Optical Tweezers Across Scales in Cell Biology

Abstract

Optical tweezers provide a noninvasive approach for delivering minute physical forces to targeted objects. Controlling such forces in living cells or in vitro preparations allows for the measurement and manipulation of numerous processes relevant to the form and function of cells. As such, optical tweezers have made important contributions to our understanding of the structures of proteins and nucleic acids, the interactions that occur between microscopic structures within cells, the choreography of complex processes such as mitosis, and the ways in which cells interact with each other. In this review, we highlight recent contributions made to the field of cell biology using optical tweezers, and provide basic descriptions of the physics, the methods, and the equipment that made these studies possible.

The use of light for physical measurements and manipulations.

Photons, the essential particles of light, have no mass, but they nonetheless carry energy in the form of momentum. As they move through space and interact with the world, they can transfer this energy and therefore physically affect their environment (Figure 1A). Half a century ago, Arthur Ashkin and colleagues first used this characteristic to apply forces to and manipulate transparent objects on the micrometer scale. Originally interpreted as the harvesting of radiation pressure[1], this technology is now called optical tweezers (OT, see Glossary) or optical trapping[2, 3]. Since Ashkin’s original discoveries, OT has become an invaluable tool for creating or measuring small forces (in the range of pico-Newtons) within cells or in vitro preparations. In this review, we aim to deliver an introduction, accessible to cell biologists, to the physical principles underpinning OT and the ways in which this approach has shed light on diverse topics about cell structure and function. Other recent reviews have provided excellent perspectives on the physical application of OT [4, 5], the specific contributions of OT to molecular-scale studies [6, 7] and single cell biophysics [8, 9].

Figure 1: Photons at work.

(A) Schematic of a photon (white comet) being scattered as it travels through a transparent particle (blue sphere). As the photon enters and exits the particle, its direction and momentum change. The difference between its original momentum and its altered momentum is reciprocally transferred to the particle, resulting in a force F (orange arrow). (B) Schematic of a focused laser beam, comprising numerous photons, applied to a transparent particle at equilibrium in the center of the trap and (C) a transparent particle displaced from the center of the trap and undergoing a force drawing the particle towards the trap’s center. (D) Schematic of a folded protein attached to a trappable particle. The particle is centered in the trap in (i), so there is no net force placed on it by the beam. With increasing OT forces (ii-iii), the protein is gradually unfolded. (E) Schematic of two optical traps holding (i) and stretching (ii) a red blood cell. The careful measurement of OT forces (See Box 1) yields information about the protein’s secondary and tertiary structures (D) and the cell’s viscoelasticity (E).

OT can generate minute forces (on the pico-Newton to nano-Newton scale) by focusing light, in the form of a laser beam, tightly to a single spot, where an optical trap can occur (Figure 1B and C)[6]. The tight focus is generally achieved using a microscope objective with high numerical aperture (NA≥1.2), and the tightness of the focus, as well as the shape and material of the target object[10], are important for determining the OT forces that will result. For non-transparent objects, photons are absorbed, heating the object but also pushing it away from the optical trap[11, 12]. Objects with a high degree of transparency are therefore most suitable for OT. In terms of a target’s shape[13], curved surfaces induce strong beam deflection and therefore a large momentum transfer (Figure 1A), while more planar structures refract light less efficiently. It is therefore difficult to apply strong OT forces to structures such as membranes. Another important parameter is the target object’s refractive index relative to its surroundings[14]. Objects such as glass, plastic, and some cell structures and proteins, have high refractive indices, and are therefore readily controllable with OT. Objects with lower refractive indices, especially when the index is close to water’s, are more difficult to trap effectively in aqueous biological contexts. In some cases, particle coatings are introduced to allow more efficient and therefore stronger trapping capabilities[14, 15].

The particle’s position within the trapping beam is the final important factor. As shown in Figure 1C, for a spherical particle, the further away the particle is from the center of the trap, the larger the beam deflection is, and therefore the larger the resulting force. Above a certain distance, however, photons from the trap will start “missing” the particle, resulting in the decrease of momentum transfer, and eventually the particle’s escape from the trap. These relationships between particle positions within the trap and the OT forces that result have been explored in detail elsewhere [16–18].

There are several basic methods for delivering OT (Box 1), each suited to specific types of experiments. In some studies, it is necessary to direct OT to different positions within a stationary sample. This requires beam steering, which can be achieved using a beam steering mirror (manual such as gimbal mirrors[19], or motorized such as galvo mirrors [20]) or a spatial light modulator (SLM) (Box 1)[21]. Other studies require multiple traps to be placed on different objects or within the same object, and this can be achieved by methods such as beam splitting [19], time sharing [22], or holography[23].

BOX 1: Methods for delivering OT.

