Exploring cell and tissue mechanics with optical tweezers

Robert Hooke coined the term ‘cell’ in 1665 and formulated his famous law of proportionality between stress and strain in 1678. We now know that all animals, plants and fungi consist of these discrete units, the cells. However, their response to mechanical stress is more complicated than Hooke's law. Cells are composed of molecular complexes that are constantly turned over, reshaped and regulated by active processes that sustain life. The mechanics of these complexes, their interactions and activity provide cells with unique material properties that allow them to resist and transmit mechanical information. But what are these properties?

The cell is a heterogeneous, asymmetric, anisotropic and active medium (Charras et al., 2005; Efremov et al., 2019; Latorre et al., 2018; Turlier et al., 2016), for which no unified theory exists (Dumont and Prakash, 2014). Consequently, we need to broaden our views of cell mechanics away from continuum mechanics and advocate that mechanical properties need to be measured and characterized inside cells with spatial and temporal control. However, only a few techniques allow the measurement of subcellular mechanics non-invasively, for instance with three-dimensional (3D) degrees of freedom to apply strains and compressive, tensile and shear forces over a broad force and frequency range that matches molecular and cellular processes (Roca-Cusachs et al., 2017; Wu et al., 2018).

Here, we review the recent advancement in optical trapping technology, which allows force measurements with piconewton resolution and mechanical manipulations with nanometer precision directly in living cells, organs, and animals. Importantly, to visualize how manipulations affect cell behavior, these measurements can be combined with optical microscopy and other complementary techniques. We discuss relevant challenges and pitfalls and how they can be controlled to unlock the full potential of in vivo optical trapping.

Art (Arthur) Ashkin invented optical tweezers and first succeeded to trap and manipulate live cells (Ashkin and Dziedzic, 1987; Ashkin et al., 1986, 1987; see also the Nobel Lecture 2018 at https://www.youtube.com/watch?v=vb58pd2ycKQ). To form an optical trap for biological applications, a near infrared laser is focused into or near a cell using a high numerical aperture (NA) microscope objective (Box 1, left-hand side of figure; for detailed descriptions see, for example, Bustamante et al., 2021; Gieseler et al., 2021; Mahamdeh et al., 2011; Neuman and Block, 2004). Near the focus, transparent particles – often microspheres with a diameter of ∼0.5–1.5 µm – can be held by the light as it transfers its momentum, an optical force, onto the particle. These optical forces are proportional to small displacements of the trapped particle, independent of direction so that optical tweezers are 3D Hookean springs. For example, for the x-direction, the force is F=kx, where k is the spring constant, the so-called trap stiffness. Although the trap stiffness scales with laser power and depends on direction (Bormuth et al., 2008), typical values for a trap stiffness are 0.1–1.0 pN/nm. With Hooke's law being valid up to ≈100 nm (Box 1, right-hand side of figure), forces up to 100 pN can be exerted or measured and higher forces up to the escape force can be measured beyond the linear regime (Box 1, 2) (Farré et al., 2017). Robust calibration techniques now enable reliable force and displacement measurements inside cells (for details see Box 2). Furthermore, optical tweezers can be combined with a wide range of imaging or other force-measurement techniques, allowing researchers to obtain complementary, multimodal data to determine the mechanical properties of cells.

The cellular response to force typically depends on the amplitude, force profile, history, and frequency or duration of the stimulation. Various modalities of optical tweezers have been established to probe and map the mechanical properties of cells (Fig. 1). To extract the stiffness of a sample, a trapped microsphere is driven into a region of interest with a constant speed defining a force ramp (Han et al., 2018; Venturini et al., 2020; Falleroni et al., 2018). The probe is retracted when a user-defined setpoint force or displacement from the trap center is reached (Fig. 1A). In this manner, a force–deformation curve is recorded. Imaging of large cell deformations by spinning disk confocal (Català-Castro et al., 2021) or light-sheet microscopy (Chardès et al., 2018) aids the real-time analysis of force-induced cell shape changes. Depending on the mechanical behavior of the biosystem, either a certain continuum model can be used to obtain quantitative information, or model-independent parameters need to be extracted to compare different experimental conditions. In general, a larger slope in the force–deformation curve indicates a higher sample stiffness and, thus, readily provides insight into structural strength of force transmission (Han et al., 2018, 2020) and mechanical resistance of the cytoplasm (Hu et al., 2017). The same approach can be used to measure a viscoelastic response, for example, by driving the trap with increasing velocities into the structure of interest.

