2.3.1 Cell cortical stiffness measurements

The first reports that cancer cells are softer than normal cells were published almost two decades ago [30,33,34,40,41]. Since then, multiple studies have been conducted on the mechanical analysis of established cancer cell lines and patient-derived cells. A meta-analysis on a broader collection of studies contrasting cancer and normal cell stiffness is presented in Figure 1. It comprises a total of 30 studies across six different cancer types: breast (n = 14), pancreas (n = 3), bladder (n = 8), prostate (n = 3), and ovarian (n = 2) cancer. Although there are many more published studies that report on differences depending on disease state, we focused here on studies reporting comparable parameters such as the apparent Young’s modulus E_App. Among the selected studies, AFM-based methods dominate, since AFM is by far the most commonly employed technique to assess cell mechanical properties. In Figure 1, apparent Young’s moduli are plotted for the abovementioned tumor types. Most studies focused on the assessment of established cancer cell lines, mostly breast cancer cells, and a high degree of redundancy becomes apparent. In many cases, even the same cell lines were measured by different research groups, suggesting a relatively high reproducibility of results. However, the limited number of studied models also limits the possibility of drawing more general conclusions. Of note, the prevalence of cancer cell lines compared to patient-derived material appears critical when cell mechanical parameters are to be evaluated with regard to their clinical relevance. The underlying reasons for using cell lines are obvious, since the use of clinically relevant samples is comparatively more challenging since it involves ethical approvals, collaborations with clinicians providing samples, and patient-to-patient heterogeneities, among other challenges. Nevertheless, it will be important to expand studies on patient-derived cells in the future.

Figure 1. Meta-analysis of cell indentation measurements on whole cells conducted by AFM. E_app measured for different types of cancer cells are given, with data taken from studies on breast [40,43–54], pancreas [31,54,55], bladder [30,56–62], prostate [42,46,63,64], and ovarian cancer cells [54,65,66]. Cell lines and patient-derived cells are displayed separately. Dots represent average or median values taken from respective studies. Where multiple studies report on the same cell lines, multiple dots are shown, and horizontal lines indicate medians.

In most cases, there is accordance that normal cells have higher apparent Young’s moduli compared to cancer cells. In some cases, the more invasive cell lines or metastatic primary cells had even lower stiffness values compared to less invasive cells. There are some exceptions, however, e.g., where increased cellular stiffness values are shown for invasive compared to low invasive prostate cancer cells [42]. Of note, E_App across studies widely ranges from approximately 100 to up to 30.000 Pa. Studies reporting higher values up to 100.000 Pa were excluded here, as they appear unrealistic for soft biological tissues. Generally, AFM indentation tests with sharp indenters on spread cells can result in higher values than measurements on rounded cells. In some cases, especially when spread cells with flat cellular extensions are tested, the underlying stiff substrate also might have affected the obtained values.

Although the viscoelastic nature of cells is well known [67,68], most cell mechanics surveys to date have been limited to an apparent elastic response (>80% of all studies), with only a small fraction addressing cytoplasmic viscosity properties. Here, we extended our literature search to studies that assess cell viscosity, aiming to explore its potential as a biomarker for cancer diagnostics.

2.3.2 Whole-cell viscosity

To measure the whole-cell apparent viscosity η_App, researchers employ microrheology techniques inspired by stress relaxation or creep tests, commonly used in conventional rheology [38]. Our meta-analysis includes nine studies across six different cancer types: breast, kidney, prostate, thyroid, ovarian, and liver. Overall, the data reveal a strong preference for AFM-based methods [43–45,48,69–72]. Other techniques are micropipette aspiration [72] and micro-electromechanical systems (MEMS) resonant sensor [45]. Whole-cell viscosity measurements exhibited significant variability, with η_App ranging from 3 to 500 Pa · s (3,000 to 500,000 times the viscosity of cell culture medium [73]. This variation is likely due to differences in the cell and cancer types studied, and in the case of AFM, also influenced by the specific protocols and tips employed [30,74]. The viscosity ratio η_Canc/η_Norm between cancerous and healthy cells was calculated for six cancer types to allow for comparison (Figure 2A). This figure consistently reveals a clear pattern: cancer cells exhibit lower viscosities than healthy cells. Overall, this reduction in viscosity averaged 44%, comparable to the decrease observed at the intracellular level (Figure 2B). For cancers with data on cells of increasing metastatic potential, only ovarian cancer cells (early, intermediate, and late MOSE cells) show a distinction between low and high invasiveness. This finding indicates that cancer cells undergo a fluidization of their mechanical properties compared to normal cells.

Figure 2. (A) Analysis of the viscosity ratio between cancerous and healthy cells. Whole-cell viscosity ratio η_Canc/η_Norm between cancerous and healthy cells obtained from AFM [43–45,48,69–72], micropipette aspiration [72] and micro-electromechanical systems (MEMS) resonant sensor [45] on six different cancers: breast [43–45,48], kidney [71], prostate [82], thyroid [70], ovarian [69], and liver [72]. Error bars in the left panel indicate the standard error of the mean from four different surveys on MCF-10A, MCF-7 and MDA-MB-231 cells; *- p value < 0.05. (B) Intracellular viscosity ratio η_Canc/η_Norm between cancerous and healthy cells obtained for breast [75,77–81], pancreas [80], and brain [76] cancers, where η_Canc and η_Norm denote the static or η_App, respectively. The first bar in each histogram is set to 1 by definition. The techniques used for measuring the cytoplasm viscosity were particle-tracking microrheology [78,80,81], magnetic rotational spectroscopy [77] and optical tweezers [75,76]. Error bars in the left panel indicate standard errors of means from six different surveys on MCF-10A, MCF-7, MDA-MB-468, and MDA-MB-231 cells.