The nature of the desired OT measurements or manipulations dictates the complexity and configuration of the OT setup. Some common applications, with associated optical equipment, are summarized in this box.

Simple OT microscope (Panel A)

A simple OT setup is composed of two systems: a trapping system and an imaging system (Figure IA). The trapping optics comprise a laser beam (Figure IA, red) expanded by a telescope (two lenses positioned two focal lengths apart) to overfill the back aperture of a high numerical aperture microscope objective (trapping objective), a dichroic mirror to direct the laser beam towards the microscope objective, and the microscope objective itself.

The imaging optics comprise an illumination light (A, blue), a lens or another microscope objective (called a condenser) to create a collimated beam of light in the sample, the trapping objective, a lens to focus the sample image, and a camera. The imaging light (A, blue) passes through the dichroic mirror, rather than being reflected. The optics are placed such that the plane where the OT is located is imaged onto the camera, which allows the live visualization of the trapped object and its environment. Filters are often required in front of the camera to block the intense back-scattered light from the OT.

In this configuration, the laser beam is focused to a tight diffraction-limited spot. The OT within the sample chamber is stationary, and the physical displacement of the target particles or sample can be applied by moving the sample stage. If the goal is to rotate, rather than to apply a directional force to the target object, this can be done using a birefringent object and a circularly polarized laser beam. Such rotational forces allow for the optical alignment of the object [28] and viscoelastic measurements of the target’s medium[29]. A convenient method to displace the optical trap rather than the sample would be to introduce a steering mirror between the laser beam telescope and the dichroic mirror [30].

OT with force measurements (Panel B)

As we have discussed, many studies require the precise measurements of OT forces. These measurements can be achieved either by analyzing of the trapped particle’s absolute position on the videos captured by the camera, or by measuring the relative position of the trapping beam on a position detector[31]. In the latter case, a second dichroic mirror reflects the trapping laser beam through a lens to the position detector (typically a quadrant photodetector or a position-sensitive detector) (Panels B and C). The advantage of using a position detector is its high speed and precision compared to most cameras.

Holographic OT (HOT, Panel C)

Light-shaping approaches including holography make it possible to create multiple dynamic traps within a 3D volume. Practically, this involves adding a light shaping element, such as a spatial light modulator (SLM), to the OT pathway prior to the sample. The SLM shapes the beam’s wavefront, creating patterns of constructive and destructive interference within the sample, with bright spots at the sites of constructive interference serving as traps. The additional optical elements for this change are a mirror and a telescope (Panel C). This addition adds complexity to the setup, as the holograms displayed by the SLM need to be calculated. Methods and codes for performing these calculations have been shared and can be found online [32].

Microscopy methods that are compatible with OT

Due to their flexibility, OT systems are relatively easy to combine with existing microscope systems, and past studies have combined OT with techniques including epifluorescence[33], confocal fluorescence[34], TIRF and FRET [35], Raman spectroscopy[36], differential interference contrast (DIC)[37], interference reflection microscopy (IRM)[38], and super-resolution microscopy (STED)[39]. Commercial suppliers of optical traps also offer OT with IRM, confocal, STED, and others.

For many biophysical studies using OT, it is critical to quantify the force that is being delivered to the targeted object. In principle, the force applied to a particle can be gauged by measuring the net angle of the light after refraction by the object (Figure 1B and C), which can also be measured as a relative position of the particle to the center of the focused laser beam after passing through a lens. This measurement is usually made by collecting the beam in a high numerical aperture objective after it has passed through the sample[24]. In practice, measuring the relationship between the relative position of the particle and the net angle of the exiting beam is complicated by a number of factors, including the medium surrounding the particle, the properties of the laser beam, and optical elements including coverslips and microscope objectives. As a result, tailored calibration methods (reviewed comprehensively in [25]) must be applied before the trapping forces can be calculated from the object’s relative position.

The source of the OT beam is another important factor, and for most applications, OT is created using a continuous wave laser with a wavelength in the visible or near-infrared region. To limit laser absorption in biological material, wavelengths in the near infrared region are preferable [26]). The laser power needed usually sits between 10 mW and 1 W, but this varies dramatically depending on the target and the OT system’s optics. For biological applications, it is also important to consider the heating effect and potential damage caused by the beam (Box 2).

Box 2: OT Heat Effects and Phototoxicity.

While low doses of light, especially the near infrared light often used in OT, are generally well tolerated by cells, concerns about phototoxicity arise when traps are strong or used for long durations[96]. Studies in S. cerevisiae have shown that 60s of 25mW trapping using a 785nm diode laser and a 100x objective does not have discernible impacts on the cells’ health, but can delay subsequent budding [97]. A thorough analysis of a 1064nm laser’s impact on S. cerevisiae showed that 15 minutes at 19, 38, 76, and 95mW through a 60x objective led to no effect, delays in budding, 50% mortality, and 90% mortality, respectively[98].