If the optical trap is equipped with a force-feedback system (Bugiel et al., 2017), the force can be clamped to a user-defined setpoint (Lang et al., 2002) and duration while the resulting probe displacement is recorded (Fig. 1B). After the response time and assuming a linear-elastic (Hookean) solid that is probed with a constant area, the strain is directly proportional to the stress (see Glossary), with the proportionality constant being the Young's modulus (see Glossary) of the sample (Fig. 1B). In a purely viscous (Newtonian) fluid (see Glossary), a constant stress results in a continuous flow of the material. Without restoring forces, the initial position will not be recovered once the stress is released and the applied force is set to zero. In a viscoelastic material (see Glossary), a constant stress results in a larger deformation as the material flows and results in a creep compliance curve J(t). To determine the frequency-dependent rheological spectrum (see Glossary), this curve can be transformed (see below). After release of the force at the end of the experiment, the deformed material partially relaxes back towards its starting geometry (Fig. 1B). However, due to energy dissipation, the starting position might not be recovered. The force clamp is also obtained in the so-called optical stretcher – a setup of two counter-propagating laser beams that is frequently used to retrieve the mechanical footprint of non-adherent cells in a high-throughput manner (Ekpenyong et al., 2012; Guck et al., 2005).

Experimentally easier to implement is the trap position clamp (Fig. 1C). Here, the sample is indented to a constant value by a fast movement of the trap to a new position while the resistance and force relaxation is recorded. The relaxation time (see Glossary) can be extracted and provides insight into the viscosity of the sample. Note that the time to relocate the trap needs to be shorter than the relaxation time of the viscous or viscoelastic sample response. The conversion of the force into a stress yields the stress-relaxation curve (Khalilgharibi et al., 2019; Tassieri et al., 2016).

A special implementation of a position clamp is membrane tether extrusion force spectroscopy (Fig. 2A). While recording the force, a lipid membrane nanotube is pulled from the plasma membrane of cells in vitro (Das et al., 2021; Datar et al., 2015; Nussenzveig, 2017; Raucher and Sheetz, 2000), or inside living animals (Johansen et al., 2016). The obtained force curves contain rich information about the membrane–cortex interface. The initial formation of the membrane tube is hindered by an energy barrier (Koster et al., 2005) that provides insight into local membrane curvature (Ma and Berro, 2021). The force needed to further extrude the membrane tube provides important information about membrane–cytoskeleton coupling (Sheetz, 2001) and the lipid reservoir (Dai and Sheetz, 1995a; Raucher and Sheetz, 1999), as well as membrane tension, bending rigidity and membrane viscosity (Hochmuth et al., 1996). With higher pulling speeds, forces increase as lipids need to diffuse into the tether, membrane viscosity contributes significantly and proteins accumulate at the tether neck (Tian and Baumgart, 2009). Importantly, once the trap is stationary, the extrusion force relaxes to a constant value (Fig. 2A) that depends on the membrane tension and is largely independent of the pulling history (Dai et al., 1998). Using an appropriate mathematical model, it is possible to relate the force and velocity to the density of attachment points between the membrane and the cytoskeleton (Brochard-Wyart et al., 2006).

Because the mechanical properties emerge from their constituent parts, every organelle has a distinct rheological spectrum that depends on the measurement frequency (Chen et al., 2010), location and age (Khan et al., 2019; Majumdar et al., 2018). Owing to the small probe size and 3D navigability, optical tweezers can perform spatially resolved measurements to map mechanics within the same cell (Mandal et al., 2016). In an active microrheology experiment, the frequency-dependent viscoelastic properties can be derived from a trapped microsphere that is oscillated at various frequencies with small displacement amplitudes that fall within the linear regime (Box 1) (Guo et al., 2014; Tassieri, 2019). Depending on the mechanical property of the material, the measured force either follows the microsphere position perfectly (purely elastic) or lags behind resulting in a phase difference or lag δ (Fig. 1D). In a purely viscous material, this phase lag is exactly 90° or π/2, meaning that the force is proportional to the speed. In general, for a viscoelastic material, the phase lag has an intermediate value and is recorded for different frequencies, yielding information about the complex shear modulus (see Glossary) of the material (Fig. 2E). The complex shear modulus = at a given angular frequency has a real and imaginary part that describe the capability of the material to store, i.e. the storage modulus [], or dissipate, i.e. the loss modulus [], mechanical energy, respectively (Ayala et al., 2016).