2.3.3 Cytoplasmic viscosity

To measure the cytoplasmic viscosity of healthy and cancerous cells, particle-tracking microrheology, optical tweezers, and magnetic rotational spectroscopy were applied across seven studies covering three cancer types: breast, brain, and pancreas [75–81]. In these techniques, micron or submicron-sized particles are dispersed in the cytosol, and their passive or active movements are tracked, allowing local viscosity measurement. Here again, breast cancer cells dominate the data, specifically four cell lines with increasing metastatic potential, MCF-10A, MCF-7, MDA-MB-468, and MDA-MB-231, respectively, accounting for 75% of all intracellular assays. For conventions, the static shear viscosity, such as that obtained in particle-tracking microrheology and magnetic rotational spectroscopy, will be noted η₀ [38], and η_App, will be used when the experimental conditions are not low-frequency shear.

According to literature data, particle-tracking data [78–81] show viscosity values ranging from 0.3 to 6 Pa · s – 300 to 6,000 times that of cell culture medium at physiological temperature [73], while active methods report significantly higher viscosities, between 10 and 50 Pa · s [75–77]. These data heterogeneities could be attributed to the size of the probes used, from the order of a hundred nanometers for the former and a micron for the latter [77]. To account for the differences in absolute values across techniques and to allow for comparison, we expressed existing intracellular results as the ratio: η_Canc/η_Norm, where η_Canc and η_Norm represent the viscosities of cancerous and normal cells, respectively. Figure 2A displays the η_Canc/η_Norm for breast, pancreas, and brain cancer types, with the first bar in each histogram set to 1 by definition. The figure highlights a consistent decrease in η_Canc/η_Norm across all cancer types, corresponding to an average 51% reduction in viscosity. This suggests a general trend toward increased cytoplasmic fluidization in cancerous cells.

For breast cells, Young’s moduli of MCF-7 and MDA-MB-231 tumorigenic lines were lower than those of healthy cells, but the difference was insufficient to differentiate between malignant and non-malignant cells clearly. Magnetic rotational spectroscopy did, however, reveal a significant difference, with MCF-7 cells showing five times higher intracellular viscosity than MDA-MB-231 [77]. This underscores cytoplasmic viscosity sensitivity to metastatic potential, suggesting that it could serve as a mechanical biomarker for cancer cells with high metastatic potential. The preceding analysis shows that findings on cancer cell viscosity have been insufficiently documented and that further research is required, especially by expanding the scope to include more diverse cell lines and cancer types. In the following sections, we will discuss the techniques utilized to measure the viscoelastic cell properties, which are summarized in Table 1.

Table 1. The comparative analysis of the methodologies to assess viscoelastic cell properties.

2.4.1 Atomic force microscopy (AFM)

Of all the mechanobiology techniques available today, AFM is by far the most widely used [30,34,41]. AFM is an ultrasensitive technique to measure forces down to the piconewton range [83,84]. It relies on a force sensor, the AFM cantilever that can be equipped with a nano- or macro-sized indenter (e.g., a pyramidal tip or also larger e.g., a colloidal probe in the range of 5–10 μm diameter). To probe cell mechanical properties, the cantilever is brought into contact with a cell (either round or in adherent state) with a pre-set force or indentation depth, and from the resultant force-distance curve the force-indentation relation can be derived. By fitting the force indentation curve to a model of choice (most commonly a Hertz model for simple indentation tests [85,86], an apparent elastic modulus E_App can be quantified, which represents the most commonly reported parameter for cell and tissue mechanical measurements. Alternative models include, for instance, the Oliver Pharr model [11], and the liquid droplet model [87]. In a configuration where the stress/strain response is measured over time, AFM can also be employed to determine the apparent viscosity of single cells [43,44,48,70,82]. Such measurements include stress relaxation [88], creep compliance, or oscillatory probing protocols [44,89,90]. In such cases, viscoelastic mechanical response models, such as the Kelvin-Voigt, Standard Linear Liquid or Standard Linear Solid models, or more complex rheological models can be employed to derive viscoelastic cell parameters from force versus time traces [33,89]. Alternatively, information on the viscoelastic properties can be obtained from the indentation part of the curve using viscoelastic relaxation models [91,92]. The strength of AFM lies in its ability to integrate high-resolution topographical imaging with force mapping, an essential configuration for biomechanical studies. AFM can also be combined with optical imaging or spectrometric techniques, e.g., advanced optical techniques such as confocal microscopy [93], total internal reflection fluorescence (TIRF) microscopy [93], calcium imaging [42], or Raman microscopy [94] to obtain additional information on cell state. While AFM is versatile with regards to cell shapes (both rounded and adherent), it is typically restricted to probing the cell surface, and furthermore relatively low in throughput (typically 100 cells/hour).

2.4.2 Micropipette aspiration

This technique applies negative pressure to draw a part of a single cell into a narrow glass pipet [95]. The living cell, suspended in solution, is first immobilized at the pipet tip before suction is applied. Using optical microscopy, the portion of the cell entering the pipet for a given pressure is monitored over time. The kinetic time course of cell deformation resembles creep experiments in rheology, where stress is applied and deformation is recorded over time [38]. Using viscoelastic models to fit the length time series, the cell viscoelastic properties, including its apparent viscosity η_App, can be calculated [72].

2.4.3 Microfluidic techniques

To overcome the limitation of low throughput inherent to some aforementioned assays, microfluidic techniques can be attractive, e.g., for cell deformability cytometry [96–99], microconstriction arrays [100], or shear flow deformation cytometry [101] studies. These techniques have in common that suspended cells are passing through a microfluidic channel where they deform under fluid shear stress [99,101] or have to squeeze under pressure through narrow channel constrictions [100]. Resultant cell deformations or passage time are quantifiable parameters that can be assessed in timelapse videomicroscopy recordings. When the applied shear stresses are known, the viscoelastic properties of the cells can be inferred. High rates of deformation in the millisecond range (as in deformability cytometry) typically result in increased elastic moduli values compared to deformations in the second range [102]. Advantageously, these techniques can also be combined with fluorescent detection, which allows the mapping of mechanical parameters to specific markers [98]. Recent real-time deformation cytometry (RT-DC) based studies have revealed mechanical changes of transformed mammary cells [103,104], mechanical effects of different drugs on leukemic cells [100], and native and malignant cells in pleural fluids [99]. In addition, differences between normal and cancer cells were seen in cells derived from solid tumor tissues through digestion or mechanical dispersion [105].