Heating is one likely mechanism for this toxicity, and recent work has helped to quantify such effects during OT. Under typical parameters for OT (100mW 1064nm laser through 60x or 100x objectives) OT caused a microsphere-bearing aqueous medium to warm by 4–5 °C in focused locations, although these effects were highly variable depending on the sizes of the beads and the optical configuration [99]. More powerful OT (at or above 300mW 1064nm laser through a 100x objective) regularly damages or kills red blood cells, possibly by creating a thermal gradient across the membrane surface that causes membrane permeabilization and cell rupture [100].

Others have proposed to harness OT photodamage as a therapeutic agent. By scanning an OT to produce low frequency vibrations in cell nuclei, they produced necrosis in the leukemia cells under conditions that left healthy leukocytes intact [101]. They propose that nuclear vibration leads to disruption of the cytoskeleton and the membrane cortex, leading to irreversible membrane permeability and necrosis. Structural characteristics of leukemia cells, such as increased nucleus-to-cytoplasm ratios and greater cytoplasmic stiffness, are cited as possible reasons for these cells’ enhanced susceptibility to vibrational OT.

It is worth highlighting that photo- and thermal damage varies dramatically depending on the wavelength and power of the laser; the duration of the OT; the optical configuration; and the structure, physiology, and absorptive properties of the target cell. Researchers seldom measure the OT power at the specimen, so it is difficult to make apples-to-apples comparisons of OT damage across studies. As a result, the best practice is to make careful observations for evidence of photo- or thermal toxicity, and to take these effects into account while interpreting results from OT studies. There are also approaches for preventing the toxic effects of heating with OT, for instance by using gripper formations of dielectric beads to avoid direct light exposure[102], or by cooling the specimen using temperature-controlled stages [103, 104]. The impacts of free oxygen radicals that can arise from irradiation can be mitigated by using oxygen scavenging reagents[105, 106] or carefully selected laser wavelengths [107].

Of course, OT experiments must take all of these considerations into account, and studies of diverse cellular targets have led to a wide array of OT methods and instrumentation (summarized in Box 1)[18, 27]. Often the primary consideration for a physicist, however, is the size of the object being targeted, which has profound implications for the efficacy of OT, and therefore for the specific methods that will be most effective (see Table 1). From the cell biologist’s standpoint, the target’s size is a logical first consideration because most research questions spring from an interest in a particular cellular object, with a predetermined size. Given the different optical approaches and biological questions that apply to OT targets of different sizes, we will use the following three sections to explore recent work across three scales: molecular, subcellular, and cellular. We will then address the particular challenges of performing OT in vivo.

Table 1:

Challenges and methods for OT at different scales

Regime	Size Range	Examples of targets	Challenges to OT	Solutions	Examples of Applications	Example References
Molecular	<10nm	Proteins Nucleic acids	Targets are much smaller than the optical trap, and cannot be confined efficiently.	Tether targets to trappable beads	Binding and interaction force measurements between sub-micrometer elements Full energy profiles of a molecules’ tertiary structures Strength and stepping dynamics of molecular motors	[7, 10, 37, 117–121]
Subcellular	~10nm-1μm	Organelles Membranes Vesicles	Often smaller than the opticaltrap Flat shapes not amenable to OT Similar refractive index to surroundings	Tether to trappable beads Use adaptive optics to maximize OT force	Membrane tension, pressure, and stiffness Viscoelasticity of membrane and liquids Drug delivery	[74–80, 122, 123]
Cellular	~1–10μm	Whole cells	Thermal or photodamage resulting from OT laser	Minimize laser power and OT duration	Cells stress and health Cells interactions and behavior Growth and displacement forces	[91, 124–129]
In vivo	various	Whole cells Macroscopic biological structures (e.g. otoliths)	Structures are much larger than the optical trap, and cannot be confined efficiently Scattering and absorption in vivo causing difficulty in light penetration Stronger OT forces required to move massive objects	Use optimal wavelength of OT light for deep penetration Use adaptive optics to correct for scattering Deliver light with Optical fiber.	Cellular transplantation. Alterations of blood flow. Fictive sensory stimulation.	[108–110, 125–127, 130]

OT at the molecular level: studying proteins and nucleic acids

Molecular dynamics within cells involve interactions on the scales of angstroms to nanometers, and on nano- or milliseconds timescales. This regime is challenging for direct optical manipulation because the diffraction limit of conventional OT systems restricts the minimum trap size that can be achieved. A common solution is to use a trappable anchor particle attached to the target molecule (Figure 1D). By manipulating molecules indirectly through these anchors, OT can probe movements, interactions, and forces among molecules that cannot themselves be trapped.