The rheological spectrum for viscoelastic materials shows that the shear modulus is not a material constant, as it changes with the strain rate – the modulus strongly depends on the frequency. Thus, the material can be soft at low frequency and impenetrably hard at high frequencies. It is common to use empirical power-law models, for example ⁠, to extract information from the measurements. For each modulus, the exponent α can be obtained from the slope of the curve in a double logarithmic plot (Fig. 1E). Its magnitude informs whether liquid (α→1) or solid (α→0) properties dominate in the material (Fig. 1E). Intermediate values indicate that the material is viscoelastic. Furthermore, the exponent itself can depend on the frequency (Blehm et al., 2016; Bonfanti et al., 2019; Hurst et al., 2021) (Fig. 1F). In this case, a simple power law is not sufficient to describe the mechanical response, and more sophisticated mechanical models need to be considered (Bonfanti et al., 2020). The lag δ is related to the ratio of the loss over the storage modulus ⁠, also called the loss tangent. Values larger than one indicate that energy dissipation is dominant corresponding to a more liquid-like behavior. For example, fluidization (tan δ>1) at higher frequencies might indicate a destabilization of intermolecular bonds under stress (Hurst et al., 2021; Lee et al., 2010). Conversely, stress-stiffening owing to reorientation and entanglement (Khan et al., 2019) or catch-bond kinetics and motor-based prestress of cellular material (Chen et al., 2010; Han et al., 2018) might lead to a frequency dependent loss tangent.

Most cells are sufficiently transparent so that stable optical traps can be generated inside the cytoplasm of isolated cells. Unlike atomic force microscopy, which is limited to the cell surface (Krieg et al., 2019), with the cantilever obstructing simultaneous optical access, optical tweezers offer the opportunity to explore individual subcellular compartments directly with minimal mechanical and optical interference of other cellular structures (Fig. 2A–F).

Much of our understanding of how biological membranes respond mechanically has been derived from the extrusion of lipid tethers (or nanotubes; Fig. 2A) from cells (Dai and Sheetz, 1995b; Hochmuth et al., 1996; for a recent review, see Nussenzveig, 2017). In particular, tether extrusion showed that local differences in plasma membrane tension regulate many cellular processes, including polarized cell migration (Houk et al., 2012; Lieber et al., 2015), stress relieve (Sinha et al., 2011), parasite infection (Kariuki et al., 2020) or phagocytosis (Berghoff et al., 2021; Masters et al., 2013), and exo- and endo-cytosis (Gomis Perez et al., 2022), as well as mechanosensitive ion channel gating (Das et al., 2021). Despite these insights, it is still disputed how stress distributes within the membrane (Cohen and Shi, 2019).

Many cells have a large plasma membrane reservoir, stored as caveolae or other membrane invaginations, that is able to buffer changes in membrane tension (Raucher and Sheetz, 1999; Singh and Lamaze, 2020). However, the existence of such a reservoir is inconsistent with a long-range membrane-based transmission of mechanical stress, as any difference in tension would act as a sink for superfluous membrane. Thus, local alterations in tension are rapidly damped over long distances and could prevent a fast mechanical signaling across the cell surface (Shi et al., 2018).

The combination of optical-tweezer-based tether pulling with multimodal imaging has afforded new insights into the regulation of force-activated ion channels. By exerting forces locally on the membrane, the spatial extent and the direction of the tension gradient can be investigated. High-tension gradients during membrane extrusion reduces the basal Ca²⁺ activity in proprioceptive neurons through activation of K⁺ ion channels, whereas membrane tether relaxation caused a strong increase in local Ca²⁺ signaling through activation of TRP ion channels (Das et al., 2021). Such experiments point to mechanically compartmentalized axons that can process multiple stimuli at the same time (Krieg et al., 2022).