2.4.4 Optical stretcher

Using the optical stretcher, cells are mechanically deformed in a contact-less manner using a dual-beam laser trap [106]. In a configuration combined with a microfluidic pump system, cells enter the optical flow chamber one by one, where they are trapped and stretched along the axis of the opposing laser beams [107]. The resultant cell elongation under stress is monitored over time and information about the cells’ viscoelastic properties of the cell can be quantitatively derived [108]. Using the optical stretcher, the mechanical properties of the breast [16,106], cervix [16], and brain cancer cells [26] were measured and compared to normal cells. Automated setups also provide the opportunity to assess the mechanical properties of hundreds or thousands of cells and to distinguish subpopulations to reveal cell mechanical heterogeneity [26,29].

2.4.5 Brillouin microscopy

While most of the above-mentioned techniques require physical contact between the probe and the material to be probed, optical elastography probes have the advantage that they can interrogate cells in a rather non-invasive manner. Brillouin scattering, discovered nearly a century ago, has recently been exploited for biomechanical measurements of biological specimens by Brillouin microscopy [36,109–112]. Brillouin microscopy measures the inelastic Brillouin scattering that arises from the interaction of photons with acoustic vibrations in a material. In most biomechanical studies, the Brillouin frequency shift, which is related to the longitudinal modulus, is used as a quantitative parameter to describe the elastic properties of biological specimens. Despite its non-invasiveness, an advantage of Brillouin microscopy is its high spatial resolution (in the subcellular range, depending on the optical setup), allowing for in situ 3D mapping of the mechanical properties of biological specimens. This technique can be employed on single cells but advantageously also on transparent tissues or multicellular structures such as organoids [36,111,113,114]. Calculation of a longitudinal modulus from the Brillouin frequency shift requires measurements of the refractive index, e.g., measured by optical diffraction tomography [115]. A limitation of the technique is that it is not possible to directly convert the longitudinal modulus at high frequency (GHz range) to shear moduli obtained by common probing techniques, e.g., AFM [110]. Nevertheless, previous studies have revealed correlations between results obtained by AFM and Brillouin, e.g., when osmotically perturbing cells and cell aggregates, and when applying drugs that interfere with major cellular structural elements, such as F-actin [113,116]. Differences between normal and cancer cells were also reported [36].

2.4.6 Particle-tracking microrheology

Particle-tracking microrheology has been a well-established method for several decades [39,95] to measure the static viscosity η₀ in confined environments, particularly in living cells [117]. Particle-tracking microrheology employs high-speed optical microscopy to track the movement of particles in the 100 nm range embedded in the cytoplasm. To enhance particle detection and prevent their internalization into endosomes, which can lead to directed movement of the probes, fluorescent particles are used and introduced via ballistic injection [80,118]. The particle trajectories are analyzed in terms of the mean-squared displacement, and interpreted thanks to the Stokes-Einstein equation that relate the diffusion coefficient to η₀ [119]. The technique is non-invasive, requires minimal material (<1 pL), and can be applied in both passive and active matter.

2.4.7 Optical tweezers

More recently, advanced methods such as optical tweezers [75,120] have expanded the range of intracellular rheological techniques for actively probing cytoplasmic viscous properties. This technique uses a fixed optical trap and microscope to study intracellular mechanics [120]. An infrared laser traps a micron-sized bead within the cytoplasm, allowing steady or oscillating manipulation. For viscosity measurement, the bead is displaced by moving the microscope stage, and the optical tweezers apply a spring-like force, causing the bead to relax back toward the trap center. This time-dependent relaxation is analyzed using the Standard Linear Liquid model [38] and allows simultaneous determination of the elastic modulus G₀ and apparent viscosity η_App of the intracellular medium. Experiments on breast [75] and brain [76] cancer cells reveal mitigated viscosity and elasticity properties compared to healthy ones.

2.4.8 Magnetic rotational spectroscopy (MRS)

This technique leverages the hydrodynamic instability found for anisotropic magnetic objects in a rotating magnetic field [121]. Berret et al. adapted MRS to intracellular environments using micron-sized magnetic wires that are spontaneously internalized in the cytosol and non-toxic to cells [122,123]. As the angular frequency ω increases, a transition from synchronous to asynchronous regime occurs at the critical frequency ω_C, which varies inversely with the static viscosity η₀. In viscoelastic media, the high-frequency wire oscillation amplitude varies inversely with the elastic modulus (G₀). MRS hence enables the simultaneous measurement of local viscosity and elasticity, making it a powerful tool for probing local rheological properties of fluids or solids [124,125], including those of living cells.

