially to the time scale, the thermal conductivity distribution within a single cell can be mapped (Figure e).

1.

Principle of the previously proposed transient heat transfer microscope system. (a) Schematic diagram of the principle of thermal conductivity measurement within a single cell. (b) Schematic illustration of the strategy for monitoring the generation and decay process of the nanolens effect around the gold nanoprobe. (c,d) Color (c) and general dark-field images (d) of a HeLa cell during the cooling process of the gold nanoprobes. (e) Remapping of thermal conductivity distribution in a single HeLa cell.

2.2. Heterogeneous Thermal Conductivity in Cells

It is well-known that even in such a small single-cell unit, different regions have different subcellular structures and functions, which may lead to different heat generation and transfer characteristics. For instance, we have preliminarily found that the thermal conductivities in the edge region of cells are higher than in the center region. In order to accurately study the heat transfer characteristics of cellular microregions, here, we designed a cell internalization experiment by varying the coincubation time with nanoprobes, through cellular endocytosis. As visualized in Figure a, the gold nanoprobes are taken up by the cell during the formation of a vesicle through the inward depression of the cell membrane. Then, the gold nanoprobes were transported to the lysosome and even the endoplasmic reticulum and other organelles near the nucleus region. This is evidenced by the colorful dark-field images that we captured over time (Figure b–d).

2.

Variations of thermal conductivities with the spatial location within single cells. (a) Schematic diagram of the cellular endocytosis process of gold nanoprobes. (b–d) Color dark-field images of the HeLa cell after incubation with gold nanoprobes for 6, 8, and 10 h. (e) Relative distance of intracellular gold nanoprobes from the nucleus in HeLa cells with different incubation times. (f) Thermal conductivity distribution measured from the gold nanoprobes at different incubation times in HeLa cells. (g,h) Position labeling image and fitted scattered light variation plots for four characteristic gold nanoprobes in a single HeLa cell.

Most of the swallowed nanoprobes distributed near the cell membrane region at the initial stage (after about 6 h of coincubation). Subsequently, the nanoparticles were gradually transported to the inner zone after further prolonging the incubation time. For statistical analysis, we defined a parameter of relative distance r _t = r ₁/r ₂ to represent the distance of the nanoprobe from the nucleus, in which the r ₁ and r ₂ are the distance of the nanoprobe to the nucleus center and the radius of the nucleus in the nanoprobe direction, respectively. As the transmission process of each nanoprobe is not synchronous, the relative distances r _t are not strictly uniform but spread over a wide range, regardless of the incubation time (Figure e). But the median values of r _t gradually decrease with the extension of incubation time, indicating that most nanoprobes are transported to near the nucleus region.

We then measured the corresponding thermal conductivity values (κ values) of the microregion around the nanoprobes during the internalization stage by using the transient microscopy method described above. The κ values exhibit a wide distribution at each incubation time, demonstrating that the thermal transfer characteristics inside the cell are heterogeneous and closely related to the environment of the microregion (Figure f). The median κ values gradually decrease with the inner migration of the nanoprobes, and the averaged κ values are about 0.17 W·m^–1·K^–1 after incubation exceeds 10 h. This value is very close to the results of 0.2 W·m^–1·K^–1 reported by Carlos’ group. In addition to the averaged value, the thermal conductivities of some typical areas were exhibited in Figure g,h. We found some areas within the cell with particularly low thermal conductivity, ranging from 0.035 to 0.11 W·m^–1·K^–1. The value of 0.035 W·m^–1·K^–1 we observed near the perinuclear region is the lowest value reported so far, which is also much lower than the result of 0.11 W·m^–1·K^–1 reported by Suzuki’s group. Notably, there are microtubule organizations, membranous organelles such as the endoplasmic reticulum around the nucleus, and many macromolecules involved in the exchange of substances in and out of the nucleus as well as local phase separation zones. All of these may be responsible for the high degree of heterogeneity in the thermal conductivities. The extremely low κ value measured may result from the fact that the nanoprobes are trapped by membranous organelles like the endoplasmic reticulum or Golgi apparatus. Another possible reason for the heterogeneity of the thermal conductivity is the heterogeneity of energy metabolism in different regions within the cell.