Actin, microtubule and intermediate filament networks have all been demonstrated to contribute to cytoplasmic strength, organelle stability and flow in a cell type-specific manner (Guo et al., 2013; Hu et al., 2019; Mandal et al., 2016). Actin-based traction forces inside lamellipodia and filopodia can be measured using optical traps, while simultaneously observing protein dynamics (Leijnse et al., 2014; Mehidi et al., 2021). Indeed, forces during actin dynamics and their coupling to the WAVE regulatory complex (WRC), a complex that links actin to various cell surface receptors (Chen et al., 2014), have recently been mapped in the lamellipodium of cultured cells. By specifically attaching functionalized microspheres to modified cell surface receptors, bound to the WRC on the intracellular side, the specific interaction strength of the receptor with the cortical actin cytoskeleton during protrusion formation was measured with an optical trap (Mehidi et al., 2021). Forces above 2 pN, close to the polymerization force of individual actin filaments, lead to detachment of the receptor from WRC, suggesting that force on the WRC controls lamellipodium formation and migration (Mehidi et al., 2021).

Optical tweezers have also been used to directly explore force transmission pathways through the actin cytoskeleton. For instance, microspheres functionalized with the actin-binding drug phalloidin have been delivered into cultured cells by microinjection and used as direct handles to pull on actin stress fibers (Hayakawa et al., 2008). To visualize the consequences of this force, intracellular calcium dynamics were simultaneously recorded by total internal reflection fluorescence (TIRF) microscopy in a multimodal setup. A pulling force of 5.5 pN opened mechanosensitive transduction channels close to focal adhesions, suggesting that mechanical stress was directly transmitted through the actin cytoskeleton to activate mechanosensitive ion channels (Hayakawa et al., 2008).

Because microinjection of microspheres into cells is laborious, force measurements are frequently performed on phagocytosed microspheres (Blehm et al., 2013; Hendricks et al., 2012; Rai et al., 2013) (Fig. 2C), endogenous cellular lipid droplets that bidirectionally move along microtubules (Shubeita et al., 2008; Sims and Xie, 2009) or artificial organelles (López-Quesada et al., 2014), all of which can be trapped easily based on their refractive index mismatch with the surrounding medium. Compared to endogenous vesicles, microspheres have a larger refractive index and enable higher forces. Independently of whether phagocytosed microspheres or lipid vesicles are used to manipulate their motor-driven transport, the stepping forces and kinetics has been shown to be consistent with in vitro dynein and kinesin experiments (Fig. 2C) (Rai et al., 2013; Sims and Xie, 2009). Remarkably, it is possible to obtain mechanical information inside cells with single-molecule precision down to forces of 1 pN (Rai et al., 2013). The effect of the collective motor force generation can also be directly measured with active rheology, and the cytoplasm of cancer cells becomes more fluid-like upon inhibition of kinesin motors (Mandal et al., 2019).

The cellular response to forces is determined by the interaction kinetics between the proteins within the force transmission pathway. Because individual interactions can react to force either with an increase or decrease in lifetime (see Box 3), the application of a piconewton force might influence the turnover of proteins in a cellular complex (Novikova and Storm, 2013). Note that based on the rate-dependent kinetics of single-molecule interactions (Merkel et al., 1999), mechanosensitive processes might be more sensitive to certain frequencies but insensitive to others. For example, focal adhesions grow larger if force is applied at a higher frequency, even if the magnitude of the applied force is unchanged (Fig. 2D) (Andreu et al., 2021). To account for this effect, experiments need to be carried out over a large range of frequencies, strain rates or loading rates.

The cytoplasm has complex mechanical properties, and it is far from being a homogenous medium. Its spatiotemporal organization depends on membrane-enclosed organelles and phase-separated membrane-less condensates. Even though these condensates await detailed mechanical characterization inside cells, in vitro optical tweezers experiments have shown that many of these are viscoelastic materials with a strong frequency-dependent storage and loss modulus (Alshareedah et al., 2021; Ghosh et al., 2021; Jawerth et al., 2018, 2020). Many have been shown to stiffen under mechanical stress (Shen et al., 2020), while others change their relaxation time as they age. This property could have profound consequences on their biochemical functions (Jawerth et al., 2020) and mechanotransduction (Sanfeliu-Cerdán et al., 2022 preprint).