2.6.1 Interstitial fluid viscosity

To increase the viscosity of the extracellular medium in 2D cultures, researchers used biopolymers like alginates, polysaccharides, mucins, and polyethylene glycols, adjusting their molecular weight and concentration [149,153–156]. These experiments aimed to replicate the properties of interstitial fluid in the tumor environment, achieving viscosities as high as 2 Pa s (2,000 times that of cell culture medium) [73]. Gonzalez-Molina et al. used viscosity-enhancing polymers to simulate extracellular fluids, conducting wound healing and cell spreading assays on liver cancer cells [153]. Surprisingly, despite the increased hydraulic resistance, cell motility was significantly enhanced on 2D substrates for both healthy and cancerous cells. Structural changes at the cell level, including cytoskeleton rearrangement, cytoplasmic expansion, and nuclear flattening were also observed [153]. Maity et al. later confirmed these findings, showing a 2.5-fold increase in migration speed compared to conventional cultures [154,157]. More recently, Bera et al. conducted an in-depth study, suggesting that the increased cell motility in high-viscosity environments is driven by a mechanotransduction pathway involving cytoskeleton-ion channel interactions, calcium influx, and enhanced cell contractility [149]. Importantly, Bera et al. demonstrated that this increased motility occurs in vivo using zebrafish, chick embryos and animal models, and affects metastasis. These findings suggest that higher local extracellular fluid viscosity could be a factor in cancer cell migration within the tumor microenvironment. Given the current knowledge of interstitial fluid viscosity, technologies for measuring their flow and stress in vivo for prognostic purposes are still unavailable to patients. However, developing such methods could offer valuable insights into key aspects of metastatic biology [34]. This contrasts with blood, which also plays a crucial role in the metastasis spread and has been extensively studied for its hemorheological changes related to cancer [25,158]. Well-established research shows that rheological alterations occur in advanced cancers and are often linked to disease stage and prognosis. For various cancers, including breast, lung, ovarian, and cervical cancers [158], it has been consistently observed that whole blood viscosity is higher in cancer patients compared to healthy controls [150]. This increase is attributed to elevated plasma viscosity and red blood cell aggregation, both of which are associated with metastasis development. A recent study on whole blood viscosity in patients with hepatocellular carcinoma found that both systolic and diastolic whole blood viscosities increased by 16% and 25%, respectively, compared to healthy individuals [150]. These increases are significant enough to improve clinical diagnosis for a range of cancers.

2.6.2 Interstitial fluid shear stress

Interstitial fluid flow in tumors is driven by pressure differences between blood and lymphatic systems. The flows of interstitial fluids and blood are governed by distinct physical models based on the medium they traverse. Interstitial fluid follows Darcy law for heterogeneous media, characterized by low velocities (1 μm s⁻¹) and shear stresses in the range 0.01–0.1 Pa. Blood, in contrast, follows a pulsatile Poiseuille flow with velocities 10 to 10⁵ times higher than interstitial flow, with shear stresses range from 0.1 to 10 Pa [25]. As a result, the behavior of cancer cells within a tumor, or circulating in the bloodstream varies significantly [24,34]. Recent studies have shown that cancer cells can respond to a wide range of interstitial stresses [159–163], with experiments conducted using in vitro microfluidic devices designed to replicate the tumor microenvironment. Most commonly, single microfluidic channels are used, where cells line the channel walls and a controlled flow is applied. Relevant work includes that of Calibasi Kocal et al., who developed a microfluidic platform capable of generating low shear stress around 10^–⁵ Pa on esophageal cancer cells, showing that laminar flow induces epithelial/mesenchymal hybrid transition and increases mechanotransduction protein expression, in contrast to static cultures [159]. Other recent reports explored how shear stress in the range of 0.005 Pa to 0.5 Pa influences ovarian, breast, and prostate cancer cell behavior through the activation of mechanosensitive sodium and calcium ion channels, enhancing motility and tumor growth, or allowing tumor cells to evade high shear regions during intravasation [162,163]. The impact of fluid shear stress on cervical cancer cells also revealed that moderate shear stress boosts cell proliferation and resistance to chemotherapy, while high shear stress suppresses growth, indicating that fluid shear stress significantly influences cancer cell behavior and drug resistance during metastasis [160].

3.1.1 pH

Dysregulation of pH balance, intracellularly and in the tumor microenvironment, contributes to uncontrolled proliferation and immune evasion [183,184]. Variations in pH values become fundamental to the survival of cancer cells. Note, that the intracellular pH in healthy cells (6.99–7.05) is kept lower than the extracellular pH (7.35–7.45), while cancer cells display a “proton gradient reversal” with extracellular values (6.2–6.8) more acidic than the intracellular pH (7.2–7.8) [185]. Unlike normal cells, tumor cells thrive in environments with high lactate and hydrogen ions (H⁺) production, conditions that would typically induce cell death [186]. However, tumor cells counteract this threat with the enhanced expression of several intracellular pH-regulating systems [187]. In contrast to tumors, healthy cells can barely survive in this hostile environment [186]. pH changes in cancer cells influence the activity of cellular channels, leading to different ionic concentrations and depolarizing the cell membrane potential [175]. Additionally, studies have shown a relationship between intracellular pH changes and the dynamics of microtubules and actin filaments, further highlighting the impact of pH on cancer cell function [188,189].

3.1.2 Role of dysregulation of ion channels

Dysregulation of the functions of ion channels and transporters is a condition that supports the uncontrolled growth and spread of cancer [177]. The role of ion channel malfunction in cancer progression and metastasis is currently of great interest, considering that not all tumors share the same pattern of ion channel expression [183]. Overall, changes in calcium, sodium, and potassium promote a cascade of events that result in uncontrolled proliferation, abnormal cell physiology, and ultimately contribute to the formation of aggressive, metastatic tumors [183].

Precise concentration and localization of calcium (Ca²⁺) ions govern cell proliferation and apoptosis [196]. Cancer cells capitalize and maintain calcium levels that avoid death. Calcium levels in tumors also favor cell proliferation and angiogenesis, which enhances nutrient distribution and growth [196–198]. Cancer and wounded cells are known for excessive intake of water and sodium (Na⁺) [199,200]. This feature affects the cell towards a bigger size and more spherical shape. This altered shape influences cell signaling contributing to cancer progression and metastasis [200]. Cancer cells often exhibit enhanced expression of potassium (K⁺)-calcium (Ca²⁺) channels [183]. Altering the cell's response to hypoxia and modifying cell adhesion, migration, and apoptosis, tumor progression is favored [201–203]. The aforementioned ionic changes also influence the membrane potential. Cancer cells have a more depolarized resting membrane potential (V_m) compared to healthy cells, contributing to an increased proliferation and migration capacity of tumor cells [204].