2.3. Change of Thermal Conductivity along with Cell Cycles

Cell duplication and division, which can even extend back in time to the beginning of life, is a simple fact but the key process for keeping multicellular mammals surviving by replacing cells that die. Cell cycles are divided into four typical stages of G₁, S, G₂, and M phases during the cell duplication and division process, and the energy metabolism and component preparation for each phase are different. Then, clarifying the thermodynamic process during different cell cycles can help us to master the changing rule of heat transportation, which may contribute to active control of body temperature.

To recognize the cycle phase of each cell, we developed a fluorescence-assisted cell cycle imaging method based on cell staining and image analysis. G₁, S, G₂, and M phases are the four main stages in the cell cycle. Among them, the G₁ and G₂ phases are the two gap phases before and after the DNA synthesis process (S phase), through which the DNA content in cells gradually doubles. Similar to the principle of flow cytometry, we used the DNA staining dye (DAPI) to stain the nucleus and analyzed the DNA content of each cell to determine the cell cycle (Figure a,b). The slight difference is that flow cytometry is done by recording the fluorescence intensity of each cell that flows through, while our method measures the fluorescence intensity of the nucleus of each cell by fluorescence imaging analysis. In this way, we could obtain the cycle in which each cell is in and at the same time measure the thermal conductivity of that cell. After the fluorescent intensity of each cell in the visual field was extracted, we then obtained the cell cycle mapping image of the cells (Figure c). Subsequently, we validated the accuracy of the developed fluorescence-assisted cell cycle imaging method using a commercial flow cytometer as a reference. Briefly, we randomly divided the same cell sample into two and fixed both samples at the same time point. The overall distribution of their genetic material in each sample was measured and counted by our developed method and by a flow cytometer. As shown in Figure d,e, the distribution curves obtained by the two techniques are almost the same, indicating that the developed method can accurately obtain the cycle stage of individual cells. It is noteworthy that a conventional flow cytometer usually cannot distinguish between G₂ and M phases because they have the same DNA content. However, our imaging method allows such distinction because the morphology of M phase cells has a unique dual-nuclei feature (Figure c).

3.

Principle of the developed fluorescent flow imaging system for characterizing the cell cycle. (a,b) The color dark-field image (a) and fluorescence image (b) of a HeLa cell after being stained with nucleus dye (DAPI). (c) Calibration results of the cell cycle phases depending on the cellular morphology and fluorescence intensity. (d,e) Cell cycle distribution plots by using the developed fluorescence-assisted cell cycle imaging system (d) and the commercial flow cytometry (e).

Then we investigated changes in cellular heat transfer characteristics with the cell cycles. The four strips in Figure a represent the statistical analysis result of thermal conductivity in cells that are in four different cell cycles (G₁, S, G₂, and M phases), and the dots in the middle of the strips are the median value. Likewise, the intracellular thermal conductivity still shows a broad distribution, regardless of which cell cycle they are in. Interestingly, the median value of intracellular thermal conductivity is not constant across cell cycles, in which the G₁ phase exhibits the maximum median κ value of about 0.196 W·m^–1·K^–1. Then the median κ value tends to decrease when the cell enters the S (0.147 W·m^–1·K^–1) and G₂ phases (0.161 W·m^–1·K^–1). The lowest median κ value was found in the M phase, which is about 0.133 W·m^–1·K^–1. Notably, the four phases of the cell cycle are not evenly distributed, in which the G₁ phase lasts for the longest time of about 9 h. Therefore, without the use of cell cycle synchronization reagents, most of the measured intracellular heat transfer data originated from the G₁ phase. Furthermore, we have also investigated the effect of different cell cycle synchronization reagents on cellular thermal conductivity. We chose hydroxyurea (Hu), colchicine (Col), and thymidine (TdR) to control most of the cells stalled in the same cell cycle, and the synchronization efficiency is more than 85%. As exhibited in Figure S3 (Supporting Information), the differences of κ values caused by the different reagents are negligible. It is demonstrated that the regulation and fluctuation of thermal conductivity in different cell cycles are intrinsic features of the cell and are not affected by the use of synchronization reagents.