Historically, and especially in recent years, the measurement and application of minute, controlled forces through OT have been important for describing the tertiary structures of proteins[40–45] and interactions between proteins and nucleic acids[46–52]. This section, however, will focus on dramatic recent progress toward understanding the molecular forces involved in cell division, both in the tethering and separation of sister chromosomes and chromatids, and in the molecular motors and cytoskeletal elements that mediate chromosome segregation and division itself. In many cases[53–56], in vitro assays have extended beyond measurements of individual motor proteins into more physiological preparations in which teams of motor proteins act on bundles of cytoskeletal components, thus giving a more realistic readout of their activities in intact dividing cells[57].

For mitosis to occur correctly and to avoid aneuploidy, the chromosomes must be carefully positioned between the centrosomes, the sister chromatids must be aligned and bound together early in mitosis, and then must be separated and pulled in opposite directions (Figure 2A). Each of these steps involves numerous minute and carefully orchestrated physical forces, and it is exactly these sorts of forces that can be effectively measured or applied using OT. In the past few years alone, OT studies have revealed several key mechanistic features of the mitotic machinery and the ways in which it applies the forces needed for faithful chromosome segregation[57].

Figure 2: Insights on mitosis, made possible by OT.

(A) A schematic cell that, for the purposes of illustration, shows different stages of mitosis (early to late, top to bottom). Letters and boxes refer to processes highlighted in panels (B-D). (B) A schematic representation of typical of OT experiments on motor proteins. In this case a MT is tethered to a bead, and OT of the bead measures the force generation or braking strength of the motor proteins tethering the trapped MT to another MT. This preparation is typical of recent studies in which teams of numerous individual motor proteins are studies working on their natural substrates, such as MTs [53–56]. (C) An illustration of the in vitro preparation for studying ultra-fine bridge (UFB) recognition complexes (ref [65]). Pairs of beads allowed for OT tension to be applied to the synthetic UFB linking them. Various protein complexes (small circles) were tested for their effects on the synthetic UFBs’ stability and strength. (D) Laser cutting of one chromatid, which produces a fragment tethered near the telomere of the sister chromatid (right, tether not shown), leads to aneuploidy if the fragment is pulled in a cross-polar direction into the wrong daughter cell. This preparation serves as a model for naturally occurring aneuploidy and for the roles that tethers play in chromosome segregation and mis-segregation. OT can be used to pull the fragment away from its sister chromatid, thus allowing measurements of the strength and progressive dissolution of tethers as mitosis proceeds [64]. (E) With the NDC80 complex (green attachment) bound to an anchored MT and a trappable bead, OT was used to measure the forces harnessed by NDC80 as the MT depolarized, helping to explain how this kinetochore protein complex assists in transporting chromatids to the spindle poles during mitosis [68].

An early step in mitosis is the establishment of two opposing poles for the mitotic spindle, and the force pushing poles apart largely comes from the Kinesin-5 Eg5 motor protein. Eg5 is a plus-end motor that pushes two antiparallel microtubules along each other, allowing polar microtubules (MTs) sprouting from opposite centromeres and overlapping in the spindle equator region to be pushed in opposite directions (Figure 2B). An in vitro study of these motors used OT of tethered beads to measure the direction and force of Eg5-driven MT movements, as well as another motor protein, the Kinesin-14 HSET’s, influence over these forces[58]. In contrast to Eg5’s powerful and unidirectional force generation, HSET was found to slide across antiparallel MTs in either direction (thus stabilizing regions of MT overlap), and also to moderate the force delivered by Eg5 (presumably to balance the forces pushing the spindle apart and pushing it together). The correct balance of these forces is necessary for the establishment of two opposing spindle poles, which sets the stage for chromatid segregation.

The next step is the alignment of the chromosomes at the equator, and it is important that sister chromatids remain bound until this alignment is complete. It is only then that an orchestrated separation of the chromatids proceeds, providing one copy of all genetic material to each daughter cell [59]. When structures binding sister chromatids, such and mitotic tethers[60, 61] and ultra-fine bridges (UFBs)[62, 63], fail to release, chromatids or their fragments can be pulled in a cross-polar direction to the wrong pole, leading to aneuploidy. This controlled release of sister chromatids was addressed with an innovative combination of laser cutting of chromatids and tethers, and OT of the resulting chromatid fragments [64]. Predictably, when tethers between chromatids were cut, there was a reduction in cross-polar forces (Figure 2D). Furthermore, following laser cutting of chromatids, fragments attached to the kinetochore consistently moved toward the correct pole while fragments attached to the telomeres at the ends of the chromatid were often dragged in the cross-polar direction. These experiments showed that the tethers provide a binding force between the sister chromatids, that they are located at the tips of the chromatids, and that this binding must be overcome by forces applied to the kinetochore in order to segregate the chromatids correctly.