3.1.3 Membrane potential

The membrane potential, determined by the ion concentrations, is highly affected by ion channels' permeability, expression, and activity. Cancerous cells show a tendency to have a more depolarized V_m compared to healthy cells. This depolarization is associated with poorer patient prognosis. Depolarized cells increase their proliferative behavior in cancer, influencing the cell cycle progression [204,205]. Hyperpolarization is required to initiate apoptosis, and the depolarized state disrupts this process increasing the survival rate of cancer cells [204,206]. Furthermore, the depolarized phenotype of cancer cells is associated with stem-cell-like behavior characterized by self-renewal, inhibition of cell differentiation, and migration. All these factors contribute to increased tumor aggression and metastasis [204].

3.3.1 Optical sensing techniques

Fluorescence microscopy and spectroscopy are well-developed and state-of-the-art techniques for imaging the inner cell processes at high resolution. These techniques, relying on light absorption and emission at a larger wavelength, are carefully adapted to interact with specific cell components [216]. By analyzing light intensity, wavelength, and distribution, it is possible to extract information about concentration, binding state, and environmental conditions experienced by the probe in the tumor or cancer cell (Figure 3B) [217,218].

Fluorescent pH indicators, such as SNARF-1 and the genetically encoded pH sensor SypHer2, have been used to study the reversed pH gradient in cancer and assess the efficacy of pH-regulating therapies [186,219]. By calculating the fluorescence ratio at two different wavelengths of the ratiometric dye SNARF-1, the extracellular pH of a tumor was measured in vivo [186]. In addition, genetically encoded pH nanosensors can be designed to trace single extra- to intracellular gradients and monitor intracellular pH response to different stimuli, such as treatment with pharmaceuticals. One recent example represents the nanosensor consisting of biocompatible polystyrene nanoparticles loaded with the pH-inert reference dye Nile red, and is surface-functionalized with a pH-responsive fluorescein dye. The nanosensor is equipped with a targeting moiety and can adhere to cell membranes, allowing direct measurement of extracellular pH at the cell surface (Figure 3F) [190].

Ion dynamics can also be studied through fluorescence methods. Ion concentration transients and waves have been detected through calcium-related markers probing the long-distance communication between cancer cells in human colon and prostate cancer cells [218]. Further efforts to quantitatively evaluate the tumor microenvironment concentration of potassium ions in vivo have been realized through the combination of fluorescent nanosensors and photoacoustic emission (Figure 3E) [192,193]. This analysis revealed significantly higher levels of potassium ions (29 mM) in the microenvironment compared to the surrounding muscle tissue (19 mM).

By tracking the expression of K⁺ channels tagged with green fluorescent protein (GFP) probes, researchers have studied the effects of manipulating the membrane potential in cancer cells. It was found that hyperpolarization of the membrane potential through K⁺ channels overexpression increases breast cancer cell migration, invasion, tumor growth, and metastasis [202]. The increased polarization creates conditions to upregulate cadherin-11, leading to metastatic behavior [202].

3.3.2 Magnetic resonance imaging

The electrical properties of cancer cells have been studied through the magnetic resonance imaging (MRI) technique that utilizes the interaction in high-intensity magnetic fields of magnetic moments of atoms and radiofrequency pulses [220]. It can utilize contrast agents with optimized nuclei relaxation times that enhance the contrast between specific ions, molecules, and targets (Figure 3D) [191,221].

While traditional imaging often overlooks ions, researchers have recognized that altered ion levels within tumors hold valuable diagnostic and prognostic information. Previous investigations have hinted at a prevalence of higher concentrations of sodium ions Na⁺ in cancer cells compared to the surrounding healthy tissue [199]. After analyzing orthotopic breast cancer regions with MRI combined with diffusion-weighted imaging (DWI), it was possible to confirm this observation and identify that the excess of Na⁺ is concentrated intracellularly in cancer cells (Figure 3I) [195]. By employing both techniques, enhanced levels of sodium ions were linked to increased cellularity in breast cancer models that serve as a promising biophysical marker in early tumor development [195].

Advanced MRI-based chemical exchange saturation transfer (CEST) has been shown as a promising tool for the diagnosis of metastases potential in primary tumors based on spatial acidosis studies (Figure 3H) [194]. A pH-responsive dual-mode MRI contrast agent was also demonstrated to be effective in detecting metastasis in liver tumors with sizes of 0.5 mm when using the T₁-T₂ dual-modal MRI contrast agent for cancer imaging [191].

3.3.3 Microelectronic probe

While previous methods can infer cellular electrical properties, using microelectronic techniques provides a direct characterization of these parameters (Figure 3C). Direct electrical measurements offer the advantage of real-time quantifiable dynamics of localized events [222]. Based on the fact that each biological system has distinct dielectric properties, it is possible to use them as relevant quantitative biophysical markers in cancer prognosis [223].

Microelectrodes have been engineered in recessed-tip, and double-barreled configurations to facilitate negligible perturbations during sensing with small tip sizes and simultaneous measurements of the membrane potential and pH, respectively [224,225]. Due to the compatibility with microfabrication techniques, it has been shown that it is possible to fabricate arrays of sensors for high throughput tracking of the heterogeneity of pH changes in cancer cells [226]. The activity of voltage-gated ion channels has also been related to metastatic potential in prostate cancer cells through the patch clamp technique [227]. The direct influence of bias voltage and sensing through integrated sources makes possible the control and regulation of specific ion channels [228].