4.

Energy metabolism and thermal regulation in the single cell. (a) Changes in intracellular thermal conductivity distribution during cell cycle evolution. (b,c) Glucose uptake (b) and ATP accumulation (c) during the cell cycle. (d) Changes in the thermal conductivity of the matrix with different phases and biomolecular doping density.

2.4. Energy Metabolism and Thermal Regulation within the Cell Cycle

The stabilization of the thermal environment inside a cell is the combined result of heat production and heat dissipation. We then investigated the energy-supplied material uptake and energy transformation ratio within the cell cycle, aiming to unravel the mystery of how cells maintain the balance of their thermodynamic environment. Glucose is the main energy-supplied substance for cell activities; especially, more active cancer cells consume more glucose than normal cells. We measured the glucose uptake during the entire cell cycle (Figure b) and found that the total cellular uptake of glucose is increasing from the G₁ to M phase, which is consistent with the result of HeLa cells characterized with fluorodeoxyglucose (FDG) tracers. This is understandable because glucose is the main source of energy, and the cell grows progressively larger as the cell cycle continues and therefore has a greater need for energy. , The ingested glucose needs to be converted into energy currency (ATP) that can be directly used for cellular biological activities. We found that ATP/ADP exhibits the highest value in the S phase (Figure c), indicating that cells require large amounts of energy to complete the replication of genetic materials. Interestingly, we found that the relative change in the ATP/ADP value is much larger than the increase in glucose uptake from the G₁ to S phases. This may be due to the fact that aerobic glycolysis and the metabolism of amino acids such as glutamine are also involved in cellular energy supply. Except for the S phases, the ATP/ADP value in the G₂ phase is also higher than that in the G1 phase. It is probably because more energy is still required at this stage to facilitate the synthesis of RNA and proteins required for cell division.

In principle, the thermal conductivity of a medium depends on the composition and phase of the medium. Unlike homogeneous liquids, the presence of a large number of biomolecules in the cytoplasm may lead to a phase transition between the liquid and gel phases. , Then, we performed a phase transition experiment in agarose and trehalose solutions and gels to simulate the cytoplasm, aiming to clarify the influence of the phase transition on the heat transfer property. As displayed in Figure d, the value of thermal conductivity decreases sharply from 0.30 to 0.12 W·m^–1·K^–1, as the trehalose solution transits to the gel phase without any change in solution concentration. Moreover, the thermal conductivity will be further reduced to about 0.09 W·m^–1·K^–1 when the biomolecules (BSA) are further added to the gel to form a local phase separation.

Combined with previously reported temperature changes in the cell cycle, , intracellular energy metabolism and heat transfer regulation together constitute thermal regulation to maintain robust control of the temperature in cells. Although we do not yet have a full picture of intracellular temperature regulation, our current results clearly indicate several important pieces of evidence and perspectives. First, the thermodynamic environment inside the cell is not homogeneous but highly heterogeneous. The local temperature of the nucleus has been reported to be significantly higher than that of the cytoplasm. , Here, we find a lower thermal conductivity near the perinuclear region, which contributes to maintaining the nuclear temperature gradient. Second, cells could dynamically adjust heat production and dissipation based on the level of energy metabolism (Figure S4, Supporting Information). Here, we demonstrate the dynamic regulation of thermal conductivity during the cell cycle according to glucose uptake and ATP levels. Third, the regulation of the thermal conductivity in the cell may be closely related to the phase transition and phase separation in the cell. Our results clearly show that the formation of gel phases and multiphase systems significantly affects the thermal conductivity properties.

2.5. Substantial Metabolism within the Cell Cycle

Changes in the content of metabolites are the signs and material basis for fluctuations in energy metabolism within the cell. We then investigated the substantial metabolism changes during the cell cycle by using vacuum ultraviolet laser desorption/ionization reflection time-of-flight mass spectrometry (VUVDI-RTOF-MS). , Combined with the fluorescence-assisted cell cycle imaging, single-cell mass spectra at different cycles can be obtained, followed by omics analysis. The mass spectrometry samples were first tested by fluorescent imaging to identify the cell cycle distribution (Figure a–c). The same cell sample was then analyzed by the VUVDI-RTOF-MS system, with a scanning step of 500 nm, and the mass spectrum information on each pixel was obtained by laser ablation three times. After data integration and image redrawing, we could map the intracellular distribution of substances corresponding to any m/z (Figure d) and the mass spectra corresponding to any pixel inside the cell (Figure e).