UFBs are connections comprising DNA and proteins that link sister chromatids [62, 63]. The assembly, strength, and disassembly of these bridges has been addressed using pairs of trappable beads linked by synthetic UFBs [65]. The OT served two purposes in this preparation. First, it permitted controlled movements of the preparation through different chambers of a microfluidics device, allowing for the sequential formation of protein complexes on the DNA bridges. Different fluorescent labels were used on different UFB proteins in the microfluidics device, allowing the exact composition of the synthetic UFBs to be controlled and validated (highlighting the compatibility of OT with fluorescence microscopy). Second, by applying opposite forces on the two beads, OT tested the strengths of the bridges with different protein compositions and DNA configurations (Figure 2C). While the biochemical fundamentals of UFBs had previously been described, this work showed the protein composition, stoichiometry, and sequential construction of UFB recognition complexes. The work also helped to explain how these protein complexes disassemble UFBs without creating uncontrolled DNA breaks just prior to cell division.

Segregation of the chromatids depends on the force applied by MTs to the kinetochores of each chromatid. This process does not, however, directly involve molecular motors; the force is generated as the kinetochore maintains contact with the depolymerizing end of MTs as they shrink back toward the spindle pole [66, 67]. The binding between kinetochores and MT tips is mediated by the nuclear division cycle 80 (NDC80) complex of proteins, but until recently, it was unknown whether this complex simply pulls the kinetochore passively along the MT as it depolymerizes, or whether it harnesses the energy released in MT depolymerization to power its translation down the MT lattice. This issue was addressed by assembling kinetochore modules with NDC80 complexes on glass beads, by applying OT to those beads as they translated along depolymerizing MTs, and by measuring the forces that were pulling the beads along the depolymerizing MTs [68] (Figure 2E). The results supported the idea that the kinetochore is actively driven down the MT using the force released by depolymerization, thus providing a major portion of the force required to move each chromatid into the appropriate daughter cell.

The aforementioned Kinesin-5 Eg5 motor protein pushes antiparallel polar MTs across each other, thus providing much of the force for establish spindle poles and separating sister chromatids [53, 69]. It is surprising, therefore, that mammalian cells exposed to Eg5 poisons can survive and divide. In these cells, it is another motor protein, Kif15, that redistributes itself from its normal location on kinetochore MTs to play a wider role in the mitotic spindle[70, 71]. OT of glass beads attached to Kif15 motors allowed for a structure-function analysis of the Kif15 protein and its ability to produce sliding forces in antiparallel pairs of MTs[54]. This work described the step size and force generation of the motor protein, accounting for its ability to expand its role in spindle dynamics and cell division when Eg5 is missing.

Finally, in-vitro OT has been used to describe the force generated by the Kinesin-14 motor protein Ncd in the mitotic spindle. It was known that Ncd dimers actively step on one MT with their motor domains while sliding diffusively along another MT with their passive tail domains [56]. Recent OT work showed that Ncd delivered rapid and powerful stepping on MTs when it was immobilized, but that most of this power was lost when Ncd was linking two antiparallel MTs [55]. Interestingly, when two linked MTs were non-aligned, the diffusive tail was less prone to slippage, and force generation between the MTs was restored [55]. This means that Ncd applies forces to non-aligned MTs, but not to antiparallel MT pairs. Suggested roles for this motor, therefore, include the transport of new unincorporated MTs, the pliable and force-free linking of antiparallel MTs, or the rearrangement of nonaligned MTs into an antiparallel orientation.

Subcellular OT to study membranes, stiffness, and fluidity

Just beyond the molecular scale sits a subcellular scale where membranes, organelles, vesicles, and other structures perform many of the crucial functions of cell biology. As a result, great efforts have been made toward observing and manipulating these structures. Many of these studies include the use of OT[72, 73]. While OT can operate on the scale of tens to hundreds of nanometers in a vacuum[5] or low scattering media[2], OT of biological systems on these scales is challenging. This is due, in part, to the scattering and absorption of light in cells and tissues, but also to the irregular shapes of many biological target objects, and especially to the small differences in refractive indices between most cellular structures. The target object and its surroundings will therefore dictate the feasibility of direct or indirect (bead tethered) OT within the system. Regardless of whether the trapping has been direct or indirect, OT has proven to be particularly useful in studies of cell signaling pathways; cellular interactions and behavior; and membrane tension, pressure, and stiffness[74–80].

Mechano-transduction is a process by which mechanical forces from the environment elicit intracellular signaling and adaptive responses in the affected cell. The sensitivity and nature of these responses can be probed using controlled OT forces that mimic naturally occurring mechanical stresses. For instance, an optically-trapped bead can be used as an indentation tool, in one case applying controlled piconewton forces in vitro to mouse neuroblastoma cells [74]. This approach showed that soft mechanical stresses (equivalent to the force produced by a moving lamellipodium) induce intracellular calcium signaling [74], suggesting that mechano-transduction pathways are sensitive to physiologically-relevant mechanical forces in individual cells.