4.2.1 Calorimetry

Calorimetry is a label-free technique used to measure the heat generated or absorbed by biological samples, providing insights into their metabolic activity [252]. The heat generated can reflect physiological processes in normal cell cultures but can also be employed to detect cellular responses to various stimuli, such as drugs, environmental changes, or stressors, offering a versatile tool for studying cellular behavior in different conditions. Growth of hepatocarcinoma tumor microtissues (based on HepG2 cell line) releasing a maximum thermal power was monitored, which corresponds to a diameter increase of the microtissues from ca. 100 to 428 μm [252]. Isothermal microcalorimetry, in particular, was also used to investigate the influence of enriched media compared to normal medium on neuroblastoma cells [253]. The spiking of fructose, glucose or fructose/glucose to the medium led to increases in the metabolic activity of the cells [253]. These studies also show that tested glioblastoma cells (Kelly and SH-EP Tet-21/N) preferred fructose metabolism over glucose metabolism, a beneficial adaption of these cancer cells under low oxygen and low nutrient conditions [253]. While differential scanning calorimetry (DSC) is known as a sensitive technique in the field of phase transition studies, during the last decades, it has also been recognized as a tool for cancer diagnosis and monitoring [254–256]. DSC has been tested as a tool for lung cancer patient diagnosis and prediction in clinical settings [257]. Different lung cancer subtypes and stages were successfully distinguished from control patients [257]. With further optimization DSC could provide an accurate, non-invasive, and radiation-free strategy for state-of-the-art low-dose CT scans [257].

4.2.2 Heat transfer method (HTM)

HTM is a thermal transducer platform that monitors the thermal interface resistance R_th between the solid chip and the supernatant liquid [258]. When combined with surface-imprinted polymers (SIPs), HTM has been used to distinguish between human breast cancer cell lines MCF-7 and leukemia Jurkat cells [259], and to detect differences in their glycosylation patterns using modified Chinese hamster ovarian cells [260]. The model allowed excessive expression of the transmembrane protein mucin-1 (MUC1) and control of its glycosylation by varying culturing protocols [260]. The selective differentiation between cells expressing the MUC1 protein and cells that do not express MUC1 and MUC1 glycosylated and MUC1 non-glycosylation indicate significant differences in the imprints [260]. HTM has also been effective in monitoring the quality of cell cultures over time, as demonstrated by comparisons between adherent breast cancer cell culture (ZR-75-1a) and suspension culture (ZR-75-1s) [261].

4.2.3 Modified transient plane source (mTPS) sensor

Another noteworthy technique is the mTPS sensor systems, which can combine aspects of calorimetry with the monitoring of changes in thermal properties at the chip interface. Although not yet applied to cancer detection, mTPS has been used for real-time quantification of yeast cell numbers and the metabolic activity of cultures [262]. The label-free and indirect sensing strategy is suited for bi-modal sensing, revealing both thermal and electrical information about the sample [263,264]. The integration of mTPS into microplates for high-throughput screening or microfluidic chips has also been explored [265], and therefore holds great potential for future life science applications.

5.2.1 The mechanotransduction signaling

Tumor growth is accompanied by biomechanical changes in the tumor microenvironment, including an increase in matrix stiffness generated by ECM components, and the accumulation of solid stresses from compression or tension during tumor growth [288]. Analysis of the genomic mutations in cancer of different types suggested that tumor tissue stiffness correlates with the scale of genomic variations and intratumoral heterogeneity [289]. This increased genomic variation can be associated with increased DNA damage in response to the cell and nuclear squeezing upon cell proliferation in a stiff microenvironment or migration through tiny pores of the collagen-enriched ECM or basement membranes [289,290]. The stiffness of tumor tissues depends on the deposition of the ECM components such as collagen, laminin, elastin, fibronectin, and glycosaminoglycans (GAG) by cancer-associated fibroblasts (CAFs) and tumor cells and ECM remodeling by enzymes such as matrix metalloproteinases (MMP), procollagen-lysine, 2-oxoglutarate 5-dioxygenase (PLOD) and lysyl oxidase (LOX) secreted by tumor cells and tumor-associated stroma cells [22,291]. Mechanical stress induces actin stress fiber reorganization in the tumor cells, consequently affecting cell elasticity [292,293]. Mechanical stress also induces large-scale chromatin remodeling and changes in gene expression profiles [294]. Mechanical stimuli can activate cell membrane mechanosensors such as transient receptor potential (TRP) channels and Piezo ion channels [295], integrin receptors [296], G-protein coupled receptors (GPCR) [297] and cadherins [296,298], and corresponding cytoplasmic mechanotransducers including RhoA-mediated actin dynamics [299], focal adhesion kinase (FAK) complexes, transcription factors (e.g., yes-associated protein (YAP)/WW domain-containing transcription regulator 1 (TAZ), mechanistic target of rapamycin complex 1 (mTORC1) [300], phosphoinositide 3-kinase (PI3K)/Akt [301], and β-catenin [298], which induce other downstream signaling mechanisms in tumor and stroma cells [302,303]. The mechanotransduction signaling is activated in most types of cancer. It is involved in the regulation of different aspects of cancer progression, including tumor growth, therapy response, immune evasion, metastasis, and stemness [304].

5.2.2 Mechanical properties of CSCs

Indeed, for many cancer types, CSCs are shown to be the dynamic and plastic cell populations. The epigenetic changes, newly acquired mutations, and microenvironmental factors such as hypoxia, nutrient availability, physical parameters including stiffness, shear stress, and microenvironment architecture, as well as different types of treatments, can induce a bi-directional transition between CSC and non-CSC states. CSC surface markers such as CD44, CD133, integrin α2β1, and integrin β4 directly interact with collagen and GAG hyaluronan (HA) and induce intracellular signaling critical for CSC maintenance such as FAK, Akt/mTOR, MAPKs and β-catenin [305–308]. Indeed, increasing stiffness leads to the enrichment of CSC populations in breast [281], liver [309], and lung tumors [282], as reviewed recently [310]. High stiffness triggers TAZ-mediated mechanotransduction [281] and activates the YAP/β-catenin-dependent transcriptional program and expression of the reprogramming factors Nanog, Oct4, and Sox2 [281,282]. Integrin-mediated activation of FAK and downstream pathways, including AKT/mTOR signaling, contribute to CSC regulation through matrix stiffness [311]. Mechanotransduction mediated by Piezo ion channels is critical for the dissemination of cancer cells [162,312–315], and the high expression of the Piezo mechanoreceptors was correlated with worse clinical outcomes and metastases in many cancer types [313,315–319]. Piezo proteins, the mechanosensitive ion channel, are sensing cell deformation and transforming mechanical stimuli such as pressure or shear stress into biochemical processes. In particular, activation of Piezo channels leads to the influx of extracellular ions, mainly Ca²⁺, triggers the RhoA pathway regulating actin cytoskeleton [299,320], and activates different intracellular signaling mechanisms, including integrin/FAK, Akt/mTOR, and MAPKs [321–323]. Piezo proteins regulate the maintenance of CSC populations. Genetic deletion of Piezo1 inhibits CSCs in glioblastoma and colon cancer preclinical models [317,324]. The expression levels of Piezo1 or Piezo2 correlate with CSC-related protein markers and transcriptional signatures in patients with colorectal and gastric cancer, respectively [318,324]. TRP family of mechanoreceptors also plays a critical role in the regulation of CSCs, as recently reviewed [325].