To investigate the local mechanical environment in the vicinity of the probe, embedded microspheres or lipid vesicles can be used as passive force probes (Turlier et al., 2016; Zia, 2018). Here, the optical trap is stationary, and the Brownian and cell-generated motion are detected (Mason and Weitz, 1995; Tassieri, 2019). By a power spectral analysis of the tracked motion, the frequency-dependent response of the surrounding cytoplasm, for example, the power-law behavior of its complex shear modulus, can be obtained. The resulting exponent α is characteristic for the material properties of the viscoelastic medium (α=0 elastic, α=1 viscous, and α>1 active, non-equilibrium material). Often, an exponent of α=0.75 is found, which is characteristic for semi-flexible polymers, such as the actin or extracellular matrix (ECM) network, for example, in the tail and brain of zebrafish (Blehm et al., 2016). This approach has provided insights into the local mechanical properties in cellular protrusions, such as axial tension, bending flexural rigidity and plasma membrane tension (Gárate et al., 2015; Valentino et al., 2016), as well as how cytoplasmic rheology correlates with the migratory cell behavior of cancer organoids (Reuten et al., 2021; Wullkopf et al., 2018), or tissue formation in developing animals (Dzementsei et al., 2022). However, passive microrheology approaches are not valid universally; especially at low frequencies, active energy consumption of the cell increases the mechanical motion of the tracer particles (Guo et al., 2014). This leads to a violation of the fluctuation–dissipation theorem (Turlier et al., 2016), meaning that ATP consumption by motor activity increases the fluctuations compared to those expected at thermal equilibrium. Also, the small, passive amplitudes limit detectable displacements in very rigid environments. Thus, care needs to be taken when interpreting passive microrheology measurements (Mizuno et al., 2008).

An alternative to passive is active microrheology; this is where a trapped microsphere is actively oscillated while force and displacements are simultaneously recorded (see Fig. 1D above). Active microrheology has led to new insights into the physics of living matter and how energy consumption modifies the viscoelastic response of the cytoplasm (Guo et al., 2014; Mandal et al., 2019) and its mechanical organization (Ahmed et al., 2018; Hu et al., 2017). For example, it has been shown that the cytoplasm fluidizes during cell division (Hurst et al., 2021). Fluidization reduces the drag on chromosomes while they move through the viscous material and facilitates cell division (Fig. 2E). Conversely, the cytoplasm stiffens during water efflux, due in part to an increase in intracellular molecular crowding (Guo et al., 2017).

The nucleus is the biggest cellular organelle and has received much attention in mechanobiology. It is a transducer of mechanical stress and significantly contributes to the mechanical properties of the whole cell (for reviews, see Nader et al., 2021; Niethammer, 2021). Atomic force microscopy, micropipettes and magnetic tweezers have been used, but these methods also deform the cell membrane and cytoplasm. In contrast, optical tweezers can directly probe the nucleus without unwanted activation of cell-membrane-mediated mechanotransduction. For example, the mechanical properties of the nucleus have been shown to be the origin of a proprioceptive signal in response to 3D compression, which triggers early events during confined migration in development (Venturini et al., 2020) (Fig. 2F). It has been found that nuclei deform non-linearly and display rapid viscoelastic relaxations, a property that appears to be unaffected by modest 3D confinement (Català-Castro et al., 2021). How confinement in general affects the rheological response of the nucleus is unknown.

Changes in nuclear mechanics also relate to cancer progression, for example, in the transition from grade III to IV glioma (Alibert et al., 2021). Based on active microrheology measurements of the nucleus, nuclear stiffness increased in grade IV compared to grade III glioma, whereas the bulk cytoplasm and actin cortex softened (Alibert et al., 2021). This counterintuitive result highlights that the mechanical properties of different cellular organelles and compartments might be differentially affected during a physiological or pathological process. Such local changes are difficult to measure by other techniques. Although most of these experiments use spontaneous microsphere internalization, subcellular particles are sometimes also used. For instance, in mouse oocytes, the nucleolus has a different refractive index to the surrounding nucleoplasm. Thus, nucleoli could be trapped and used as a microrheology probe for the nucleolus–chromatin complex (Syrchina et al., 2020). This approach is particularly valuable, as delivery of microspheres into the nucleus is inherently difficult.

In plants, tip growth depends on the nucleus position (Ketelaar et al., 2002). Its manipulation in root cells with optical tweezers stopped growth. Here, nucleus positioning likely did not involve actin bundling, hinting at a novel mechanotransduction pathway between the cell surface and the nucleus (Ketelaar et al., 2002).