5.

Changes in intracellular substantial metabolism during the cell cycle. (a–d) The bright-field image (a), color dark-field image (b), fluorescence flow imaging (c), and single-cell mass spectrometry intensity map (d) of the different charge-to-mass ratios for a HeLa cell. (e) The full mass spectrum of a certain position in the single HeLa cell. (f) PCA of the single-cell mass spectrometry results in different cell cycles. (g,h) Heat maps of the substances with the highest weights along the PCA1 (g) and PCA2 (h) axes change during the cell cycles.

The mass spectra data corresponding to cells in different cell cycles were then performed via PCA. As shown in Figure f, the S and M phases have a large overlap with the G₂ phase and are more clearly distinguished from the G₁ phase. This phenomenon indicates that the metabolism of cells in the G₁ phase is significantly different from that of the other three phases, which is similar to the results of the changes in the cellular thermal conductivity.

Subsequently, the weight values of each axis of the PCA graph were extracted separately and heat maps were drawn for detailed analysis. Variations of substance content with the cell cycle for the top 11 mass spectrometry peaks with the highest weights along the two axes are shown in Figure g,h. The results of the single-cell mass spectrometry analysis fully illustrate the shift in the metabolic process during the cell cycle. First, the branched-chain amino acids (BCAAs), such as leucine/isoleucine (Leu/Ile) and valine (Val), are downregulated when the cell cycle enters the S phase (Figure g). It suggests that most BCCAs are catabolized during this phase to supply the N– and C– resources, as well as the energetic resource, for synthesis of a large number of substances in this phase. At the same time, an upregulation of alanine (Ala) in the S and G2 phases is observed (Figure h), predicting an increase in the rate of aerobic glycolysis in both phases. A more detailed omics analysis is shown in Table S1 and the related discussion in Supporting Information. According to the metabolic results, we find that the metabolism of substances and energy changes dynamically along the cell cycle, and these changes ultimately lead to fluctuations in intracellular temperature and thermal conductivities.

3. Conclusion

Using transient photothermal microscopy, we systematically investigated the distribution of thermal conductivities inside cells and their variation with the cell cycle. First, the thermal conductivities within cells are highly heterogeneous; in particular, the maximum difference could be up to several times at different locations. This phenomenon may arise due to the fact that different regions within the cell consist of different types and contents of biomolecules, leading to different liquid–liquid and liquid–gel phases in the localized regions within the cell. Second, the heat transfer properties of membrane organelles (e.g., endosomes and lysosomes) were measured using the cell endocytosis model by varying the internalization time of nanoprobes. The mean value of thermal conductivities is about 0.16–0.21 W·m^–1·K^–1, which is close to the result of 0.2 W·m^–1·K^–1 reported by Carlos’ group. Remarkably, we found that some sites in the perinuclear region have a very low thermal conductivity of 0.035–0.088 W·m^–1·K^–1, which is the lowest κ value reported thus far. This may go some way to explain the higher temperature of the nucleus than the cytoplasm discovered by some research groups. , Third, we found that the cell could regulate the heat dissipation property according to the energy level and metabolism during the cell cycle to maintain an appropriate internal thermodynamic environment. Metabolites and metabolic pathways related to energy metabolism during the cell cycle were further confirmed by metabolomic analysis based on single-cell mass spectrometry. These results are fundamentally vital to enlighten the study of intracellular thermal adaptation and regulation, which is crucial to decipher the mechanisms behind diseases such as fever and even actively regulate our body temperature.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/cbmi.5c00180.

• Experimental details, materials, and methods, including gold nanoprobe preparation steps and characterization results; brief building procedures of the homemade imaging system and the VUVDI-RTOF-MS system, as well as the measurement protocols; and detailed cellular metabolism analysis results (PDF)

∥.

H.G. and P.S. contributed equally to this work.

The authors declare no competing financial interest.