Given the importance of mechanotransduction for CSC maintenance, it is unsurprising that CSCs can be identified and isolated based on their intrinsic mechanical properties: several studies demonstrated that CSCs are a mechanically soft population compared to non-CSC counterparts. The soft tumor populations were isolated from breast cancer cell lines by using microfluidic devices and characterized using in vivo models. These experiments demonstrated that soft tumor cells are more metastatic and tumorigenic in vivo and possess CSC properties [326]. The previous studies reached a consensus conclusion that metastatic tumor cells are much softer than non-metastatic cells [11,47,65,106], and as discussed in Section 2. Tumor cell softness was associated with an oxygen deficiency called hypoxia, another feature of aggressive, metastatic, and therapy-resistant tumors [11,327]. In each individual tumor, different spatially defined areas may be perfused differently by functional blood vessels [328]. Therefore, hypoxia has substantial inter- and intratumoral heterogeneity [329]. The hypoxic microenvironment promotes CSCs and metastatic properties, and one of the key mechanisms of this regulation is the activation of the transcriptional program driven by hypoxia-inducible factors (HIF). Both hypoxia and ECM stiffness induce CSCs, tumor growth and metastases through the mechanotransduction pathways such as β1-integrin/integrin-linked kinase (ILK)/PI3K/Akt mechanism [330]. In a feedback loop of this interplay, Piezo1-mediated mechanotransduction is shown to induce HIF1α expression [313].

5.2.3 Dynamic changes of the CSC mechanical properties in the metastatic cascade

A subset of CSCs determines metastatic growth through sequential steps, including tumor cell intravasation into the bloodstream, dissemination of the CTCs to distant organs, extravasation at the distant site, and initiating metastatic tumor development [246,331]. At the initial stage of the metastatic progression, immobile cancer epithelial cells acquire mesenchymal phenotype and the ability to migrate in a process called epithelial-mesenchymal transition (EMT) [332,333]. The EMT process is associated with generating a range of intermediate cell phenotypes, including both epithelial and mesenchymal features. EMT is positively associated with other tumor hallmarks, such as genomic instability and hypoxia. This developmental mechanism is associated with an epigenetic reprogramming of the bulk tumor cells into CSCs [333,334] and contributes to the heterogeneity of CTCs [335]. Therefore, EMT is considered to be a key driver of intratumoral heterogeneity [336]. The ECM stiffness activates the EMT program via Piezo1/2-mediated Ca²⁺ influx and mechanosensitive ephrin type A receptor 2 (Epha2)/Lyn protein complex followed by the activation of the TGF-β and Akt signaling pathways and transcriptional program driven by Twist1 and Snail [214,316,337–340]. Consequently, the activation of the EMT program softens the cytoplasm [341,342], induces actin reorganization, and increases actomyosin traction forces [342], playing a key role in cell migration [343], and, therefore, facilitates tumor cell invasiveness. At the intravasation stage of metastatic progression, the highly migratory and invasive tumor cells leave the primary site and invade through the basement membranes to the blood of lymphatic vessels. The stiffness of the ECM is a limitation encountered by invading cells, and only a subpopulation of tumor cells resistant to deformation and able to move through stiff microenvironment and dense basement membrane might survive and propagate metastases. As discussed in Section 2, highly metastatic cells have lower elasticity and viscosity compared to less metastatic tumor cells. The EMT program induced by stiff microenvironments makes prostate tumor cells more migratory than in soft substrate, enabling them to escape. The physical cell compression upon squeezing through narrow gaps in the wall of blood vessels is associated with nuclear envelop rupture (NER), temporary disruption of a lipid bilayer surrounding cell nuclei [290]. NER is associated with DNA damage and is potentially leading to the genomic heterogeneity of the metastatic cells [344]. The recent studies described the adaptive longitudinal dynamics of single tumor cell viscoelasticity under different pressures in the microvessels and softening of tumor cell nuclei during transendothelial migration [345,346]. The cells with softer nuclei have advantages in cell migration but a higher possibility of DNA damage and genome instability [347]. In addition to its effect on DNA integrity, the stiff microenvironment also induces mechanical adaptations of tumor cells through epigenetic reprogramming. A study of tumor and non-tumor epithelial cell lines revealed that their exposure to a stiff matrix induces migration driven by the activation of the YAP transcriptional program. The cells retained enhanced actomyosin expression and nuclear YAP translocation induced by a stiff matrix even after exposure to a soft secondary matrix, suggesting that matrix stiffness induces mechanical memory through epigenetic changes [348,349]. The high viscosity of the tumor interstitial fluid is associated with the leakage of the lymphatic vessels and increased degradation of ECM, leading to the accumulation of macromolecules [149]. This elevated interstitial viscosity induces TRPV4 activation and RhoA-mediated cell contractility. Breast tumor cells exposed to elevated viscosity exhibited increased migration and extravasation potential and high lung colonization in murine xenograft models. Tumor cells previously exposed to elevated viscosity retained high migratory potential after switching the medium viscosity to baseline level. This viscous cell memory was mediated by TRPV4 and Hippo signaling [149]. Thus, mechanical memory in the form of epigenetic adaptations induced by the primary tumor microenvironment can influence tumor cell survival during the entire metastatic journey and colonization of distant organs [149,349].