OT has also been employed to study cells’ impacts on their surroundings, specifically regarding how contracting cells influence the tension and stiffness of the extracellular matrix (ECM) that surrounds them (Figure 3B) [81]. The approach involved embedding trappable beads in synthetic ECM in the vicinity of cells that were exerting large mechanical stresses. OT measurements of the forces necessary to trap or move ECM-embedded beads revealed the relative stiffness of the ECM at various locations around contracting and relaxed cells. The results show that cell-induced stresses result in significant stiffness gradients across the extracellular matrix, including at magnitudes and distances that would allow nearby cells to sense and respond to these gradients [81].

Cancer cells’ interactions with their surroundings are of particular interest, as changes in the cells’ stiffness, membrane fluidity, and mobility have implications for metastasis. OT is particularly well suited to measuring stiffness (through in an indentation assay with a trappable bead) and has specifically been used to measure how stiffness varies across various breast cancer cell lines and in different contexts [78]. While these different lines of cells have characteristic levels of stiffness in isolation, their stiffness changes dramatically when they are placed on synthetic ECM of various rigidities. More aggressive cells, in particular, gain stiffness when on matrix, and this stiffness increases further when they are in contact with other cells in the matrix [78]. This shows that the mechanical structure of the environment is a key factor in these cancer cells’ stiffness, and therefore their potential for mobility. By extension, this mechanism could mediate intercellular mechanical communication and collective durotaxis [78].

In another study of cancer cells’ mechanical properties, trappable beads were embedded in human cancer cells’ membranes, and OT was used to stretch these tethers away from the cell in discrete steps [82]. This allowed the membrane fluidity to be modelled based on the relaxation curves of the membrane after each stretch. This study found that transfection with a cancer-associated microRNA led to an increase in membrane fluidity that paralleled increased migration and invasion by the cells in vitro [82]. A similar study used OT to measure the force profiles of membrane tethers pulled from human ovarian cancer cells [83]. The results indicated that the force required to separate the membranes of cancerous cells from the cytoskeleton was significantly lower than the force required for non-cancerous cells, again with implications for metastasis [83]. –These two studies, along with others studying molecular interactions in cancer cells [84, 85], demonstrate OT’s particular suitability for measuring cells’ mechanical properties, providing useful diagnostic methods for the metastatic potential of cells, and for the potential effectiveness of targeted therapeutics.

Optical manipulation of whole cells

As we move to the scale of whole cells, we move into a range where the objects’ physical properties are often sufficient to allow efficient optical traps. As such, most OT studies of whole cells do not use beads or other anchors to facilitate OT. The cells themselves are trappable and OT-based studies of their interactions or behavior can be performed directly[86, 87].

In some cases, OT simply provides a tool for capturing, transporting, and precisely placing individual cells, allowing for the controlled construction of arrays or assemblies of selected cells. This was an effective approach, for example, for capturing specific types of neurons from a population of dissociated tiger salamander retinal cells and for delivering these selected neurons, one by one, onto precise positions of a multielectrode array (MEA) [88]. This permitted the controlled assembly of an in vitro circuit mimicking the circuits found naturally in salamander retinas. Electrical stimulation of targeted neurons within the network, using the MEA device, led to activity across the network that could be recorded using calcium imaging, thus revealing the functional properties of this synthetic network [88]. Another proof-of-principle study used photonic-crystal-enhanced OT to trap individual human pluripotent stem cells, delivering them precisely into assemblies of 8–10 cells, and using OT forces on the assemblies to test their adhesion to their substrate [89]. In other cases, the objective has been the opposite: to study individual cells in isolation. One example has been the creation of cultures of Salmonella bacteria, each derived from an individual Salmonella cell. This was made possible by a digital microfluidics device in which each well contained one magnetic bead fused to single Salmonella cell. The cell in each well was assessed for fluorescent labelling, after which the authors optically trapped the beads attached to fluorescent cells, isolated them, and used them to seed a colony [90].