Owing to their excellent spatial resolution, optical tweezers offer the possibility to manipulate organelle-organelle interactions, map membrane contact sites and dynamically reposition individual components with respect to one another (Fig. 2B) (Gao et al., 2016; Guet et al., 2014; Kim et al., 2015; Osterrieder et al., 2017). In tobacco leaf epidermal cells, it was recently shown that mitochondria interact with the endoplasmic reticulum (ER), which affects mitochondrial fusion through the small GTPase Miro (White et al., 2020). Furthermore, chloroplasts strongly attach to the ER with forces exceeding 400 pN, presumably reflecting the site of inter-organelle lipid exchange (Andersson et al., 2007). Less is known regarding organelle mechanosensation in mammalian cells. In a pioneering study, locally applied force at the membrane of mesenchymal stem cells activated Ca²⁺ dynamics originating from ER-based ion channels (Kim et al., 2015). Such an activation might be important for stem cell differentiation. Taken together, the versatility of optical-tweezer-based force measurements has afforded unprecedented insight into intracellular mechanics. Future work will help to dissect the contribution of different organelles to cellular mechanotransduction (Feng and Kornmann, 2018).

In contrast to various other force spectroscopy techniques (Roca-Cusachs et al., 2017; Wu et al., 2018), optical tweezers have the ability to measure and apply forces inside cells of tissues, embryos and moving animals (Fig. 3). In transparent animals, optical tweezers even enable measurements through tissue boundaries, for example, the characterization of complex fluid dynamics of native blood flow (Harlepp et al., 2017).

Therefore, complementary to microscopy methods, optical tweezers enable the characterization of the relationship between structure and mechanics, and how this relationship tunes the force sensitivity of the cellular functions within their respective native environments. For instance, to investigate the forces that control stability and morphodynamics during embryogenesis and record how local deformations propagate throughout the tissue, optical tweezers have been used to deform cell–cell junctions in combination with light microscopy (Fig. 3A) (Bambardekar et al., 2015; Ferro et al., 2020). In these experiments, no microspheres were used, as the cell junction itself could be trapped and deformed by the trapping laser. Creep compliance measurements and the resulting viscoelastic properties of cell junctions have helped to explain morphogenetic events, such as cell intercalations during germ band extension in fruit flies (Bambardekar et al., 2015) or primitive streak formation in chicken embryos (Fig. 3A) (Ferro et al., 2020).

Light-sheet microscopy together with optical tweezers has been used to manipulate mechanically sensitive processes, while measuring behavior and whole-brain Ca²⁺ activity inside living zebrafish (Fig. 3B) (Favre-Bulle et al., 2017, 2018). Like other vertebrates, zebrafish use vestibular otoliths, composed of a gelatinous matrix, in each ear for 3D orientation in their environment. A 5 pN force exerted on the otoliths evoked complex calcium dynamics in the brain and a behavioral response, visible as eye movement and compensatory tail displacement in a restrained zebrafish (Favre-Bulle et al., 2017).

In intact plant tissue, optical tweezers have been used to manipulate organelle–organelle interactions (Fig. 3C) (Leitz et al., 2009). Mechanical deformations of the ER with trapped intracellular statoliths leads to a long-lasting mechanoresponse that affects gravitropism in root cells of Arabidopsis thaliana, possibly through gating of ER-resident mechanosensors (Leitz et al., 2009; Sparkes, 2018). The force necessary to reposition and deform other plant organelles are in the lower piconewton range, making optical tweezers, for example, combined with confocal microscopy an ideal tool to investigate gravitropism and other mechanical plant processes (Abe et al., 2020).