Tumor cells can spread from the primary site as single cells or clusters and can be detected in the bloodstream in the form of single or clustered CTCs [328]. A minimally invasive method of the CTC enumeration in peripheral blood samples called liquid biopsy can be used as a prognostic marker in patients with breast, prostate, and colorectal cancer [350]. It is based on immunomagnetic capture of cells positive for the epithelial cell adhesion molecule (EpCAM) in the blood samples. However, some CTCs are EpCAM negative as a result of their phenotypic plasticity, such as EMT [351]. This plasticity is induced by different microenvironmental stimuli, including mechanical stresses. The CTCs in the bloodstream are exposed to hydrodynamic shear stress (HSS), which can affect the metastasis-initiating properties of these cell populations. Indeed, HSS has been shown to trigger EMT phenotypes [159,352], and induce CSC properties through the inhibition of the extracellular signal-related kinase (ERK)/glycogen synthase kinase (GSK)3β signaling [353]. Shear stress induces a number of signaling pathways, including PI3K/Akt, MMPs, and YAP/TAZ [354–356], and promotes tumor-forming and migration potential [353,355,357]. Therefore, CTC exposure to the HSS in the bloodstream might induce their reprogramming toward more aggressive metastasis-inducing populations. In contrast to the highly invasive soft tumor cells, CTCs are mechanically robust [299], allowing them to withstand HSS. Some CTCs are protected from HSS-mediated damage by inducing RhoA signaling, increased formation of cortical F-actin, and activation of myosin II [299]. Despite the adaptive mechanisms, metastatic colonization is a rare event, and less than 0.01% of tumor cells entering the bloodstream give rise to macroscopic metastases [358]. Different microenvironmental factors, including HSS and cell compression upon squeezing through the wall of blood vessels and small-sized capillaries, might induce NER and cell death. On the other hand, mechanical deformation selects the tumor subpopulation with resilience to cell death induced by mechanical deformation. This cell population possesses highly activated DNA damage response, upregulated proliferation, and resistance against chemotherapeutic drugs [359]. Thus, the biomechanical characteristics of tumor cells are dynamically changing through epigenetic reprogramming or selection processes depending on the applied physical forces. This biomechanical plasticity provides tumor cells with high adaptability during tumor growth and metastatic dissemination in the dynamic and heterogeneous microenvironment [34,360] (Figure 6).

Figure 6. Tumor cell adaptability to the mechanical stresses. Dynamic changes in the tumor cell properties provide high adaptability to mechanical stresses during tumor growth and metastatic dissemination. CAF: cancer-associated fibroblasts; CTC: circulating tumor cell; ECM: extracellular matrix; EMT: epithelial-mesenchymal transition; FAK: focal adhesion kinase; HSS: hydrodynamic shear stress; MAPK: mitogen-activated protein kinase; MMP: matrix metalloproteinases; mTORC1: mechanistic target of rapamycin complex 1; LOX: lysyl oxidase; PLOD: procollagen-lysine, 2-oxoglutarate 5-dioxygenase; TAZ: WW domain-containing transcription regulator 1; TRP: transient receptor potential channels; YAP: yes-associated protein 1. Created with BioRender.

5.2.4 Electrical features of CSCs

Electric fields regulate the key biological functions of cells, including proliferation, stemness, and differentiation [361,362]. In addition to their distinct mechanical properties, CSCs have unique electrical characteristics. As discussed in Section 3, cancer cells possess unique bioelectric properties, making them distinct from their normal counterparts. The bioelectric properties of tumors and CSCs are defined by the above-mentioned Ca²⁺ ion channels TRP, Piezo as well as K⁺, Na⁺ and Cl⁺ channels, the intracellular and extracellular pH, and mitochondrial characteristics including pH, membrane potential, and ion concentrations [183,202,325,363,364]. Consequently, the seminal studies demonstrated the unique bioelectrical features of CSC cells.

A depolarization of the cell membrane was suggested as a marker of CSC populations [204], whereas induction of cell membrane depolarization by a suitable electrical environment was associated with CSC differentiation in glioblastoma (GBS) and was proposed as one of the therapeutic approaches to eliminate CSC populations [365]. The electrostatic potential difference (EPD) measured on cryosections of different tumor tissues correlates with tumor spread and CSC abundance, suggesting the prognostic potential of EPD measurement [366]. Furthermore, galvanotaxis study revealed that CSCs migrate to the regions with negative electrostatic potential [366].

A single-cell impedance measurement revealed that CSC and non-CSC populations from liver and oral squamous carcinoma cell lines can be differentiated within the impedance magnitude ranging between 2 and 20 kHz [283]. The stem cell phenotype of these cells was verified by the analysis of CSC biomarkers. Another study measured the bioelectrical impedance of non-cancerous breast epithelial cells and breast cancer cells with different metastatic potential in the range of 200 kHz-5 MHz. This study revealed a high correlation of the conductivity (σ) and permittivity (ε) to the tumor cell characteristics, including metastatic potential and migration capacities, and expression levels of different markers indicative of cell proliferation, microenvironment hypoxia and acidification, and glucose metabolism. Highly metastatic breast cancer cells have shown higher relative ε than non-metastatic tumor cells, and this difference is even more prominent compared to normal epithelial cells. These distinct electric properties can be attributed to the activation of the ion channel proteins such as Na⁺/H⁺ exchanger 1 (NHE1), which transports a proton to the extracellular space, high lactate secretion and, correspondingly, increased extracellular acidification [165]. Hence, this methodology can be used to monitor CTCs in the bloodstream as a prognostic marker of metastatic tumors. The importance of electrical signaling for tumor progression was demonstrated by finding that metastatic potential can be inhibited by non-contact induced electric fields (iEFs), suggesting a new treatment strategy to prevent tumor dissemination [367].