Finally, OT can be used to probe the physical properties of whole cells, and this has been particularly beneficial for the study of red blood cells (RBCs)[8, 91], which rely on being flexible and resilient as they undergo strong fluid forces and squeeze through capillaries. When RBCs lose their deformability or aggregate inappropriately, this can block capillaries and other small vessels, with dire consequences for circulation. Flexibility can be measured by placing two optical traps, pulling in opposite directions, on opposite sides of an individual RBC (as depicted in Figure 1E). The degree to which an RBC deforms can be expressed as a deformability index (DI), which is the cell’s proportional increase in diameter when stretched in this way. This approach has been used, for instance, to show that RBCs from diabetic patients have lower DIs than control RBCs do [92], suggesting one possible mechanism for vascular occlusive disorders in this cohort. Using a similar approach with a single OT, it has been shown that Atorvastatin, a commonly prescribed statin drug, directly increases the elasticity of RBCs when applied in vitro [93]. When studying aggregation of RBCs, OT can be used to measure the interactions among RGCs and the forces that cause them to adhere to or be repelled by one another (Figure 3C) [76, 94]. This approach has been used to study whether nanoparticles, which could be embedded in RBC membranes as a drug delivery method, affect the aggregation properties of the RGCs carrying them. Using OT to assess aggregation forces among RGCs, researchers have demonstrated that nanodiamonds, but not other types of nanoparticles, cause increases in RGC aggregation forces and the resulting formation of larger RGC aggregates in vitro [95].

Optical manipulation in vivo

Ultimately, analyses of biological processes on the microscopic scale are pursued to better understand large scale phenomena, including at the organismal scale. These topics might include cell migration, system-scale physiology (blood flow, for instance), and sensory perception. Such studies, by their very nature, are best performed in vivo. The scattering properties, size, and opacity of many model organisms provide obstacles to in vivo OT, although methods to reduce light distortion can mitigate these effects in some cases [108–110].

One example of such in vivo work is the use of OT for the targeted transplantation of single cells. This approach has been used to study the behavior of individual hematopoietic stem cells (HSCs) when transplanted into bone marrow (BM). After capturing single stem cells with OT, performing laser microsurgery on the bone to expose BM in live mice, and seeding the individual stem cells into the BM, researchers could repeatedly return to the site to image the engraftment of the HSC and its interaction with the BM [111]. This provides proof of principle for a method that could be used in the future to study the reconstitution of blood cells with transplanted HSCs or the mechanisms underlying tumor formation by cancer stem cells.

Making use of the optical properties of small transparent zebrafish larvae, researchers have investigated cardiovascular development by controlling blood flow with OT. Specifically, they modelled hereditary hemorrhagic telangiectasia (HHT), which can cause disruption of blood flow and internal bleeding [112]. HHT model larvae have little or no blood flow through their small intersegmental vessels (ISVs), and OT was attempted to alter the RBCs’ trajectories and redirect them into empty capillaries. This approach led to a significant increase in the number of RBCs entering the ISVs, demonstrating the effectiveness of these dynamic traps. In the future, further advances in this method could assist in studying blood cell-to-vessel interactions and in the design and implantation of synthetic capillaries.

Optical trapping on an even larger scale has helped to reveal the brain-wide networks of neurons that allow for hearing and vestibular perception[20, 113, 114]. In larval zebrafish, both senses are reliant on otoliths, or ear-stones, which are disk-shaped calcium carbonate crystals roughly 30–50μm in diameter. Rapid vibrations of the otoliths relative to the rest of the ear provide the fundamental signal for hearing, while slow gradual movements underlie vestibular perception. Because they are transparent and have a high refractive index, otoliths are readily trapped, producing piconewton-scale forces that simulate natural acceleration and elicit behavioral responses normally used by the animal to counteract rolling movements[113]. The use of OT provides two benefits in this context. First, out of the four otoliths in zebrafish larvae, OT permits the targeted stimulation of individual otoliths or pairs of otoliths (with dual OT), thus revealing their specific contributions to behavior. Second and more importantly, it can drive vestibular processing in the brains of stationary animals, facilitating calcium imaging in the brain. Such OT of the otoliths, combined with light-sheet microscopy for calcium imaging, revealed patterns of brain-wide activity during vestibular perception[114]. Finally, oscillating OT can produce vibrations of the otoliths at controlled frequencies, allowing the controlled delivery of fictive auditory (high frequency) and vestibular (low frequency) stimuli, and further permitting calcium imaging to describe their distinct but overlapping sensory pathways in the brain[20]. The observed auditory responses in the brain, despite the rudimentary nature of the Bio-Opto-Acoustics (BOA) stimuli in this initial study, closely resembled elements of brain-wide auditory processing networks stimulated by actual sound [115, 116].

Concluding Remarks

OT has provided valuable insights into biological form and function on scales ranging from single molecules to whole organisms (see Table 1). In many cases, basic OT has been cleverly extended by advanced optical methods, thus increasing its reach and precision. However, there is still room for improvement, and recent advances in optics and computation hold the potential to drive the next generation of OT.

Computational modelling of OT has been used for designing efficient optical traps, understanding optical forces and torques, and modelling the dynamics of the objects being trapped [131]. Often, for the purposes of modelling forces, the object is represented as a sphere or given another simple geometry. Similarly, the object’s environment is usually simplified as non-scattering and homogeneous. This is because the complexity of the object and environment considerably increases the complexity of force calculations and associated computational time. Therefore, increased computational power, as well as advances in the algorithms used to model optical forces, will be beneficial to ongoing advancement of OT technology. As an example, recent developments in machine learning have led to faster methods for simulating light scattering [132] as well as simulating particles in OT[133]. Further progress along these lines will be important for designing optical traps with stronger trapping forces or more nuanced and physiological force delivery.