During zebrafish development, optical-tweezer-based active and passive microrheology has shown that gut progenitors acquire higher elasticity than their surrounding cells. These changes were attributed to their microtubule cytoskeleton (Dzementsei et al., 2022), while differences between brain and tail vasculature might be due to their specific ECM (Blehm et al., 2016; Staunton et al., 2019). The emergence of certain properties is not restricted to development but also occurs during pathological processes, such as cancer morphogenesis. To understand the mechanical behavior of cancer metastasis, rheology measurements have been carried out in primary tumors (Reuten et al., 2021; Staunton et al., 2016) and cancer spheroids (Han et al., 2020; Wullkopf et al., 2018). In such spheroids, different tissue regions adopt distinct mechanical properties (Fig. 3D) that correlate with their invasive potential (Wullkopf et al., 2018). For example, independently of the cancer type, the invading tip cells displayed a higher scaling exponent α than the cells located in the center of the organoids, suggesting that a higher fluidity correlates with invasiveness (Wullkopf et al., 2018). These experiments support the hypothesis that cells with a higher deformability have a higher propensity to migrate through tissue barriers and confined spaces. Finally, the ability of optical tweezers to perform non-invasive operations with high precision have been exploited in nano-surgical applications. Blood clots have been removed from obstructed arteries of living mammals (Fig. 3D) (Zhong et al., 2013). These proof-of-principle experiments demonstrate the versatility and power of optical tweezers.

Despite the promises optical tweezers hold for intracellular force and displacement measurements, several important points need to be considered when planning, executing and analyzing optical mechanics measurements inside cells and animals (Català-Castro et al., 2021; Pompeu et al., 2021). Those include changes in light momentum due to the refractive index mismatch inside cells, the choice of the probe particle size and its refractive index, laser-induced heating, optical aberrations when working in deep tissue, and the rheological response of active matter.

The cytoplasm has a higher refractive index (n_cyto is ∼1.38) compared to that of cell culture medium (n_m of ∼1.33, close to that of water) (Schürmann et al., 2016). Thus, the trapping laser is deflected at the cell surface, eventually leading to a change in light momentum. Even though this deflection can be exploited to apply forces to objects in cells without a microsphere, the additional deflection when using microspheres may lead to an underestimation of force. This effect can be noticed when approaching a cell from the outside as an attractive force prior to actually making contact with the cell membrane (Riesenberg et al., 2020). When working with cells in vitro, refractive index-matching media could be used to reduce this mismatch (Boothe et al., 2017).

Even though measurements of light momentum enable calibration-free force measurements on particles of arbitrary size and shape (Bui et al., 2018; Català et al., 2017a), the maximum force is limited by particle size (Bormuth et al., 2008). However, large microspheres are difficult to target directly inside cells, might interfere with biological processes and have a nonlinear force-displacement relation (Box 1, right-hand side of figure). Conversely, inside cells, it is difficult to trap small particles with a diameter much smaller than the trapping wavelength. First, Brownian forces might exceed optical escape forces. Second, when focusing through layers having different refractive indices, aberrations might weaken the trap. Depending on whether a high trap stiffness or a high escape force is desired, the optimal choice of particle size can be different. The highest trap stiffness occurs at the Mie resonance, where the particle size matches the trapping laser wavelength in the medium. Thus, for a 1064-nm laser, a microsphere diameter of 800 nm in water or 770 nm in cytoplasm results in the highest force within the linear optical force regime (Box 1) (Bormuth et al., 2008; Mahamdeh et al., 2011). However, force–displacement relationships are nonlinear when approaching maximal forces (Box 1, right-hand side of figure). Within the linear range, forces can be increased up to 40% based on the choice of laser polarization matching the measurement direction (Bormuth et al., 2008). Thus, depending on the application, different microspheres might be desirable. To guide the range of accessible parameters, several computational optical tweezer toolboxes allow the exploration the degrees of freedom in silico (Callegari et al., 2019; Lenton et al., 2017; Nieminen et al., 2007). To aid the investigations of mechanical properties of specific cellular structures, microspheres can be coated with biomolecules. Such functionalized microspheres can then be enriched at actin cables or at cell surface receptors (Hayakawa et al., 2008; Mehidi et al., 2021; Meloty-Kapella et al., 2012).

Compared to other manipulation techniques, a major drawback is the low force limit of optical tweezers – usually less than 400 pN for a laser power of <1 W in the laser focus. As a consequence, deformations are small with the benefit that linear elastic regimes of biological materials are probed. Higher forces, or the same forces with less light and photodamage, can be achieved by using cleverly engineered, high-refracting micro- or nano-spheres (Cordova et al., 2018; Ferro et al., 2016; Jannasch et al., 2012; Sudhakar et al., 2021), large microspheres trapped by structured light (Taylor et al., 2015) or off-resonance trapping of upconversion nanoparticles (Shan et al., 2021). Such particles have not yet been used inside cells, but they should be well-suited to apply high forces onto organelles or probe higher frequency regimes in a rheology experiment with lower laser power as compared to a direct manipulation of cellular structures with lower refractive index. Generally, the refractive index of the trapping probe should be as high as possible. For example, microspheres composed of melamine (n=1.68) have a higher trap stiffness than those of polystyrene (n=1.59) having the same size (Català et al., 2017a). However, backscattering of light from microspheres with a very high refractive index limits stable trapping, and antireflective coatings are necessary (Bormuth et al., 2008; Jannasch et al., 2012).