While computational tools have enabled more elaborate beam shapes for OT, this does not resolve the problem of scattering that occurs once the beams enter biological tissue. As discussed earlier, wavefront shaping is a promising method for counteracting scattering and reaching deeper into biological systems with OT. While light shaping methods have traditionally been applied to the improvement of imaging quality, they should prove just as suitable for refining OT in the future.

Another way to reach deep into a scattering tissue is to use an optical fiber, and recent advances in the materials and structures of fiber optical tweezers (FOT) suggest that this could be the most realistic future direction for delivering OT into highly scattering tissues like the brain[134]. Specifically, the development of multimode fibers and computational tools that allow real-time beam shaping [135, 136] hold great potential for applying OT to neuroscience. Similarly, the ongoing development of more sophisticated fibers, including hollow core and graded index fibers, as well fibers with integrated lenses[137], will extend the power and precision possible for FOT.

As physicists and biologists continue to adapt methods such as adaptive optics and light shaping from microscopy, adopt more advanced fibers and detectors, and utilize more powerful computers and improved algorithms for modelling light, we expect OT’s range and sensitivity to continue to expand, opening up further questions (see Outstanding questions box) in cell biology and beyond.

Outstanding questions.

• Is OT the best method for characterizing mechanobiology? OT has been enormously valuable in mechanobiology, especially at microscopic scales. Nonetheless, each study needs to identify the strength and limitations of individual and combined optical methods to judge which will provide the best and most reliable results for each experimental context.

• Would more advanced modelling of scattering tissue support adaptive optics? Adaptive optics, which have the capacity to counteract scattering caused by biological tissue, can be informed by prior modelling or by empirical refinement in each specimen. The randomness of biological matter may never allow effective correction of distorted light propagation based on computational modelling, and in situ measurement and correction methods may prevail.

• Can optical fibers be further refined for OT? Optical fibers provide a means of delivering light deep into a specimen without scattering, but basic optical fibers do not deliver optimal OT. Future advances in fiber technology may permit fine control over trapping light for effective FOT deep in biological tissue.

Figure I: Common elements constituting optical tweezer setups.

(A). a simple OT microscope composed of trapping optics (along the red beam) and the imaging optics (along the blue beam). C: Camera, La: Laser, L: Lens, M: Mirror, DM: Dichroic Mirror, O: Objective, S: Sample, LS: Light source. (B). OT microscope composed of trapping optics and force measurement optics (along the red beam) and the imaging optics (along the blue beam). PD: Position Detector. (C). OT microscope composed of highly flexible trapping optics and force measurement optics (along the red beam) and the imaging optics (along the blue beam). SLM: Spatial Light Modulator.

Highlights.

• Optical tweezers can measure or apply minute forces inside cells noninvasively.

• Recent developments in optical tweezer technology have allowed insights into the structure, function, and behavior of cells.

• Particular progress has been made in understanding mitotic machinery, cellular viscoelasticity, and metastasis.

• Emerging methods in optical physics promise to extend the breadth and utility of optical tweezers in the years to come.

Glossary box

Aneuploidy

a condition where a cell has missing or additional copies of a chromosome or part of a chromosome

Chromatid

one of two identical halves of a replicated chromosome

Diffraction limit

smallest spot size achieved by an optical illumination instrument, or resolution of an optical imaging instrument

Durotaxis

a form of cell migration in which cells are guided by the stiffness gradients of the extracellular matrix

Holography

the study or production of three-dimensional images by modifying the amplitude and phase distributions of a light wave

Kinetochore

a protein assembly on the centromere of a chromosome, providing load-bearing attachments for sister chromatids during cell division

Metastasis

the process by which cancer cells migrate away from an initially formed tumor

Mitosis

a type of cell division where a single cell divides into two daughter cells, each having the same number and kind of chromosomes

Mitotic spindle

a cytoskeletal structure that mediates chromosome segregation during mitosis

Optical fiber

a thin, flexible, transparent fiber with a glass (silica) or plastic core through which light signals can be sent with very little loss

Optical tweezers

optical instruments that use highly focused laser beams to create optical traps

Optical trap

a focused spot of light that draws in and holds transparent objects of high refractive index

Refractive index

a value related to the refraction angle of a light path when entering a material and calculated from the ratio of the speed of light in a vacuum to that in a second medium of greater density

Spatial light modulator

optical instrument that imposes spatially varying modulation on the wavefront of a beam of light. This can result in a hologram once this beam is subsequently focused

Wavefront

an imaginary surface over which the phase of the wave is constant