With increased laser power, sample heating may cause a heat-shock response (Hörner et al., 2017; Leitz et al., 2002) and a decrease in viscosity that might affect calibration (Peterman et al., 2003). Most trapping probes are transparent and heating is mostly due to the absorption of the trapping laser by water (≈10 K/W; Peterman et al., 2003). Metallic particles, however, should be avoided, especially inside cells (Andres-arroyo et al., 2015). Thus, overall heating can be minimized by the choice of the most efficient optical trapping probe and laser wavelength (Català et al., 2017b; Leitz et al., 2002).

Thick or highly curved specimens might deflect the trapping laser due to the cell curvature or an optically dense particle within the optical path. Opaque samples or turbid tissues also limit the penetration depth of the trapping laser (Favre-Bulle et al., 2019). Thus, an important requirement for the use of optical tweezers is the transparency of the sample and optimal optics to generate a stable trap. Stable traps require a high NA lens for focusing. However, high-NA oil immersion lenses introduce spherical aberration when trapping inside aqueous environments, degenerating trap quality (Fig. 3Dii; Zhong et al., 2017). With optimal aberration correction, high-NA water immersion lenses, and index-matching media, deeper trapping beyond 150 µm should be feasible (Dzementsei et al., 2022). Also, wavefront-shaping techniques using adaptive optics can be applied to minimize aberrations due to scattering, and thereby allow better trapping at increased penetration depths (Favre-Bulle et al., 2019).

The mechanical complexity of biological cells poses a challenge for the interpretation of rheological data (Bonfanti et al., 2020; Verdier, 2003). The non-linear, non-isotropic and history-dependent active response (see Box 3) can be teased apart in force spectra, but care needs to be taken in their evaluation. For example, non-linear visco-plastic effects can be tracked by an incomplete recovery of the shape compared to baseline deformations or forces (Català-Castro et al., 2021). In addition, strain stiffening or stress softening can become visible upon an increase or decrease in the force during repetitive indentations, respectively (Andreu et al., 2021; Mehidi et al., 2021). Likewise, active processes violate the fluctuation–dissipation theorem, and material properties cannot be derived from passive microrheology experiments alone (Turlier et al., 2016). Thus, the age and metabolic state of a tissue culture needs to be carefully controlled to prevent artifacts and ensure reproducibility of measurements. In summary, more advanced models to relate force and deformation in complex media and at interfaces, for example, at the cell surface or the nuclear envelope, are needed, which combine genetics (Das et al., 2021), pharmacological perturbations (Venturini et al., 2020) and clever experimental interrogations (Mehidi et al., 2021) to decipher the mechanical footprint of living materials.

Optical tweezers can be used to create 3D viscoelasticity maps inside living cells and complex mechanical environments. The small size of a laser focus and the ability for subdiffraction sensing and manipulation with nanometer precision and subpiconewton force resolution make optical tweezers a perfect tool for the study the mechanics of cells. Direct momentum-based force measurements enable quantitative measurements inside cells and allow force measurements beyond the linear Hookean regime. The temporal resolution and fast response time of optical tweezers with high refractive index nanospheres allow the mapping of the microsecond dynamics of walking molecular motors (Sudhakar et al., 2021). Furthermore, the non-invasive operation with infrared light enables deep cell penetration, can resolve single-molecule forces in living cells, and does not spectrally overlap with common genetically encoded fluorescent biosensors. Thus, multimodal interrogation of force and function in living specimens are possible. In the future, the use of novel nano- and micro-particles as trapping probes, combined with adaptive optics, will allow researchers to measure even deeper and faster with higher forces to enter an as-yet unexplored territory of cell biology and tissue biophysics.

We would like to thank the members of the NMSB lab for discussion and sharing unpublished material.